(707a) Process Intensification of the Alkoxylation of Fatty Amines

Process
Intensification of the Alkoxylation of Fatty Amines

Pia Müller^1*, Wyatt Winkenwerder²,
John van der Schaaf¹

1 Department of Chemical Engineering
and Chemistry, Eindhoven University of Technology, P.O. Box 513,  5600MB, the
Netherlands; 2 Nouryon, Brewster, New York 10509, United States

*Corresponding author: p.muller@tue.nl

j.vanderschaaf@tue.nl

wyatt.winkenwerder@nouryon.com

1. Introduction

Alkoxylated fatty amines are widely used
as non-ionic surfactants, leading to a steady increase in industrial demand
over the past six decades. These compounds are key components in products of numerous
industrial branches, e.g. oil drilling, paper manufacturing, textile industry,
as well as fiber manufacturing, paints, dyes and plastics [1]. These
compounds are typically produced by reacting fatty amines with ethylene oxide
(EO) or propylene oxide (PO) in large fed-batch reactors, often reaching the
size of 30 m³. The process is intentionally de-intensified, due to its
strongly exothermic nature (Δ_RH = -120 kJ/mole of EO) and the safety risks associated with EO.
Therefore, the epoxide is slowly added to the fatty amine to keep the
exothermic reaction under control. In the past, attempts have been made to
intensify the reaction by using venture-loop or spray-tower reactors [2], [3]. However,
due to the required active mixing, these processes necessitate a power input higher
than the typical set-up, and fail to provide significantly improved product
output. The aim of our work is to transfer this industrially important reaction
to using modern technology to a 21^st century process. One possible
option to be is shown in Figure 1.

Figure
1: Schematic of
possible, intensified reaction setup. In total two mole EO (or other epoxides)
react with the fatty amines to form a di-alkoxylated amine. The process includes
a CSTR in series with a plug flow reactor.

This study focuses on gaining detailed
kinetic insights and their application to the development of continuous flow
systems to improve the reaction efficiency whilst maintaining safety standards.

2. Methods

To achieve our aim a kinetic study was
necessary. In a first step we used model compounds such as butylene oxide (BO)
or propylene oxide (PO) with octylamine (OA) in a lab scale reactor. A
temperature window of 120 °C to 160 °C was explored by analyzing the mixture
with GC, NMR and inline FTIR. The process imitated the industrial process by
adding the epoxides in a lower molar ratio stepwise to the amine.

In a second steps the collected data were then
applied to the design of a flow set-up utilizing microchannels as perfect plug-flow
reactors. A comparable flow system was also employed to determine the kinetics
of ethylene oxide (EO) with OA over a temperature range between 120 and 170 °C.
For safety reasons and accurate contact times it was important to work in a
fully liquid system. The system was therefore pressurized by a back pressure
regulator to 12 bar (PO and BO) or 60 bar (EO). Such pressure for PO was chosen
after a large negative deviation from ideal mixing (8 bar measured at 150 °C to
22 bar in mixture) behavior was observed and a NRTL model was designed. For EO
a high tolerance of the theoretical vapor pressure was chosen.

Finally, the reactions with PO and OA was transferred
to different set-ups of potential industrial interest. Options like spinning disc
reactors or other CSTR types were tested, as well as the recycling of an
intermediate mixture 

3. Results and discussion

For the reaction of BO and PO with octylamine experimental data
were collected in semi-batch and flow:

Figure
2: Experimental data
in fed-batch (left) was collected over four addition steps of PO of each 0.5
mole per mole OA. In flow (right) the molar ratio initially was 2 mole PO per
mole. Both experiments were performed at 130 °C. An auto-catalytic model (upper
right) has been used to simulate both experiments.

The reaction is initially very slow in
both batch and flow. The reaction accelerates after a certain amount of product
is formed, which is typical for auto-catalytic reactions. The advantages of going
from fed-batch to flow are clearly visible in Figure 2. What is impossible to
achieve in batch, i.e. high conversion in only one step, is feasible in flow.
Through the use of microchannels the volume could be decreased from 100 to 7
ml. Thanks to the intensified mixing the reaction time for achieving full
conversion could also be decreased by a factor of 1.5-3.0 depending on the
epoxide and conditions. Furthermore, the continuous set-up is superior due to
the higher saturation of the amine with epoxides achieved under pressure. The
overall reaction for PO was 1.5 times as fast as for BO. It is expected for EO
an increase in rate by at least a factor 2 compared to PO. The use of a CSTR
will lead to an additional decrease of reaction time, due to a better back
mixing, which would enhance the inherent auto-catalytic effect of the reaction.

4. Conclusions

Process intensification for the
alkoxylation of fatty amines was achieved with a test flow system. A rate
increase by a factor 1.5-3.0 could be reached for the reaction of propylene and
butylene oxides with the model substrate octylamine. Using a continuous
microchannel flow system, the intrinsic safety risks associated with the
process could be overcome. A kinetic model that fits the experimental data for
the reactions of BO, PO, and EO with amines was found. Using this model,
different process configuration options are easily compared for performance.

References

[1]   T.
M. Schmitt, Analysis of Surfactants, Second Edition. CRC Press, 2001.

[2]   L. L. van Dierendonck, J. Zahradnik, and V. Linek,
“Loop Venturi Reactor A Feasible Alternative to Stirred Tank Reactors?,” Ind.
Eng. Chem. Res., vol. 37, no. 3, pp. 734–738, 1998.

[3]   M. Di Serio, R. Tesser, and E. Santacesaria,
“Comparison of Different Reactor Types Used in the Manufacture of Ethoxylated,
Propoxylated Products,” Ind. Eng. Chem. Res., vol. 44, no. 25, pp.
9482–9489, Dezember 2005.