average residence time. It is a very special know-how to design the catalyst and the
activation process with the cocatalyst system in this way.
Using the polymeric flow model [27, 28] and assuming constant temperature
of the growing particle at the lower stable reaction point, it can be shown that
the polymerization process always runs in the kinetically controlled range, with an
overall rate constant for the ethene polymerization in the range of 100 dm
3 mol
À1 s
À1
and an effective diffusion constant for ethene inside the microreactor of about
1Â10
À4 cm
2 s
À1 at the preferred temperature of about 80
C. This means that the
growing particle (the microreactor) is at a temperature range near to the temperature
of the diluent (T 0 ), with only small temperature and concentration gradients at
the surface and inside these particles [32, 33].
In conclusion, super-active catalysts used in technical processes have a very
high activity but do not overheat at the start-up of polymerization and remain
at this high activity level over a long period of time in the range of the average
residence time of around 1–2.5 h depending on process design (see Sect. 4). Only
with those catalysts can the modern process be operated effectively.
3.2 The Mesoscale Level
As shown in Fig. 4, the mesoscale level deals with all processes in the threecomponent slurry phase. On this level, mass and heat transfer processes play key
roles. To guarantee a stable stationary state throughout the whole reactor and
to establish stable concentration and temperature gradients from top to bottom
and vice versa, the reactor must be equipped with an optimized stirring and
cooling system. This is a challenge for large volume tall-thin reactors because the
height-to-diameter ratio is much higher than 1. However, such stirring systems
are now available [34–36]. The design is characterized by several blades, impellers,
or turbines at different levels along the stirrer axis. In that way it is possible to
generate a virtual draft tube with sufficient mixing throughout the tall-thin reactor,
as described elsewhere [36]. Figure 8 shows the design of such a stirring system,
together with scale-up criteria applied for these tall-thin reactors.
With this stirrer design, it is possible to have good local and bulk mixing.
It can be estimated that using such optimized stirring systems, a bulk mixing
time can be realized within a few minutes, which is lower by at least one order of
magnitude than the average residence time of 1–2.5 h. This is sufficient because
each polymer particle passes through the whole reactor many times before leaving
the polymerization vessel. Together with effective outercooler loops, it is possible
to have only small temperature gradients inside the whole slurry phase. Further, it is
essential to generate a polymer powder with a high bulk density in the range of
400 g dm
À3 to run a high polymer content in the slurry of up to about 25% by
weight. It is important that the polymer particles have a ball-like structure to reduce
friction between the particles. As shown elsewhere, the viscosity of the slurry phase
can be limited in a range that is not more than ten times higher than the viscosity of
the diluent [32, 33]. This does not seriously influence local and bulk mixing.
68
L.L. Bo ¨hm
Précédent

- 75/261

Suivant