Bubble columns are among the most important types of reactors in the chemical industry. Yet, despite their significance, processes in multi-phase systems have so far remained poorly understood while reliable models for designing and optimising plants have been lacking. In the Multi-Phase project, Evonik is working with external partners on fundamental know-how for efficient and cost-effective multi-phase reactors.
Authors Dr. Marc Becker Process Technology & Engineering, Evonik Industries Prof. Dr. Robert Franke Advanced Intermediates, Evonik Industries Dr. Ingo Hamann Advanced Intermediates, Evonik Industries
The processes of multi-phase systems have already been studied, simulated and described in the past, yet the models and concepts currently available are insufficient for chemical practice. There are a number of reasons for this. Firstly, a lot of the literature is based on water-air systems because they are non-toxic, easy to handle and have known properties. However, as a blueprint of reality in the chemical industry, they are of only limited suitability. Commercial chemistry uses organic solvents more than any other, and most of the reactions are performed under pressure and/or at high temperatures. In addition, the gas composition changes along the length of the column because the gas takes part in the reaction itself.
Secondly, real bubble columns normally contain intermediate trays, heat exchangers or packing. These internals change the physical conditions and therefore the mass and energy transfer. Backmixing, dilution, dead zones and turbulence also play a big role for selective reactions, but that role has yet to be quantified hydrodynamically. Not least, there is almost no literature on the design of industrially feasible equipment. This is why construction and up-scaling usually involve a large number of laboratory experiments. Certain components – for gas input or the internals in the reactor, for example – must be changed when a reactor is scaled up. How to do that is still based on practical tests.
Multi-Phase will close knowledge gaps
How much simpler, more economical and resource-efficient it would be if there were tools and models that could analyse, describe and predict the processes in bubble columns. With this goal in mind, the Advanced Intermediates Business Unit and the Process Technology & Engineering Service Unit, together with partners from universities and various medium-sized companies, have initiated the “Increasing Energy Efficiency and Reducing Greenhouse-Gas Emissions through Multi-Scale Modelling of Multi-Phase Reactors” project (Multi-Phase, FKZ 033RC1102A), which is funded by the German Federal Ministry of Education and Research. The project has three objectives:
- To develop suitable measuring techniques for bubble column reactors
- To study relevant material systems in an industrial test reactor
- To provide new, broadly applicable models for multi-phase reactions
Internal and external experts in reaction kinetics, fluid mechanics and simulation have been working on these issues now for the past three years. For them, the challenge is identifying and understanding bubble reactor phenomena, which span a large scale range. The relevant reaction and transport processes occur on different scales of time and space – ranging from reactions at the molecular level through turbulence around the gas bubbles and boundary movement to large-scale flow processes. This places heavy demands not only on the measuring technology but also on the development of practical simulations for the planning and design of plants.
Testing methods successfully adapted
Up until 2013, a total of about ten different measuring methods were tested, adapted and successfully used in a pilot reactor. The engi-neers chose mainly non-invasive methods for refining and adjusting the measuring techniques. Working with the Helmholtz-Zentrum Dresden-Rossendorf, they were able, for the first time, to adapt gamma-ray tomography to a technical reactor and measure the gas content of a bubble column non-invasively. With the invasive methods they tested, the most important factor was compatibility with organic media, high pressures and temperatures. The use of Particle Image Velocimetry (PIV) from Intelligent Laser Applications allowed engineers to visualise liquid velocity fields as well as the size and distribution of gas bubbles. A Bruker FTIR spectrometer for measuring concentrations was used to determine the residence time distributions.
Measuring technology requires a real reactor to supply valid data. The planning and construction of a pilot reactor were thus impor-tant milestones of the Multi-Phase project. To this end, Evonik constructed a 4 m bubble column at the Marl site. By early 2014, engineers had tested a variety of organic solvent/gas systems, such as cumene/nitrogen and acetone/nitrogen – whose properties come close to actual systems – and recorded and measured their hydrodynamic processes. The innovation was that the reactor allowed an organic multi-phase reaction to be tested under high pres-sure (up to 40 bar) and at elevated temperatures of up to 80 °C. The evaluation of the data from the measuring phase is still ongoing and will be concluded by the end of this year. There are already indications, however, that a lot of the data from the literature does not correspond to what is actually happening in the pilot reactor.
Pilot reactor used for an actual process
From the beginning, an important aspect of the pilot reactor design was that it should not be specifically tailored to the experimental part of the project but also transferable to an actual, technical-scale material system belonging to the Group after the test phase has finished. For this reason, the reactor was largely financed by the Performance Intermediates Business Line, which has used it for hydroformulations since May 2014 and is now generating additional data with an actual reaction system for the project. The new measuring technology allows the system to be monitored over an extended period of time in order to gain a better understanding of the reactions taking place inside the reactor. After this, the equipment will be used to test new processes for the production of plasticiser alcohols and speciality chemicals with a variety of catalysts at different pressures and temperatures.
The tests on the pilot reactor are providing the basis for the development of new models that describe the processes and interactions in bubble columns far better than any others currently available. They will result in a simulation tool that allows engineers to plan and design reactors, shorten their development cycle, define the most economical operational mode from the outset and save energy and emissions during operation. This work is now in full swing. The experts are taking a two-stage approach to the modelling: firstly, they are working on short-cut models to facilitate a rough estimation of the most important parameters of a bubble column. This set of simple models will enable engineers to determine factors such as the size and dimensions of the reactor and heat exchangers, the quantity of gas and the number of trays required. Short-cut models can be transferred to almost any multi-phase reaction by using the specific measuring and substance data. They reduce experimental efforts and make it possible to evaluate the reactor and the necessary investment at a very early stage in the planning and design process. They are also suitable for optimising existing processes.
In addition to developing short-cut models, the engineers are refining and adapting computational fluid dynamics (CFD) simulations. CFD is an established method for describing flow processes based on various complex model equations. CFD simulations have been carried out since as early as the 1990s – but with the two phases as a continuum. Older models, therefore, simply describe the macroscopic structure of the flow. Only in the last few years have researchers and developers turned their focus to microscopic processes at individual bubbles in order to record material and energy transport to the phase boundaries, which are decisive for the yield and efficiency of the reaction. CFD simulations are quite detailed and require a computing time of several weeks. This makes it all the more important to validate these complex flow models through experiments as part of Multi-Phase and adapt them to organic material systems.
Multi-Phase is filling knowledge voids on a number of levels and in a variety of disciplines. By testing various measuring methods it has been possible to prove their potential for adaptation to an actual organic reaction system. The tests in the pilot reactor have delivered key knowledge about real reaction systems and process conditions. The models are closing the gap between experiments and simulation and will allow the processes in a bubble column to be calculated in an acceptable period of time, so that reactors can be planned reliably. When the project ends in April 2015, open-source software will be available for the community to use and further refine.
Even after the Multi-Phase project is over, not every question about bubble columns will have been answered. The findings will still be extremely useful, however. In the future, Evonik will be able to reduce the number of experiments needed to plan new bubble-column reactors and therefore shorten the development time. Project managers are confident that, thanks to the newly acquired knowledge, existing plants can likewise be optimised in areas such as raw material and energy consumption. And because the models describe the interaction of a gas and a liquid in general – not just in bubble columns – the results can be transferred to other reactor designs and multi-phase processes, from microbiological fermentations to wastewater purification plants.