(64f) Development and Experimental Characterization of a Pilot-Scale Falling Film Microreactor and Outlook on Production Scale

The standard
Falling Film Microreactor (FFMR-STANDARD) developed by IMM in which liquid
films of a few tens of micrometer thickness and interfacial areas up to 20,000
m²/m³ combined with an effective heat exchange can be
obtained represents a broadly used microstructured lab tool in the area of
gas/liquid contacting (see e.g. [1-3]). In the context of the German public
funded project µ.Pro.Chem (Förderkennzeichen 16SV1991) with the partners BASF
AG, Evonik Degussa GmbH and IMM one topic is to transfer this lab-scale device
which is successfully applied for the Evonik Degussa GmbH process of an ozonolysis
towards pilot scale and production scale. Targeted pilot throughput is thereby
for the liquid about of 10 l/h which represents a one-hundred fold increase compared
to the lab-scale device FFMR-STANDARD. To reach production scale a further
increase of throughput by a factor of 10 ? 100 is required.

The followed
approach by IMM to reach pilot scale is first to design a larger base unit with
an about tenfold increase of structured surface and then in a next step to
design a pilot reactor based on (tenfold) internal numbering-up. Therefore in a
first step a falling film microreactor with a larger reactor plate has been
developed and realized. The microchannel width and depth has been kept constant
compared to FFMR-STANDARD while both the number of channels and the length of
the channels have been increased by a factor of 10^0.5. The principal
design in this new falling film microreactor (FFMR-LARGE) ? removable reaction
plate in a housing bearing the feeding lines ? is similar to the one of the FFMR-STANDARD.
Some modifications have been incorporated to improve or ensure equal
distribution of gas and liquid in the reactor. The next step then was to come
to a pure plate design for this larger base unit as basis for the later
internal numbering-up. While the reaction plate structure for the gas/liquid
contacting was kept, major changes have been necessary to realize the
functionality of the housing with regard to stability and feed distribution by
only structuring plane reactor plates by wet-chemical etching. Furthermore, the
reactor plates design has been adjusted to the selected sealing process by
brazing. Corresponding prototypes (called STACK-1x-FFMR-LARGE) have been successfully
realized including a version with an open gas chamber which allows visual
observation of the reaction plate. Finally, a pilot-scale reactor consisting of
a stack of ten of these plate-based functional units has been realized (called
STACK-10x-FFMR-LARGE).

The reactor
development work has been accompanied and supplemented by experimental works
dealing with the characterization of the reactor performance and also with
studying equal distribution of gas and liquid in the reactor for all developed
reactor types. The ozonolysis reaction has been performed by Evonik Degussa
GmbH itself and can not be reported here. The reactor performance has
additionally been studied by IMM using the CO₂ absorption in an
aqueous sodium hydroxide solution and by an oxidation reaction of an organic
compound by pure oxygen. These investigations have confirmed that the
performance of the larger base unit FFMR-LARGE can be operated at a tenfold
throughput compared to FFMR-STANDARD while keeping similar reactor performance
(see also [4]). Via the CO₂ experimentation it was shown also that
the step from FFMR-LARGE to STACK-1x-FFMR-LARGE has not changed reactor
performance and thereby confirmed the design. Furthermore, just recently the
experimentation with the pilot-scale reactor STACK-10x-FFMR-LARGE has started.
So far already a reactor performance of about 70% at an again tenfold increased
throughput has been reached. The works here are continuing to confirm 100%
performance. Gas and liquid equal distribution in different falling film microreactors
has been studied by different methods: evolution of the liquid film via dosing of
a dyed liquid, liquid flow rate measurements per channel and discoloring of a with
phenolphthalein dyed sodium hydroxide solution during CO₂ absorption.
Thereby, the evaluation of the different reactor designs has been possible.

Concluding an
outlook on the approach to reach production scale by a combination of internal
and external numbering-up will be given.

[1]        Ehrich, K., Linke, D., Morgenschweis, K., Baerns, M., Jähnisch,
K., Chimia 56 (2002) 647-653.

[2]        Yeong, K.K., Gavriilidis, A., Zapf, R., Kost, H.-J., Hessel,
V., Boyde, A., Exerpimental Thermal and Fluid Science, 30 (2006) 463-472.

[3]        Löb, P., Löwe, H., Hessel, V., Journal of Fluorine
Chemistry 125 (2004) 1677-1694.

[4]        Vankayala, B.K., Loeb, P., Hessel, V., Menges,
G., Hofmann, C., Metzke, D., Krtschil, U., Kost, H.-J., International Journal
of Chemical Reactor Engineering, accepted for publication.