Separation and breakdown
Elf Atochem generates chemicals which need some careful treatment
In 1997 a pilot plant was built at Elf Atochem's Widnes factory to separate and treat such a mixed waste stream. Following successful trials and fine-tuning of the process, designer Membrane Extraction Technology (MET) has now won another three-year contract to run the plant.
The plant treats a waste stream which contains inorganic aluminium trichloride (AlCl3) at 30% plus two organic compounds, benzene and toluene. The stream arises from processes used to make aromatic ketones, known as Friedel-Crafts reactions. Aromatic ketones are used in a number of products, from sunblock to agrochemicals. For each tonne of product, 2-3t of AlCl3 solution is produced. The untreated solution is contaminated with traces of product and solvent and is classified as hazardous waste.
The system which has been installed is an extractive membrane bioreactor (EMB), which filters the stream through dense phase membranes (first made of silicon rubber). The pores in the membranes will let through bipolar organic molecules but not polar ionic molecules. Once separated the organics are broken down biologically in a 'biomedium' stream.
When first set up in 1997 the EMB system was run for over 130hrs to provide commissioning data. During the trial the plant treated over 20t of AlCl3 solution. Samples of the AlCl3 solution entering and leaving the pilot plant were taken daily and analysed for the presence of benzene and benzophenone (reactant/ solvent and product of the acylation reaction), and samples of the biomedium were taken for chemical oxygen demand (COD) analysis. When the trace organic contaminants have been removed, the AlCl3 solution can even be used as a flocculant in wastewater treatment.
After the tests ethylene-propylene-diene monomer (EPDM) was chosen as the final membrane material because it could withstand the operating conditions of the pilot plant. For six months from May to October 1998 the plant generated clean AlCl3 solution. The benzene concentration in this material was reduced by an average of 98% from 150-200mg/l to 2-5mg/l. The concentration of AlCl3 in the waste stream was higher than originally expected, as was the flow rate, so at first the benzene removal was not quite as effective as hoped. Modifications to the membrane modules were also needed which reduced their surface area by 25%.
No major problems were seen with the EPDM membranes despite the corrosive nature of the AlCl3 solution and the presence of the organic compounds. After six months of operation the membranes were tested and found to be just as strong as when they were installed in terms of burst pressure.
During the project, a number of products based on benzene and toluene were made along with some from the acid halides benzoyl chloride and 4-chlorobenzoyl chloride. The switch from one product to another did not affect the EMB system and the microrganisms were able to continue functioning. According to MET's process engineer Dr Andrew Boam: "We have most types of starter culture that are needed for treating this sort of waste stream. If you cannot find appropriate organisms you can go to an old industrial site or effluent outfall where this type of waste is present and find them." The bioreactor is designed so that if for any reason the culture gets knocked out, there is a reserve supply of microorganisms. In the case of the EMB system, a subsection containing the appropriate culture constantly feeds the main bioreactor. If there is any problem with the main bioreactor it can be restarted using the microorganisms in the protected sub-section.
From July 1998 the pilot plant has managed to process around 6t/d of the AlCl3 solution. Typically 98-99% of the benzene has been removed from the solution. Refinements to the technique have now brought the benzene concentration down to 1µg/l.
To date the most popular methods of treating AlCl3-rich wastewaters have been steam stripping and carbon adsorption. EMB is a particularly clean option as the toxic compounds are broken down rather than concentrated. Steam stripping generates a condensate contaminated with organics and carbon adsorption generates saturated carbon which also has to be disposed of or cleaned. In terms of capital and operating costs, EMB is said to be competitive with carbon adsorption and significantly cheaper than steam stripping.
On the way to mars
MET is currently developing a different membrane process, with pilot trials underway at a secret site. The new process, called MARS, which instead of breakning down the compounds removed from the waste stream, extracts them to a standard suitable for reuse.
According to Dr Boam: "In the MARS process the organic waste stream passes through a membrane into an acid solution, which changes the organic substances into salts. Phenol, for instance, changes into sodium phenate." The change reduces the osmotic potential of the phenol to zero, so the organics continue to flow across the membrane. Once the sodium phenate has been turned back into phenol using some clever chemistry it can be fed back into the manufacturing process, saving the company thousands of pounds. Dr Boam said: "You could potentially extract inorganics as well, but generally it is not worthwhile." Organic substances are much more valuable in comparison to the inorganic catalysts used in Friedel-Crafts reactions. Once trials of the MARS process are complete it may be used on a wider scale.