Catalyst Deactivation - Spent Hydroprocessing Catalysts - page 84

is believed that the chemical composition and structure of deposits at this level of metal
retention is rather complex and will change vertically from the outer surface to the bottom of
the deposit layer, which is in contact with catalyst surface. In this regard, properties of catalyst,
its porosity in particular, may play an important role.
4.7 Modeling of Deactivation
Models can be used to generate the database for selecting catalysts to match the feed and a
suitable reactor also, to predict long term performance of the system. Such a database can be
generated quite readily. Models incorporate changes in interface, intraphase and
interparticulate gradients of temperature and concentration with time on stream. The
complexity of chemical structure of heavy feeds compared with light feeds suggests that the
development of models to simulate hydroprocessing of the former is much more challenging.
It requires the determination of performance parameters such as the change in catalyst activity
for hydroprocessing reactions with time on stream, the parameters accounting for catalyst
deactivation, metal storage capacity of catalyst, pore size distribution, etc. The experimental
techniques for determining most of the required parameters are now available. Specific
parameters, which can be determined by kinetic studies include the intrinsic and apparent rate
constants, activation energies, effective diffusivities, efficiency factor, distribution parameter,
Thiele modulus, etc. With such parameters available, modeling can be conducted on two levels
of scale, i.e., catalyst particle level and active phase level. Modeling on a reactor level requires
the information on liquid holdup, height and diameter of reactor, volume of reactor and
catalyst bed, superficial liquid velocity, etc. It is noted that most of the studies were conducted
on more than one level of scale.
The studies involving a wide range of feeds have shown that models are feed- and catalyst
structure-dependent [10,49] . Therefore, the model developed for a particular feed may require
some modifications in order to predict the catalyst performance using a different feed. Models
take into account initial and steady-state deactivation by coke, which deposited on catalyst
surface, as well as a more less a linear and continuous deposition of metals from heavy feeds.
Attempts have been made to simulate deactivation by coke deposition occurring during very
early stages of the contact of feed with catalyst. The validity of models can be verified using
the data from the experiments on accelerating aging carried out in bench scale units, pilot
plants and from commercial reactors. The conflicting results obtained during the accelerating
aging experiments and those obtained in the commercial units have been noted [153] . Then,
with respect to model development, the former results may have a limited validity. However,
the accelerating aging test developed by Alvarez et al. [46] could predict the catalyst
performance in a pilot plant unit quite accurately.
Kinetic parameters determined experimentally form a basis for catalyst evaluation on the
active phase level. Generally, experimental data are compared with kinetic laws in either