(533e) Scaling a Micro Structured Reactor for Sugar Chemistry from Lab Via Pilot to Full Production Scale

Scaling a
micro structured reactor for sugar chemistry from lab via pilot to full
production scale

M. Kraut¹⁾,  G. Rabsch¹⁾
and R. Dittmeyer¹⁾

1) Karlsruhe Institute of Technology, Institute of Micro
Process Engineering, Hermann von Helmholtz Platz 1 76344
Eggenstein-Leopoldshafen, Germany

Motivation

It is well known, that micro structured devices show very
high heat transfer rates, and some processes profit immensely by fast heating
and cooling processes. Examples of production processes performed in micro
structured reactors are still quiet rare. Following Roberge’s work [1] it is
surprising that not more processes have been transferred from macro scaled
devices to micro or milli-scaled ones.

Especially processes in the food industry are often considered
not to be  compatible with micro structured flow devices due to the properties
of natural products (varying quality, unknown side products, fouling tendency…)
The challenge is taken up to make a new product using micro structured heat
exchangers to obtain high specific heating rates to achieve defined residence
times to perform a sugar transformation.

Process

The transformation of sugars at elevated temperatures is
well known for a long time (e.g. caramel making). The target product is
narrowly defined in terms of selectivity especially of undesired side products.
This quality is achieved by a very well defined residence time of the reaction
mixture at an equally well defined temperature. The reaction mixture firstly is
brought to a temperature which is safely below the onset of both, desired and
undesired reactions. Then the mixture is heated up rapidly (heating rates
>100 K/s) to the optimum reaction temperature and held there for a defined
time to obtain equilibrium yield of the desired and a minimum yield of
undesired side product. After water is separated from the reaction mixture as
steam the procedure is repeated with a second set of temperature and residence
time.

Laboratory
reactor

In Lab-scale the task was to find the parameters for
operating the process reproducibly and safely within the required quality
criteria. As a tool for preheating a symmetrical cubic heat exchanger was
applied, which was described previously [2]. To perform the reaction and
provide the necessary residence time a micro structured so called laboratory
module[3] was used (fig 1, bottom).

This module provides 30 parallel reaction channels and for
an intense heating 179 channels perpendicular to the reaction channels. The
structured foil has been made from stainless steel (1.4571 or 316 Ti) and
structured mechanically. The reactor is operated with a very high volume flow
of thermos oil to obtain the boundary condition of a “hot wall”. Since the
temperature difference in the oil is negligible a heat balance can’t be done,
but the point is to provide the reaction with optimum conditions, not the most
efficient heat usage.

Since the reaction itself has a very small enthalpy the
heating effort is small, and one could argue that the reactor is over-designed.
However, even small variation in temperature can lead to undesired effects, so
that the condition to keep the temperature constant is sufficient to design a
completely temperated reactor.

The process parameters for the reaction were found to be in
the range of a few dozen seconds in residence time and a reaction temperature
of about 200°C.

Fig.1 KIT lab module (bottom) and Pilot reactor derived from
lab module  (top)

Pilot scale reactors

The lab results were then transferred to a pilot scale. The
scaling principle was to keep the structure of the device and multiply it
according to the scaling factor, in this instance a factor of 15 (from 1 to 15
kg/h). Since the parameters of operation were known, the pilot device was
manufactured in several layers of plates for heating and cooling respectively.
The thickness of each foil was 1 mm, so the foils were diffusion welded to form
a monolithic block (fig.2). The length of the heating channels was slightly
increased to ensure diffusion welding compatibility of the device. However, geometry
for the inflow  has been altered, so that the specific pressure drop was not
increased in comparison to the laboratory module. Also a 16^th layer
of heating channels was added to obtain the same conditions for each of the
reaction channels (heated walls in both directions).

The reaction conditions were successfully reproduced at
designed thoughput using this device. No detrimental scaling effects were
observed. However, the flow distribution in the flanges was optimized during
the pilot phase to ensure a better distribution based on cfd-calculations. [4]

 To test the possibility of etching as an alternative
manufacturing technique, which for higher amounts of foils is economically
advantageous, one device was manufactured using etched foils. Dimensions were
slightly altered to obtain circular channels by face-to face stacking. That
device was also tested successfully.

Fig.2 Diffusion welded block of pilot reactor. 16 layers of
cooling foils are visible. Three cooling zones are evident.

Production
reactor

Based on these encouraging results the design of a
production unit with a further scaling factor of 60 was done. Since a reactor
with a factor of 60 in height which would be the consequent scaling option was
found to be not feasible; several conditions were altered:

·       
Crossflow heat exchange was replaced by countercurrent heat
exchange

·       
The number of parallel reaction channels was increased 

·       
The heating fluid is divided in two streams, each of which is
flowing through a half of the reactor.

Design calculations ensure the condition of “hot wall” or
constant reaction temperature along the whole channel. 

Fig.3: Design of production reactor as scaled up version of
the pilot device, reaction direction ion green, heating in red.

Conclusions

A process of sugar transformation has been investigated
using micro structured reactors and found to be feasible when using well
defined residence time at equally well defined temperature, exploiting the high
heating rates and flow conditions within micro structured channels.

A scale-up from laboratory device to a diffusion welded
pilot device has been done successfully, showing that the scale-up strategy of
using identical local conditions is feasible.

Further scale-up to obtain a micro structured reaction
system has been performed with further economic improvements.

[1] Roberge, D.M., Ducry, L., Bieler, N., Cretton, P.,
Zimmermann, B., Chemical Engineering and Technology 28(2005), pp. 318-323

[2] Schubert, K., Brandner, J., Fichtner, M., (...),
Schygulla, U., Wenka, A. , Microscale Thermophysical

Engineering5(2001), pp. 17-39

[3] K. Haas-Santo, M. Kraut, W. Benzinger, A.
Wenka, L. Bohn, R. Dittmeyer, Chem. Eng. Tech., 82 (2010), p. 1341

[4] J. Cao, M. Kraut, R. Dittmeyer, L. Zhang,
H. Xu, Int. Commun. Heat Mass 93 (2018), 60-65