December, 1934
INDUSTRIAL AND ENGINEERING CHEMISTRY
capacity-to produce acid for use in their own refineries. Kow i t develops that the modern trend in oil refining is to adopt processes that use organic chemicals of various kinds or clay or zinc chloride or liquid sulfur dioxide (a). Those refiners that installed sulfuric acid plants, a t large capital outlay, may soon be ruefully concluding that it would have been better to have purchased their acid during the past few years, and thus to have left their way clear for the adoption of an organic solvent process of refining, without the sacrifice of an acid plant. Enough possible instances have been cited to show that each proposal of this sort is a separate problem, to be dealt
1283
with individually on its merits. It seems clear that, because a manufacturer consumes even a large quantity of sulfuric acid, i t is not therefore a foregone conclusion that it will’be profitable for him to build, in preference to buying. NOTE: This paper is part of the manuscript of an A. C. S. Monograph on “Sulfuric Acid Manufacture,” now in preparation.
LITERATURE CITED (1) Chem. & Met. Eng., 41, 35 (1934). (2) Keith, P. C., Jr., and Forrest, H. O., paper presented before New York meeting of Am. Inst. Chem. Engrs., May, 1934.
RECEIVED September 20, 1934.
Studies in the Vulcanization of Rubber V I . Thermochemistry JOHNT. BLAKE,Simplex Wire & Cable Company, Boston, Mass. on the assumption that vulcaniInoestigators generally agree that there is a subzation involves two successive the heat of vulcanization stantial ecolution of heat during the vulcanization processes: (1) the disaggregawas studied over the entire of rubber to f o r m ebonite. There are, however, tion of the rubber micelles, and sulfur range by heat-of-combusdifferences of opinion regarding the thermal (2) the reaction of this disagtion measurements. Data were gregated material with sulfur. changes in the soft rubber range (0 to 8 per cent also obtained on vulcanization The equation gives the same with m-dinitrobenzene and selesulfur), and past data are inconclusive. general type of curve as the exnium. Although it had been Qualitative data have been obtained indicating perimental ones. In a second known qualitatively that there that a small but definite heat evolution occurs paper (12) he extended his work was a heat evolution during a t during vulcanization with one per cent sulfur to include a c c e l e r a t e d comleast a portion of sulfur vulcaniand larger amounts are evolved at higher perpounds. In all of these cases zation, definite values were thus the maximum temperature also obtained for the first time. centages. T h e effect of a n accelerator and the occurred when about half the The results by this heat-ofaction of dinitrobenzene and selenium have also sulfur had combined. The concombustion method are, howbeen investigated. clusion was drawn that his hyever, of limited accuracy where The results tend to confirm the preciously pothesis of the mechanism of small heat changes occur. The proposed theory that vulcanization consists of two vulcanization is correct. method consists in the determinRiding (10) studied temperaing of heat-of-combustion values successive reactions: The Jirst, or soft rubber ture changes in the center of a on a rubber compound before react ion, is aflected strongly by accelerators and mass of compound during the and after vulcanization. Their involves little or no heat interchange; the second, formation of ebonite. He found difference represents the heat or ebonite reaction, is comparatively insensitive to that t h e e v o l u t i o n of heat content change brought about accelerators and strongly exothermic. began a t the same r e l a t i v e by vulcanization and involves state of vulcanization with all the subtraction of values of the mixtures-a coefficient of vulcanization of &and that about 10,000 calories per gram. From the data obtained the conclusion was drawn that little between 8 and 40 per cent sulfur on the rubber the reor no heat is evolved below 6 per cent sulfur but that there is a action was exothermic. The results were essentially the steadily increasing heat evolution beyond this to 300 calories same as those of Perks (9). Riding found an increased per gram of compound at 32 per cent sulfur. The first three temperature change in the presence of accelerators but points of these data ( 2 , 4, and 6 per cent sulfur) gave an concluded that it was due to the greater rapidity of the average value of 6 and a mean variation from zero of 18 reaction. Hada, Fukaya, and Kakajima (5) used the heat-of-comcalories per gram, representing approximately 0.05 and 0.18 per cent of the total heat of combustion-probably the order bustion method to study the heat of vulcanization. They of the accuracy of the determinations. In the presence of an claimed that the data of the present author are composite accelerator (diphenylguanidine) a definite heat evolution was values containing: (1) heat of reaction between pure rubber found beyond 4 per cent sulfur, although none appeared at 1 and sulfur, ( 2 ) heat of reaction between resins and sulfur, and 2 per cent. Vulcanization of rubber with m-dinitroben- (3) heat of reaction between protein and sulfur, and (4)heat of combustion of free sulfur. This is true of the first three zene and selenium did not seem to give an evolution of heat. Toyabe (11) has since measured the increase in tempera- items. There was little or no free sulfur in the samples used. ture in the center of three rubber samples during vulcaniza- They claimed also that the heat of formation of the nitric tion. He found that the maximum temperature occurred and sulfuric acids produced during combustion should have after about the same time interval and also determined that been corrected for, not realizing that the method of running in each case this point coincided with the combination of blank determinations probably automatically cancels all about half the sulfur. An equation was set up for the reaction such items. S A PREVIOUS paper (8)
1284
INDUSTRIAL AND ENGINEERING CHEMISTRY
Vol. 26, No. 12
an alkaline material in the bomb, a n d a n a l y z e d completely the bomb contents for each determination. Their careful and extensive c o r r e c t i o n s should give accurate figures for the heat of vulcanization. The check values a t 2 and 6 per cent sulfur vary, however, over 70 per cent, allowing no conclusions to be drawn regarding the soft rubber -,00 IO 20 30 range. They deduced that the heat of vulcanizaFIGURE2. HEAT OF VULCANIZA-tion is a straight-line function of the sulfur content TION DATA(BLAKE) over the entire range. The heat-of-combustion method promises nothing further in the low sulfur range. FIGURE 1. HEATOF VULCASIThe temperature rise method offers a t l e a s t a ZATION DATA (HADA,FUKAYA, qualitative picture of the thermal changes during AND NAKAJIMA) vulcanization. m
y I
They purified the ram rubber b y a s o l u t i o n and precipitation method, removing the FIGURE 3. MAXIMUM TEMPERAprotein w i t h trichloroTURE RISE vs. PERCENTAGE SULacetic acid. The vulcanFUR (HADAAND NAKAJIMA) ized rubber was f r e e d of resins and free sulfur by extracting with acetone, They calculated and applied the abovementioned corrections, and t h e i r r e s u l t s are p l o t t e d i n F i g u r e 1, which may be compared with the author's d a t a as plotted in Figure 2 (the point a t 6 per cent sulfur is probably definitely in error). Their FIGURE5 . TEMPERATURE-TIME RELATIONSHIP FOR 2 PER CENT corrected data do not SULFUR VULCANIZATION AND BLANK seem to confirm ordinary experience in vulcanization since, among other facts, they obtain an endothermic reaction of 400 calories per gram a t 1 per cent combined sulfur. Two of these authors, Hada and Nakajima ( B ) , studied the question further. They resorted to the method of observing temperature changes resulting from vulcanization. They heated a small sphere of rubber-sulfur mixture, with a thermocouple in the center, in an oil bath, together with a similarly equipped sphere of the same composition which had already been vulcanized and, therefore, was thermally inactive. The temperature difference between the two samples was plotted against time, as the oil bath was heated to and held a t 160" C. The results are decidedly a t variance with their previous data, since a t no time during the vulcanization is there any evidence of an endothermic reaction. The maximum temperature increases for the different percentages of sulfur on the rubber have been estimated from their curves and are plotted in Figure 3. Lewis (8) concluded that the heat evolved was a straight line through the origin although he presented no new data. Jessup and Cummings ('7) have recently determined the heat of combustion of raw and vulcanized samples of purified rubber. They used a method with many refinements, placed
THERBIOCHEMICAL EXPERIMENTS The temperature rise during vulcanization in the center of a mass of rubber may be magnified by increasing its volume and by heating a t a relatively high temperature. The first reduces heat los,res b y c o n d u c t i o n and the second causes the v u l c a n i z a t i o n to proceed with greater rapidity, thereby forcing the heat evolution to take place during a short period of time. A number of systems for studying these temperature rises have been examined. Rubber compounds have been heated in iron molds of varying sizes (one contained over 2 kg. of rubber) in a vulcanizing press. The time-temperature relationships were followed by means of several series connected thermocouple junctions placed in the center of the mass. Similar measurements have been made on different sized cylinders of rubber compound in test tubes placed in constant-temperature baths. It was decided that the following procedure offered a convenient but sensitive method for investigating these effects: A cylinder of rubber compound was prepared by rolling a milled sheet of the appropriate dimensions. A copper-constantan thermocouple junction was placed in the center of the cylinder during its formation, about 50 mm. from one end. The assembly was slipped into a 38 X 200 mm. Pyrex test tube in such a manner that the thermocouple leads came out of the bottom of the sample. This prevented serious displacement of the junction if the rubber expanded during vulcanization. By forming the cylinder of freshly milled rubber and inserting immediately in the tube, the natural longitudinal contraction caused the sample to expand laterally and fit the tube tightly. The other thermocouple junction was placed with the tube in a 2-liter flask containing condensing c . P. aniline vapor (184.4' C . ) . The temperature difference between the center of the cylinder of rubber and the constant-temperature bath was measured at timed intervals on a k e d s & Northrup potentiometer, using the data of Adams (1). Compounds of rubber (smoked sheets) and sulfur containing 1, 2, 3, 4, 6, and 8 per cent sulfur on the rubber were mixed and heating curves obtained. A representative curve is illustrated in Figure 4 ( 2 per cent sulfur). I n order to evaluate the temperature rise due to the vulcanization reaction, it is necessary to compare the vulcanization curve with that obtained when the material is thermally inactive. This latter may be obtained experimentally by reheating the same sample after the vulcanization has taken place. It is well known, however, that if a mass of thermally inactive material is heated by a constant temperature source, the logarithm of the unaccomplished temperature change a t any point within the mass is a linear function of time. This may be represented by the equation:
+
log (Hm - H) = A Bt where A , B = constants H = temp. at time t H , = temp. after an infinite time
I N D U S T R 1-4 L A N D E N G I N E E R I S G C H E )I I S T R Y
December, 1934
A blank heating curve may thus be d e t e r m i n e d by plotting the experimentally determined unaccomplished temperature changes ( H , - H ) on a logarithmic scale and extrapolating the straight-line portion representing the heating rate of the material before the start of the reaction. d transfer of this extrapolated line to the original data allows evaluation of the changes occurring during the heating. The curve obtained in this manner is perhaps a more exact blank than that ob-
4
6
1 3(c-
FIGURE7 .
TEMPERlTURE
PHENYLGUAVDINE-
in the sample o c c u r r i n g during the first heating would be reflected in the s e c o n d s e t of data. That the two are not materially different, however, is illustrated jn Figure 5 . The experimental and extrapolated lines are practically parallel and the data clearly indicate the logarithmic character of the heating (the displacement of one line from the other is o b v i o u s l y due to the use of a different time origin which is of no consequence since it is time intervals only that are significant), The difference b e t w e e n t h e t w o c u r v e s (Figure 4) is an i n d i c a t i o n of the thermal change due to the v u l c a n i z a t i o n reaction. These differences for the rubber-sulfur compounds are plotted in Figure 6. To study the effect of an accelerator, the following compounds: Smoked sheets Zinc oxide
100 5
Diphenylguanidine
100
5
100 5
DI-
30
IO 9
B9
FIGURE8.
(I). P. G.)
p-Nitrosodimethylaniline Selenium
Tetrarnethylthiuram disulfide Selenium
TEMPERATURE-TIME REL.4-
TIONSHIP FOR
SELENIUM COMPOUND
33 36 39
12 15 16 21 24 27 30
FIGURE 9. TEMPERATURE-TIME R~~~~~~~~~~~ FOR TETRAMETHYLD
THIURAM
~
~
COMPOUND
0.75
3
12.5
The data are plotted on a logarithmic scale in Figure 8, and the straight line illustrates clearly that no measurable heat is evolved during the vulcanization. It, is possible also to vulcanize rubber satisfactorily with selenium and tetramethylthiuram disulfide. The heating curve of the following compound was determined : Smoked sheets Zinc oxide
RISE WITH
~ C C E L E R A T E DCOVPOUUDS
20
containing 1, 2, 3, 4, 6, and 8 per cent sulfur were mixed and their curves determined (Figure 7). Selenium is a vulcanizing agent for rubber, but ebonite cannot be formed through the use of selenium alone. The following compound was mixed and heated: Smoked sheets Litharge
1285
1.5
0.5
The logarithmic plot in Figure 9 indicates that in this vulcanization there is also no heat evolved or absorbed. Selenium has the ability to accelerate the vulcanization of rubber-sulfur mixtures, the selenium combining with the rubber to some extent. It is possible that selenium sulfide is an intermediate product and is a vulcanizing agent. Curves have been obtained on a compound containing 2 per cent sulfur on the rubber and one containing, in addition, 5 per cent selenium (an atomic equivalent). They are plotted in Figure 10. It is evident that, although no measurable heat is evolved when rubber is vulcanized with selenium alone, the presence of selenium results in a greater evolution of heat when the rubber is vulcanized with 2 per cent sulfur. One of the classical methods of vulcanizing rubber was discovered by Ostromuislenskii; i t involves the use of nitro compounds, and the reaction is accelerated by litharge and
b v carbonblack. T i e following compounds were mixed and heated: I
Mop
I
I
,
I
I
I
l
l
I1
Smoked sheets 100 100 Litharge 5 Carbon black 5 m-Dinitrobenzene (D. N . B . ) 3 3
FIGURE10. SULFUR
TEMPERATURE RISE WITH .4ND SULFUR-SELENIUM COUPOUNDS
The results are plotted in Figure 11 and there is considerable difference in the amount of heat evolved. It is possible that the litharge undergoes an exothermic reaction that is independent of the vulcanization. It is interesting to plot the maximum temperature rise data in Figures 6 and 7 against the percentage sulfur (Figure 12). Hada and Kakajima obtained only about a 1O C. rise in temperature a t percentages of sulfur up to 10. The sensitivity of the present method is so much greater that there is nearly a 2' C. rise a t 1 per cent sulfur and a 69" C. rise a t 8 per cent sulfur. There is clearly a heat evolution a t all percentages of sulfur. Beyond about 3 per cent the temperature rise becomes progressively larger and is approximately a straightline function of the sulfur. Below 3 per cent sulfur the lines curve to the origin a t different rates and with different values. In this range the heat evolved in the presence of the diphenylguanidine is materially less than without this accelerator. I n neither case is the plot a straight line through the origin, as implied by the data of Jessup and Cummings. In previous papers ( 3 , 4 ) it was postulated that the reaction of rubber and sulfur during vulcanization occurred in two steps: Rubber reacts with sulfur to form the elastic soft vulcanized compound which in turn reacts with more sulfur to form ebonite: Rubber sulfur +soft vulcanized rubber Soft vulcanized rubber sulfur --+ebonite
+
+
~
1286
INDUSTRIAL AND ENGINEERING CHEMISTRY
These reactions are successive in any one molecule, but in a mass of rubber both are taking place a t the same time. The first reaction is speeded up tremendously by the use of accelerators while the second is essentially unaffected by them. If, in addition, it is postulated that the first reaction is essentially isothermic and that the second is primarily an exothermic reaction, the difference between the two curves (Figure 12) is easily explained. When no accelerator is present, the greater proportion of the sulfur is consumed in the second reaction a n d a corresponding amount of FIGURE 11. T E M P E R A T U R E RISE WITH VI-DINITROBESZENE heat is When the COMPOUNDS diphenylguanidine is present, more of the sulfur undergoes the first reaction and the percentage of sulfur involved in the second reaction is lowered, with a corresponding reduction in the amount of heat evolved. The vulcanization of rubber with selenium seems to undergo onlyzthe soft rubber reaction since ebonite has never been prepared with this reagent alone. The compound containing this element and a nonsulfur accelerator (p-nitrosodimethylaniline) evolves no measurable quantity of heat, which tends to confirm the above theory. In the data of Figure 10 further confirmation is found; the selenium undergoes the soft rubber reaction to some extent, t h e r e b y reserving more sulfur for the ebonite reaction than would be avail1 1 able were the selei’5uffuR nium not present.
FIGURE1;. MAXI&M T E ~ I P E R A T U ~ EThe heating O f RISE vs. PERCENTAGE SULFUR rubber with tetramethylthiuram disulfide and selenium seems t o evolve no heat, indicating also that it is possible to produce a completely vulcanized rubber without the evolution of heat. The second reaction probably does not occur to an appreciable extent in this case. Vulcanization of rubber by m-dinitrobenzene in the absence of litharge seems t o evolve comparatively little heat. Even this may be due to a reaction other than vulcanization.
Vol. 26, No. 12
If there were no heat lost from the samples by conduction, it would be possible to calculate the heat evolved in calories per gram by multiplying the temperature rise by the specific heat of the rubber (approximately 0.5). As a rough approximation we may obt a i n s u c h figures from t h e p r e s e n t LO data and, although there is h e a t l o s t and the t e m p e r a ture rise increases 40 as t h e sample is made l a r g e r , i t i s believed that a maLo terial increase bey o n d t h e present size will give but moderately l a r g e r L 4 6 8 IO IC values. Such a cal- FIGURE 13. HEATOF VULCANIZATION c u l a t i o n gives, of course, curves proportional to those in Figure 12, with a maximum value of about 35 calories per gram a t 8 per cent sulfur. These figures are plotted in Figure 13, together with the new data of Jessup and Cummings. The values are of the same order of magnitude. It
14
ACKXOWLEDGMENT The author is indebted to Philip L. Bruce for assistance in the preparation of this paper. NOTRJ: Previous articles in this series have appeared in INDnsTRIaL AND ENQINEERINQ CHEMISTRY on pages 737 to 755, July, 1930, and pages 649 t o 555, May, 1932.
LITERATURE CITED (1) Adams, International Critical Tables, Vol. I, p. 58, McGrawHill Book Co., New York, 1926. (2) Blake, IND.ENQ.CHEM.,22, 737 (1930). (3) Ibid., 22, 744 (1930). (4) Boggs and Blake, Ibid., 22,748 (1930). (5) H a d a , Fukaya, and Nakajima, J . Rubber SOC.Japan, 2, 389 (1931);Rubber Chem. Tech., 4,507 (1931). (6) H a d a and Nakajima, J . Rubber SOC.Japan, 5, 288 (1932); Rubber Chem. Tech., 6 , 56 (1932). (7) Jessup and Cummings, Bur. Standards J . Research, 13, 357 (1934). (8) Lewis et al., paper presented before Rubber Division at Chicago Meeting of American Chemical Society, Sept. 10 t o 15, 1933. (9) Perks, J. SOC.Chem. Ind., 45, 142T (1926). (10) Riding, Inst. Rubber Ind. Trans., 6 , 230 (1931). (11) Toyabe, J. SOC.Chem. Ind. Japan, 33, 96B (1930); Rubber Chem. Tech., 3, 385 (1930). (12) Toyabe, J. SOC.Chem. Ind Japan, 33, 27SB (1930); Rubber Chem. Tech., 4, 190 (1931). RECEIVED September 15, 1934. Presented before the Division of Rubber Chemistry a t the 88th Meeting of the American Chemical Society. Cleveland, Ohio, September 10 to 14, 1934.
Dow CHEMICAL COMPANY
Catalysts for Oxidation of Ammonia t o Oxides of Nitrogen S. L. HANDFORTH . ~ N DJ. N. TILLEY, Eastern Laboratory of E. I. du Pont de Nemours &Company, Inc., Gibbstown, N. J.
C
ATALYTIC oxidation of a m m o n i a is commercially one of the most important reactions in heterogeneous catalysis, and the various materials suitable as catalysts have been the subject of extensive investigations. The stoichiometric equations are as follows :
together. As the reaction must be carried out a t a temperature above 600" C. (1112°F.) and under oxidizing conditions, only the few metals similar to platin u q will withstand the conditions and remain in m e t a l l i c form. P l a t i n u m is the outstanding metal of this group, The other catalysts are composed of various oxides. These 4NH3 502 = 4N0 6Hz0 (1) have been extensively investi4NH3 302 = 2x2 6Hz0 (2) gated by many w o r k e r s and There has been much theoriznearly all oxides and combinations that might be s u i t a b l e ing on the actual steps occurring in the reactions ( 1 , 2, 16), and have been tried. Except in one it is probable that the catalytic or two cases, the efficiency of surface actually takes active part such catalysts has been so low in some steps (16). The desiras to make their u s e u n e c o able catalyst, of course, is one nomical. During the World which will have extremely high War iron oxide c a t a l y s t s , activity for reaction 1, and a low activity with respect to reac- usually promoted with bismuth oxide (S), were used extion 2. Practically all surfaces catalyze both reactions to tensively in Germany. After the war, platinum again besome degree, and almost all conceivable materials have been came available there and almost entirely displaced oxide tried and patented (13). Most of these materials, however, catalysts. With possibly one exception (18) platinum or catalyze reaction 2 to such a degree as not to be commercially platinum alloy catalysts are used exclusively in this country. In view of these facts, the early work of the du Pont Comattractive. Two general types of commercial processes using air or pany was confined largely to the platinum catalysts. Tests oxygen-enriched air as the source of oxygen are in common on the oxide type catalyst indicated that even those reported use, I n one, the operation is carried out a t essentially at- in the literature as most promising were inferior to platinum. mospheric pressure; in the other, the operation is carried out Furthermore, the first results obtained on platinum-rhodium under increased pressure varying from 50 to 100 pounds per alloys ( 7 ) were so encouraging that most of the work theresquare inch. This higher pressure is used to increase the after was confined to the field of metallic catalysts. speed and efficiency of converting the oxides of nitrogen to The platinum catalysts lost weight and disintegrated nitric acid, and is well worth while (19) on account of the rapidly under the operating conditions first tried, but it was smaller equipment required and the higher final acid strength. known that all platinum metals oxidize and volatilize more or At these high pressures, however, i t is more difficult to obtain less rapidly a t high temperature in oxidizing atmospheres high efficiency in the oxidation of ammonia to oxides of (4, 5 , 6, 17). These investigators, however, indicated that nitrogen. Consequently, higher temperatures are used in there would be little or no volatilization of platinum below pressure oxidation in order to obtain good conversion effi- 900" C. (1652" F.). On the other hand, when used as a ciency. While the capacity is increased thereby, the de- catalyst in the ammonia oxidation process, platinum is volaterioration of the catalyst is more rapid. The results ob- tilized from the catalyst rapidly even at much lower temperatained with one catalyst operated under one pressure, there- tures. The surface of the metal becomes "etched" a t first, fore, cannot be compared with the results obtained with a and then covered with sprouts which continue to increase different catalyst a t any other pressure. until the original form of the metal is lost. This effect of a Even when operating under the same pressure, the ca- reaction greatly accelerating the loss of material from a pacity, efficiency, and life of a catalyst are interrelated and catalyst has also been noted with other reactions (20). It affected by so many factors that all must be taken into was found that this loss of metal did not depend on length of account in making any comparison. In general, the higher time or amount of platinum in service. Careful tests showed the temperature the greater is the capacity of a given catalyst, the loss a t a given temperature and with a given alloy to be but unfortunately the more rapid is the deterioration. At a proportional to the weight of oxygen in the reacting gas given capacity almost every catalyst has a maximum tempera- mixture passed over the catalyst, regardless of volume. In ture for highest conversion efficiency (9,14). the operation of large commercial plants, it has been found The catalysts for ammonia oxidation may be divided into that the rate of weakening and breaking of gauze catalysts is two broad classes, as they are different in both their chemical directly proportional to the loss of metal, provided mechanical nature and mechanical arrangement. Rletallic catalysts such damage is avoided. Within the limits of gas mixtures ordias platinum or platinum alloys are usually made up in the narily encountered, this loss is for practical purposes proporform of fine wire gauzes, several layers of which are placed tional to the ammonia used. Since the amount of ammonia is 1287
+ +
+ +
I n determining the economic value of a catalyst, it is necessary to consider all factors-namely, loss of metal, conversion eficiency, and capacityunder the actual conditions of commercial operation. Thus a n alloy which shows extremely high conversion eficiency m a y be entirely aneconomical because of the high loss of metal which is incurred under the conditions necessary to obtain high conversion eficiency. Platinum-rhodium alloys, howecer, hate been found to give a low loss of metal and high capacity under the operating conditions required f o r the maintenance of high conversion eficiency. Consequently, the pure platinum-rhodium alloys containing 5 to 10 per cent rhodium are the most advantageous and economictrl of a n y thus f a r proposed.
