Filamentous coke

As the metal particle size decreases the filament diameter should also decrease. It has been shown that the surface energy of thirmer filaments is larger and hence the filaments are less stable (11,17-18). Also the proportion of the Ni(l 11) planes, which readily cause carbon formation, is lower in smaller Ni particles (19). Therefore, even though the reasons are diverse, in practice the carbon filament formation ceases with catalysts containing smaller Ni particles. Consequently, well dispersed Ni catalysts prepared by deposition precipitation of Ni (average metal particle size below 2-3 nm) were stable for 50 hours on stream and exhibited no filamentous coke [16]. 

Mechanism 1 results in the formation of filamentous coke which has the following characteristics is strongly... 

Mechanism 3 has, as a first step, reactions between surface radicals on the coke and the acetylene, butadiene, and gaseous free radicals reactions probably also occur with ethylene and propylene. Reactions with gaseous free radicals were discussed earlier as a termination step in the gas-phase reactions. When acetylene reacts with the surface radicals, aromatic structures are formed on the surface. When the C—H bonds on the surface later break, graphitic coke is formed. The cokes produced by both Mechanisms 1 and 3 tends to be highly graphitic. Microscopic photographs have shown that Mechanism 3 thickens filamentous coke and causes spherical coke particles formed by Mechanism 2 to grow in diameter. 

When coils produced of high-alloy steels are used, the initial layer of coke contains an appreciable fraction of filamentous coke, as indicated by microscopic... 

During latter stages of coking, the coke formed is generally solid and contains mainly cokes formed from Mechanisms 2 and 3. It contains less metal, perhaps in the 0.3-0.5% range nickel, chromium, and iron are still present. On occasion, a metal fragment can be detected by the microscope with filamentous coke connected to it, i.e., a porous section of coke surrounded by solid coke. 

Numerous laboratory and plant tests have been made in coils constructed of materials with nickel- or iron-free surfaces e.g., quartz glass, silicon-coated and aluminum-coated steels, or ceramics. Filamentous coke was often completely absent, and the overall levels of coking were much reduced. Furthermore, the coking rates were essentially identical during the entire pyrolysis run. 

Butadiene at 560°C on Incoloy 800 surface both knobby and filament cokes. 

In the steam cracking of hydrocarbons, a small portion of the hydrocarbon feed gases decomposes to produce coke that accumulates on the interior walls of the coils in the radiant zone and on the inner surfaces of the transferline exchanger (TLX). Albright et identified three mechanisms for coke formation. Mechanism 1 involves metal-catalyzed reactions in which metal carbides are intermediate compounds and for which iron and nickel are catalysts. The resulting filamentous coke often contains iron or nickel positioned primarily at the tips of the filaments. This filamenteous coke acts as excellent collection sites for coke formed by mechanisms 2 and 3. Mechanism 2 results in the formation of tar droplets in the gas phase, often from aromatics. These aromatics are often produced by trimerization and other reactions involving acetylene. Some, but not all, of these droplets collect... 

Figure 15. TEM of filamentous coke deposit (particles arrowed).

A second finding is that the morphology of the coke deposited on aluminized Incoloy 800 and Incoloy 800 surfaces often was quite different. As depicted in Figures 2B, 2D, 2F, 4D, 4F, 5B> 6B, and 8B> the predominant structure of the coke observed on aluminized surfaces tended to be either a film of tar or a globular coke deposit. Filamentous coke was found on numerous Incoloy 800 samples (Figures 2A, 2C> 2E, 3B> 5C-5F, 7A, and 70 whereas filamentous coke was detected at most in only small amounts on the aluminized surfaces of only three runs (Figures 2F, 7B, and 7D). Third, EDAX analyses indicated appreciable metal, generally mainly iron and nickel, in the coke formed on (or brushed off) the Incoloy 800 surfaces analyses of coke deposited on aluminized Incoloy 800 coupons in the same runs indicated trace amounts of aluminum but no detectable iron, nickel, or chromium. 

Surface heterogeneities described earlier often were important relative to coke formed or deposited on aluminized Incoloy 800 surfaces. Figures SB, 6B, and 10B show globular and cylindrical coke which resulted preferentially in the pitted areas of aluminized surfaces in several runs. Filamentous coke formed at 700 C on an aluminized Incoloy 800 coupon subjected to a 0.05 atm. acetylene feed is depicted in Figure 7D every filament observed was in or near a pitted area but interestingly not every pitted area contained filaments. 

Figure 7 shows the coke formed on the various unpolished Incoloy 800 coupons. In all cases mixtures of filament and globular coke resulted on the Incoloy 800 surfaces certain areas exhibited relatively large clusters of globular coke from which filament coke protruded. 

For polished Incoloy 800 coupons (see Figure 8)< a mixture of filament and globular coke formed on the oxygen-treated surfaces. Filament coke was however not detected on either untreated or steam treated coupons. The coupon treated with steam at 00°C was covered rather uniformly with globular coke. The coupon treated with steam at 900°C exhibited a rippled layer of coalesced globular coke. 

In what is considered a key finding of this investigation/ significant filamentous coke also formed on the surface of the Vycor tubes holding the sulfided coupons. This coke was found to contain both iron and nickel which had obviously been transferred from the Incoloy 800 coupons. Possibly the iron and nickel sulfides originally present on the coupons were liquefied or even volatilized and then transported to the Vycor surfaces where they catalyzed the growth of coke filaments. 

---