Going through the proper channels

Cutting costs, reducing waste and conserving energy: three of the process industries' most keenly-chased goals. All of these can be achieved by the techniques of process intensification, according to the British Hydromechanics Research (BHR) Group, whose project involving the Marbond reactor is a prime example of the techniques.

The Marbond is a compact heat exchanger (CHE), designed and developed by Wolverhampton-based firm Chart Marston. Its compactness gives it exceptionally high heat transfer characteristics, and its design allows several streams to pass through the reactor many times, improving its reaction and heat transfer characteristics.

CHEs have several advantages over conventional shell-and-tube heat exchangers. They are more effective; that is, the ratio of heat transferred to the maximum possible heat transfer is higher. Their volume and weight are lower, because of their high heat transfer coefficients and the large area available for heat transfer per unit volume; this, in turn, reduces their installation costs and tightens the control of temperature they convey. They provide energy savings, because the driving temperature difference between the process and heat transfer fluid streams is smaller. Finally, they make processes safer, as smaller amounts of process fluids are needed, and the tight control of temperature makes runaway reactions far less likely.

The goal of BHR and Chart Marston's research has been to combine the advantages of CHEs with the ability to perform reactions inside the heat exchanger. Integrated chemical reactor/heat exchangers (HEX reactors) such as the Marbond allow mixing, chemical reaction, and heat exchange to be performed in a single piece of equipment, opening up new possibilities for the development of very different chemical plants to those so familiar to engineers and operators.

BHR's fluid mixing researchers now claim that the reactor is significantly more selective than traditional batch or continuous reactors with cooling jackets. Put into practice on an industrial scale, this presents the appealing prospect of smaller, less intrusive plants, which are cheaper to build, run and maintain, but are more productive and profitable and generate less waste.

The design of HEX reactors such as the Marbond is based on the process intensification approach, where the fluid dynamics of a process are matched to its chemical, biological and/or physical requirements. The criteria for PI are broad: the techniques are suitable for reactions that are fast, produce or absorb heat, or form by-products. They can be single phase, two-phase or multi-phase, such as catalytic reactions.

Inside the Marbond is a stack of plates that are etched photochemically to form a series of slots. These plates are positioned very precisely, in such a way that the slots form a series of convoluted but separate flow paths through the reactor. Solid plates separate adjacent flow paths, so different fluid streams, such as reactants and coolants, can flow parallel and adjacent to each other.

The separator plates are also perforated, with the holes aligned with the flow paths for the primary reactant. These are used to inject secondary reactants into the reactor. This design means that each heat-producing layer is next to a heat-removing layer, so the heat of reaction can be removed almost as soon as it is formed.

The plates are bonded together by a process known as diffusion bonding. The stack of plates is pressed together in a vacuum furnace, which forces the metal grains in the plates to grow into each other. This process results in joints which are as strong as the parent metal.

BHR evaluated the Marbond by using it to perform a reaction that produces an industrial azo dye, and comparing its performance with that of a conventional plate-fin heat exchanger (PFHE). This reaction produces four products at rates which differ by several orders of magnitude - similar to an industrial process that produces several products at different rates, only one of which might be useful or valuable. The distribution of these products inside the reactor depends strongly on the intensity of mixing which the design can produce. Because of this, it is often used to assess the mixing performance of high-intensity `plug flow' reactors. By choosing and controlling the reaction conditions carefully, process technologists can use the amount of one of the dye products to measure the mixing intensity in the zone where the dye was formed. The more dye produced, the lower the mixing intensity.

This experiment showed that the Marbond is much more efficient than the PFHE. At each energy dissipation rate, the former produces less dye than the latter. In industry, this would equate with producing less by-product.

Another reaction, known as the Walker Scheme, characterises both mixing and heat transfer. The scheme comprises two competing reactions. Hydrogen ions, derived from HCl, are reactants in one step, but act as a catalyst in the other. This is another phenomenon often seen in industrial processes - the products of the acid-catalysed reaction are equivalent to by-products formed by side reactions.

In this case, the reaction was carried out both with and without a coolant flow. The results showed that removing the heat of reaction can reduce the amount of side-products formed. The difference was modest, but the energy produced by the Walker scheme is barely a tenth of that produced in some industrial operations. For these reactions, the benefits of the Marbond should be greater.

Chart Marston and BHR have high hopes that industry will share their enthusiasm for the Marbond. Not only would it reduce capital expenditure, it would improve companies' environmental profiles. For example, the smaller, cheaper plants would use much less energy than their current equivalents. BHR estimates that the introduction of HEX reactors in suitable processes throughout Europe could reduce annual energy consumption by over 75GJ, and reduce annual CO2 emissions by 4.8million tpa.

BHR will explain its process intensification research at a conference in Antwerp on 25-27 October. For more information on this, contact Tracey Wheeler on 01234 750422.