Reactor Vycor

Gas-phase experiments at 155 °C. were carried out in a 250-ml. cylindrical Vycor reactor in a hot-air furnace. Later experiments were done in a 500-ml. borosilicate glass flask heated in an oil bath. With either system, di-terf-butyl peroxide, oxygen, and isobutane were metered into the reaction vessel in that order by expansion from the vacuum line the pressure of each component was measured using a mercury or oil manometer. Mercury vapor was excluded from the reaction vessel. 

The coke formed in the Vycor reactor had been produced in runs (13) conducted about one year prior to inspection. The detailed history of the reactor is reported in Table III. The inlet half of the reactor was clear, showing no signs of coke deposition. The outlet half was, however, covered with coke to make the reactor black and almost opaque. The exit end of the reactor outside the heated sand bath was covered with a... 

A second, all quartz 50 mm reactor was immediately fabricated at Princeton and flown to Odeillo to serve as backup for the 25 mm vycor reactor. 

A second test (Experiment 1.08) with cellulose in steam flowing at 3 g/min was perfomed, and no clouding of the reactor s window was observed. Data for the gasification of 0.48 g of cellulose is presented in Table II. A third test (Experiment 2.08) using corn cob material was also performed, and the data obtained is listed in Table III. Following the third experiment a small crack developed where the heater tape had fused into the wall, and work with 25 mm OD vycor reactor terminated. 

A solution of 24 g (82 mmol) of l.l.lj in 2.4 L of methylene chloride was irradiated with a low-pressure mercury lamp (Vycor-reactor, Grantzel) for 12 h. After evaporation of the solvent the products were separated by column chromatography (550 g silica gel) with pentane as eluant. After collecting a fraction of 5 g (25%) of 2,3,5-Tri-rert-butyl-2,4-... 

A solution of 20 mg (0.07 mmol) of 1.1.2a in 1 mL of 2,2-dimethylbutane/n-pentane (8 3, Rigisolve, Merck) was irradiated at -196 °C in a quartz tube with a low-pressure mercury lamp (Vycor-reactor, Grantzel quartz dewar flask, charged with liquid nitrogen). After evaporation of the solvent 12 mg (66%) of 1.1.2b could be obtained. 1.1.2b could be purified by sublimation at 0.01 Torr and 35 °C. 

The coupons after removal from the Vycor reactor were cooled in an inert atmosphere and then weighed. Coupon surfaces were analyzed using a Jeolco JSM-U3 scanning electron microscope and photographs... 

The pretreatment of the metal reactors effected the product composition. For example, as shown in Figure 1, one of the runs in the Incoloy reactor having a reduced inner surface had at a given ethane conversion a product composition similar to that for the Vycor reactor. Yet a second run in which the Incoloy reactor had a more oxidized surface resulted in a product containing less ethylene and more hydrogen the product for the second run was relatively similar to that for the stainless steel reactor. 

Figure 6 and 7 indicate typical ethylene yields as a function of ethane conversion for runs at 800°C with and without steam respectively. These yields (based on the moles of ethane that reacted) decreased with increased ethane conversion and in the following order Vycor, Incoloy, and 304 stainless steel. Higher yields were also noted in the metal reactors with dry ethane as compared to wet ethane runs the opposite was the case however, for the runs in the Vycor reactor. Corrected results indicate, however, higher ethylene yields with wet ethane feeds regardless of the reactor used. The corrected results agree well with the mechanistic model also shown in Figures 3,6, and 7 this model will be discussed later in the paper. 

Surface reactions including carbon (or coke) deposition (on the reactor surface) varied significantly in the reactors investigated. Table I and Figure 8 show carbon results for several runs in the 304 stainless steel, Incoloy, and Vycor reactors. 

Recent results of Dunkleman and Albright (1) who pyrolyzed ethane have shown that the composition of the product often varies significantly depending on the material of construction of the reactor and on the type of pretreatment of the inner surface of the reactor. Considerably higher yields of ethylene were obtained in a laboratory Vycor reactor as compared to an Incoloy 800 reactor and especially a 304 stainless steel (SS) reactor. Oxidized inner metal surfaces promote the production of coke (or carbon) and carbon oxides, but sulfided surfaces suppress such production. 

Figure 1. Product composition for propane pyrolyses at 800°C bath temperature and using wet ethane feed in Vycor reactor...

Surface reactions are thought to be relatively unimportant when propane is pyrolyzed in a Vycor glass reactor. This conclusion is based on two factors. First, surface reactions in the Vycor reactor were of fairly minor importance for ethane pyrolysis (1) and are considered to be even less significant for propane pyrolysis. Secondly, as was shown in both table I and Figure 1 and as will be described later in this paper, there was good agreement between the experimental results for the Vycor reactor and the predictions of the mechanistic model described earlier (1). 

The pyrolysis results obtained using the Vycor reactor were used to make various comparisons. Increased temperatures and decreased partial pressures of entering propane both result in increased ethylene yields. Such findings are consistent with the general trends reported by many previous investigators for pyrolyses of other hydrocarbons and specifically by those (3,4) who have investigated propane pyrolyses in metal reactors. The ore-treatments of metal surfaces also affected product composition. Less surface reactions occur on H S-treated metal surface as compared to untreated surfaces. Oxy n-treated surfaces that have metal oxides on the surface tend to promote undesired surface reactions and to produce considerably more carbon oxides. 

Several mechanistic models involving numerous free-radical gas-phase reactions were tested and compared to prooane pyrolysis results obtained in the Vycor glass reactor. Data obtained in metal reactors were not compared since surface reactions were obviously of much greater importance in these latter reactors. The model developed as a result of this preliminary testing is presented in Table II of the previous chapter of this book (1). Good agreement occurred between the predicted values and all experimental results obtained in the Vycor reactor (at both 750° and 800°C with and without steam as a diluent). Figures 1 and 2 show... 

Figure 2. Comparison of kinetic (or mechanistic) model and experimental data for propane pyrolyses in Vycor reactor...

The results obtained in the Vycor reactor are of special interest since surface reactions are of relatively minor importance in this reactor. The results of this reactor then are most useful in clarifying the mechanism of the gas-phase reactions. 

The mathematical model used to correlate the propane pyrolysis data obtained in the Vycor reactor can likely be used for at least moderate extrapolations to temperatures and pressures beyond those investigated. 

Hydrocarbon conversions can. In general, be represented fairly well by first-order reaction kinetics, and the conversion levels for runs made with a constant hydrocarbon flow rate as a general rule Increased significantly as the temperature Increased. Figures 1, 2, and 3 show typical results for ethane, propane, ethylene, and propylene. Based on first-order kinetics, the activation energies for ethane, propane, ethylene, and propylene were determined In the various reactors tested. In the Vycor reactor, these activation energies were approximately 51, 57, 56, and 66 k cal/g mole respectively. They were much lower In metal reactors especially after the reactor was oxidized. In a relatively new and unoxidized Incoloy reactor, the activation energies were 15, 47, 27, and 26 k cal/g mole respectively. 

With propane as the feed hydrocarbon, relatively little difference was noted in the conversions in the different reactors (see Figure 1). Probably gas-phase reactions begin with propane at much lower temperatures than for the other three hydrocarbons propane is the only hydrocarbon of those tested that has secondary carbon-hydrogen bonds. For propane, the metallic reactors did result in slightly higher coke and hydrogen yields than the Vycor reactor, but the conversions were similar. 

Materials of Construction. The materials of construction of the reactor clearly affect the level and probably to some extent the type of surface reactions. Surface reactions are in general much less important in Vycor reactors at least in the range of approximately 450° to 550°C (see Figures 1, 2, and 3), as indicated by the lower hydrocarbon conversions in these reactors. 

The 410 stainless reactor showed much lower activity than either of the other two metal reactors. This reactor had activities only slightly greater than those of Vycor at temperatures up to about 650°C. Yet the Vycor reactor had higher hydrocarbon conversions than the 410 stainless steel reactor at higher temperatures (between 650°C and 750°C). The 410 stainless steel reactor probably terminated more free radical reactions than did the Vycor surface. 

Miscellaneous Results. After the Vycor reactor was contacted with oxygen for several hours at 800°C, the activity of the reactor was Increased slightly. Both ethylene and propylene showed detectable reactions for almost 15 minutes at 500°C which Is 50°C lower than for the untreated reactor. Small amounts of carbon oxides were detected during this time Indicating that some oxygen had been adsorbed on the Inner Vycor wall. After all of the oxygen had desorbed, the activity of the reactor returned to a level similar to that In the untreated Vycor reactor. A steam treatment of the Vycor reactor at high temperatures did not, however, produce any noticeable Increase In the reactor activity. 

Surface deposits were noted In the Vycor reactor following the experimental runs. These dark brown or black deposits occurred primarily at the exit end of the reactor. These deposits were oily and tarry In nature, and were probably coke and condensed heavy hydrocarbons. 

Hydrogen treatment at 800°C of the Vycor reactor, which contained some coke or tarry material, resulted In the production of some methane. This result was surprising since It had been thought that methane formation would occur only In the presence of a metal catalyst such as nickel or Iron. 

The present results clearly confirm the Importance and complexity of surface reactions during pyrolysis reactions. Obviously, the composition of the Inner surface of the reactor Is of Importance relative to the level and types of surface reactions. In addition, valuable new Information has been obtained concerning the role of coke In affecting more coke formation. Although the deposition of coke on the walls of a metal reactor decreases the activity of the reactor. It Is of Interest that the surface activities of coke-covered metal reactors always remained higher than those for the Vycor reactor. Lobo and Trimm (11) have Indicated that carbon without contaminants Is Inactive. Based on this finding, metal contaminants were presumably present In the coke formed. Other Investigators (10, 11) have found both nickel and Iron contamination of various cokes. Furthermore, coke Is sometimes reported to be autocatalytic In nature. The evidence that olefins and other hydrocarbons adsorbed on the surface and... 

---