Take a walk around a typical chemical plant. Look at the towering columns, the serpentine pipework, the looming tanks and vessels. The scene hasn’t changed much for 30 years. But new engineering techniques could change the way chemical plants look, and operate, for good.
Chemical engineering is based on unit operations. Each part of the process – pre-treatment of feedstocks, reaction, separation – is carried out in standardised pieces of equipment, linked by pipework to form a plant.
However, as companies become more profit-focused, they are demanding that engineers make their processes more efficient. Traditional chemical engineering is reaching the limit of efficiency. So engineers are turning to new techniques known as process intensification (PI), to force their processes to yield greater efficiencies.
Process intensification uses engineering principles to make chemical processes more efficient. Yields are increased, energy use lowered and often, the cost of the plant is reduced because the equipment it produces is more compact than conventional plant.
It does this by matching the fluid dynamics of the process to its chemical, biological or physical requirements.
This can have several outcomes. It could involve combining two or more pieces of equipment into one. Invariably, it sends process engineers back to the drawing board, having to reconsider their old designs from first principles.
This means a closer relationship with the chemists who designed the reaction. The researchers have to investigate what factors are needed to speed up a reaction and how these can be designed into pieces of equipment.
Several factors are important to assess whether a process is suitable for PI. It is most useful for reactions which proceed by an equilibrium (where the reactants and products coexist in the reaction mixture), generate or absorb heat and produce by-products. These factors allow the reactions to be manipulated. For example, if a reaction produces by-products, removing these substances as soon as they form forces the reaction to consume more reactants and generate more products to maintain their relative concentrations.
Reactive distillation rationale
This is the rationale behind a form of PI known as reactive distillation (RD), which involves installing a catalyst bed within a distillation column. The technique, suitable for processes that generate at least two products with different boiling points, works by distilling off one of the reaction products as soon as it is formed, thus forcing the reaction to generate more products.
Reactive distillation was first developed in the 1920s, but has only been widely used relatively recently. This is partly because there was no demand for products from processes suited to the technique – until the demise of leaded petrol and the replacement of leaded additives with methyl tert-butyl ether (MTBE). This is an ideal RD process because MTBE is far more volatile than the compounds used to make it, methanol and isobutylene.
Also, RD columns are difficult to design. There are many problems, the first of which is how to hold the catalyst.
The company which first used RD commercially, CD Tech, developed the `Texas teabag’ – a porous fibreglass fabric sheet, sewn into a series of long pockets which are filled with catalyst, then rolled up into a bale resembling a Swiss roll. BP Amoco’s solution involves a folded steel mesh sheet whose convolutions hold shaped catalyst pellets.
`Structured packing like this is used inside distillation columns to give the vapours a surface to condense on,’ explains Michael Jones, director of BP Amoco’s research division at Sunbury. `We’re now looking at ways of using this packing to hold catalysts.’
Another problem is the size and position of the catalyst bed. The best position – where the temperature is ideal for the reaction – is dictated by a complex interplay of boiling points, energy levels and vapour pressures. A European Union-sponsored research project, involving Neste, BASF and Snamprogetti, and universities in the UK, Germany and Finland, is developing software to aid the design process.
`It will tell you where you should put the feeds in, where you should put the catalyst, what would be the distribution of the products,’ says Juhani Aittamaa, Neste project director. `But we need more research into reaction profiles and catalysts to make it work properly.’
With the right process, RD can save companies millions. The plants become cheaper to build, with fewer vessels and less pipework. Production costs are reduced – as much as tenfold – because the heat that drives the reaction also drives the distillation process.
Using the same scientific principles, combined heat exchanger/reactors (HEX reactors) speed up reactions by manipulating their energy profiles. These combine three processes – mixing, reaction and heat exchange – into one piece of equipment. They offer increased selectivity in reactions, forcing the reaction to produce more of the desired product.
The Marbond reactor, developed by PI specialist BHR Group and process equipment maker Chart Marston, won last year’s Manufacturing Industry Achievement Award for process innovation of the year. It consists of steel plates which are photochemically etched into a complex pattern of slots and holes, stacked so that the slots form a series of convoluted but separate flow paths through the reactor.
This ensures that a reaction stream which produces heat is running in a stream adjacent to a coolant. Removing the heat of reaction as it is formed forces the reaction to generate more product.
`Even greater efficiency can be achieved if both streams are reaction streams, one generating heat, another absorbing it,’ explains Christopher Phillips, who led the Marbond development at BHR Group. `In this case, each reaction accelerates the other.’
The twists and turns of the flow paths ensure that mixing is more uniform than in other reactors. This, and the heat removal, boosts the selectivity of the reaction. `We estimate that, if HEX reactors were used in every suitable reaction in Europe, energy consumption would fall by 75GJ and carbon dioxide emissions would fall by 4.8 million tonnes a year,’ says Phillips.
Stuart Nathan is deputy editor of Process Engineering