Catalyst Deactivation - Spent Hydroprocessing Catalysts - page 58

on morphology of the coke deposited on the catalyst surface. They concluded that most of the
coke was located far from the active phase in the form of the irregularly shaped structures
covering the entire surface.
Some information [179] showed that for the model feeds such as carbazole and alkylated
carbazoles, as well as polyaromatic hydrocarbons (PAHs), the initial coke deposition (less than
5 wt.%) occurred predominantly on the Al 2 O 3 support. Consequently, little deactivation was
observed during this initial period. However, when a VGO was used under similar conditions,
the catalyst poisoning by N-compounds and PAH present in the feed was observed few
minutes after the start of the run [180] . Similarly, the initial poisoning by N-compounds was
also observed during the hydroprocessing of the Maya heavy crude, although to a lesser extent
[181] . While using Kuwait atmospheric residue (AR), Matsushita et al. [182] concluded that
the coke formed during the early stages of the operation deposited on the support rather than
on active phase. At the same time, N-compounds in the feed adsorbed near and/or on the active
sulfides phase and as such contributed to the initial catalyst deactivation. Therefore, the
catalyst deactivating patterns observed for model compounds [179] were rather different than
those observed for the real feeds [180-182] .
Figure 4.5 shows the decrease in the H/C ratio of coke with time on stream [183] . At the same
time, the N content of coke increased linearly. This suggests that during hydroprocessing of the
diluted vacuum residue derived from the Chinese heavy crude, the accumulation of carbon and
nitrogen in the coke deposited on catalyst increased, whereas that of hydrogen decreased with
the increasing time on stream. An ultimate result of this change was the increase in the content
of graphite-like structure of the coke on catalyst. The experiments were performed at 683 K
and 8MPa in an autoclave. During these experiments, no attempt was made to decouple the
contribution of metals to deactivation from that by the coke and N-compounds. The duration
of these experiments (e.g., maximum 5 h) suggests that these observations reflect the events
Figure 4.5: Effect of time on stream on H/C ratio and nitrogen content of coke on catalyst
[From ref. 183 . Reprinted with permission].