Double your money

Distillation is one of the oldest tricks in the process engineer's book, and the towering distillation column is one of the defining features of chemical and refining complexes the world over. You might think that there are few surprises in this particular area. But interest is now growing in a technique which can transform the humble distillation column into almost a stand-alone chemical plant, combining reactor, separator, and recycling loop into one money-saving, efficiency-improving, performance-enhancing unit.

The technique, known as reactive distillation (RD), allows two important operations of a chemical plant, reaction and separation of products, to be carried out simultaneously. It's the subject of much interest for chemical and oil companies, and a group of petrochemical majors, along with European universities, is currently researching a computer system theat might bring RD into a whole new range of reactions.

The technique has two main advantages over conventional plants. First, RD plants are cheaper to build, as they use one vessel instead of two, and so need less pipework and other fittings. Savings can run into millions of dollars, and production costs can be cut as much as tenfold. Second, the technique can improve the yield and product quality from processes involving reactions that reach an equilibrium. RD helps here because equilibrium mixtures are dynamic individual molecules constantly break down and reform. If the product is removed, the system tries to correct the imbalance by converting more of the reactants into products, and eventually all the feedstocks are consumed. And removing the product is exactly what RD does.

The technique isn't new it dates back to the 1920s but it wasn't until about 15 years ago that industry started taking it seriously. This wasn't because of any deficiency with the process, says Michael Jones, who is directing BP's research into RD at its laboratories in Sunbury, but was driven by a demand for a product ideally suited to RD. As the worries over leaded petrol grew, oil companies began searching for an environmentally-friendly alternative, and the best option proved to be methyl tert-butyl ether (MTBE). The reaction which forms this product between methanol and isobutylene is an equilibrium, and the ether product is much more volatile than the feedstocks. Therefore, in an RD column, the ether formed is removed immediately, driving the reaction to completion.

Changes in the ether

The first commercial use of RD to make MTBE was in 1981, at Charter Oil's refinery in Houston, Texas, using technology designed by a local firm, Chemical Research and Licensing. This later teamed up with contractor ABB Lummus Global to form the company CD Tech, now the leader in this technology with some 70 units operated by licensees around the world.

One reason for the comparative rarity of the process is the difficulty of designing RD columns. They have to combine features of reactors and distillation columns, most importantly a catalyst bed, which has to be located at the point where the temperature best suits the reaction. But as well as facilitating the reaction, the catalyst also has to provide a surface where the vapours can condense, while allowing hot vapour to rise through the column and liquid to trickle down.

Much of BP's research into RD is aimed at tackling this problem. 'One way of doing this is to use structured packing,' says Jones. 'This is used inside distillation columns anyway to allow the separation of the liquid and the vapour. What we're looking at is holding or supporting different catalysts within different sorts of packing.'

Approaches to this differ. One of CD Tech's inventions is a device known as the 'Texas Teabag' the first system to combine structured packing and catalyst, claims its director of operations, Tom Hickey. This consists of porous fibreglass fabric, sewn into a series of long pockets which are filled with catalyst pellets. These are then wound into spirals with stainless steel mesh into bales 'they look a bit like Swiss rolls,' says Hickey and stacked at the appropriate point inside the column. Some of BP's research into RD uses a similar idea, Jones notes, with sheets of fine steel mesh folded many times and formed into a cylinder. The catalyst-bearing pellets are then dropped into the folds.

BP's structured packing work ties in with the EU-sponsored international research project, where BP's consortium partners include Neste, BASF and Hoechst, the Italian contractor Snamprogetti, and Aston, Dortmund, Clausthal and Helsinki Universities. Part of the BRITE-EURAM scheme (Basic Research in Industrial Technology Europe and European Advanced Materials III), the three-year project is aimed at developing a software tool to assess whether particular reactions are suited to an RD approach.

The consortium is blending experimental results gathered from the five industrial partners' pilot plant-scale RD columns (BP's, at Sunbury, is 5m high a commercial unit would be ten times this size) with computer modelling techniques, mostly carried out at the universities. The pilot plants are running a range of reactions with different catalysts, packings and column sizes, because the complex interplay of energy levels, boiling points and vapour pressures makes it near-impossible to predict from theory alone how any particular reaction will turn out or how the column should be constructed.

The goal is to develop a predictive tool that will allow companies to predict whether a certain process can be operated in an RD column. This will require very little information about the process, says project coordinator, Juhani Aittamaa of Neste: 'just general knowledge about the properties of the reactants and products, and the rate of the reaction.'

The second part of the package will go into more detail. 'It allows us to simulate what the column would look like,' Jones says. 'It'll tell you where you should put the feeds in, where you should put the catalyst bed, what will be the distribution of the products, how big a column you'd need, how many separation stages, and so on.' However, points out Aittamaa, more information is needed for this system, and a large amount of research into the reaction profile and catalyst would be needed.