We were welcomed to the beautiful surroundings of Jesus College by Ian Wilson, who many will know through his work on studying fouling prevention in refinery heat exchangers. Ian is a Fellow of Jesus College.

Colin Ramshaw, (colin.ramshaw@ncl.ac.uk ) who chaired the morning session, then introduced the meeting and stated that this was the last under the current EPSRC contact.

David Reay (DAReay@aol.com ) then gave his usual update on ‘happenings’ since the last PIN meeting, (see www.ncl.ac.uk/pin for the talk).

The main proceedings then began:

Oscillatory Flow Reactors for Continuous Sterol and Biodiesel Production: Adam Harvey of Cambridge University (aph15@cam.ac.uk ) described his work on using the oscillatory flow reactor (OFR) for the continuous production of sterol and biodiesel fuel. He introduced us to the oscillatory Reynolds number, and discussed the significance of the velocity ratio in the tubular reactors. If this was in the range 4-10 one is getting close to plug flow, the equivalent of ten or more tanks in series. The niche use for OFRs is for the conversion of long residence time batch processes to continuous ones. Pulsed flow reactors are one option, but are long and thin. The OFR is the best option for intensification, as it is an order of magnitude shorter.

As an example, Adam discussed sterol production. This is done conventionally in a 75m³ batch reactor with a 50m³ fill at 115^oC, molar ratio 0.9. The time of reaction is 2 hours, with a 24 hour cycle timer per batch. Pressure is 2 barg. If continuous, this could be 1/12 of the size. The incentive for change is safety, with product quality and energy savings being secondary benefits. The equivalent OBR has a residence time of 10-15 minutes at 115^oC. Models have been derived to examine how much of the desired product can be achieved under a range of conditions, such as various temperatures. These suggested a better product spec. could be achieved at 85^oC with 15 minute reaction time and 95% conversion. This model was validated.

It was concluded that for sterol the OFR operated at a lower temperature, gave consistent product quality, was more flexible and had a volume of 5m³ . A large OFR is now being put on the site of a fine chemicals producer.

With regard to biodiesel, the reaction is based upon mixing vegetable oil triglycerides with methanol leading to the biodiesel fuel. One cannot run modern Diesel engines on pure vegetable oil, as it is too viscous. The methyl ester produced by the above reaction is a direct replacement for normal Diesel. The resulting fuel is sustainable, CO₂ neutral, gives off no sulphur and reduced emissions generally. It is seen as a good use for waste vegetable oils. Glycerol is a useful byproduct of the reaction.

Experiments showed a residence time in the reactor of 15-30 minutes to produce 80% by volume of input as fuel. It was concluded that the process was viable and a unit is now being designed for a local waste oil processor. In the future it is intended to examine optimisation, to look at different feedstocks and to replace methanol, which is not sustainable, with ethanol, which is. Suspended solid catalysts will also be studied. It is hoped to follow up ideas on integrating OFRs with mixer-settlers and a distillation column in the final production variant.

There was substantial discussion. Points covered included pressure drop and power needs – energy use was 4-5 times less than a CSF reactor, flow and mixing in the OFR had been optimised experimentally, and a large scale unit might be the subject of a decision to proceed within one year. Scale up was not a problem, even at 100-fold increases in production. ICI has done large pulser units, which someone felt may be one scale-up area of difficulty. The 5m³ unit tested had the same throughput as the 50m³ unit being replaced. The OFR can be used for more viscous fluids than vegetable oil.

Process Intensification – Relevance to the Food Industry: Christina Goodacre of DEFRA (Christina.goodacre-official@defra.gsi.gov.uk ) then introduced us to the food sector, and the possible opportunities for PI technologies, and the factors we had to bear in mind in looking at applying such technologies there. Christina is with the DEFRA Food Technology Unit. There is, she said, already mention of PI in the LINK Advanced Food Manufacturing Programme, which involves DEFRA, EPSRC etc. PI is a sub-theme of the Programme.

The characteristics of the food industry are:

• Large industry
• Wide range of products
• Short shelf life products have the highest growth rate
• The microstructure of products is important (e.g. one cannot cool chocolate too quickly and interfere with the structure
• Manufacturing efficiency is important (more robust processes, less waste, less down time)
• Process flexibility is important (shorter runs and frequent product changes)

Opportunities for PI may be spotted by the fact that it could reduce the metal volume in a plant needing cleaning.

What is the objective of PI with respect to the food processor? One might identify the following: less work in progress/shorter lead times; less hold up and waste; better control; better quality (e.g. emulsification); smaller footprint (important in a bakery, where the bread is only 5% of the volume of the oven); less energy and water use. In the last case, the Climate Change Levy and Integrated Pollution Control legislation is hitting companies in the food sector.

Christina then told us about the processes in the sector:

• Lots of heating & cooling, the former to cook and for food safety, deactivating micro-organisms
• Biscuit making – one heats and then removes water
• Cooling – viscous sauces, baked products etc.
• Fruit juices – evaporation/eater removal
• Mixing – liquids, liquids plus solids, liquids plus air (e.g. ice cream, batters – these are not reactions)
• Phase change – crystallisation (sugars, fats); freezing; gelation – getting networks to form in food
• Fermentation
• Maturation – lager takes 3 weeks to mature, cheese 6 months. One cannot write precise equations for many of these processes.

With regard to chemical reactions, flavour and colour involve these, (Maillard and enzymatic reactions). The inactivation of enzymes and fermentation are also this categorised.

BUT we have to bear in mind the following:

• Microstructures are very important, thus there should be no high shear stresses in the process
• Multi-phase mixtures such as ice cream are very complicated. This contains air, fat, water and crystals.
• Diffusion paths are often the limiting factor in PI. E.g. in bread baking there is not a lot you can do. In meat processing, the cooking is limited by the rate at which you can get heat into a lump of meat.
• Opportunities for PI are possibly greater at the liquid food processing end.

Challenges are:

• Matching heat and mass transfer to the reaction kinetics. For example, biscuits take about 10 minutes to bake. This could be reduced somewhat, by say 30%
• Rapid cooling of shear-sensitive viscous liquids and of solids (bread, meat joints). The former is a wide application area. Ideally the ‘negative microwave oven’ needs inventing!
• Rapid mixing of shear-sensitive materials (later on in the production stage)
• Crystallisation (chocolate, ice cream, margarine ‘seed engineering’
• Aim for a 5 minute clean down and change over of product. (Can take hours at present) The Ice pig of Joe Quarini at Bristol was cited as an innovative example.

During discussion, pulsed electric fields, and ultrasound were offered as possibilities. The malt kilning process could be intensified, but first look at kinetics then the kiln design. Pasteurisation was queried. There is a legal requirement to keep the milk at e.g. 72^oC for 10 minutes. A higher temperature for a shorter time can be used, but flavour is affected (UHT milk).

Christina said that she was happy to discuss ideas for PI in food with PIN members.

The Oscillatory Flow Mixer applied to Water Treatment: Malcolm Mackley of Cambridge University then introduced us to the use of the OFR as a method for treating domestic waste water, using it as a biomass filter. The challenge is to treat effluent containing 50-200 mg/l which is normally passed to gravity separating tanks. Is there a better separating method? Is there a role for oscillatory flow mixing?

The concept studied, in conjunction with Anglia Water, was to capture the biomass in the effluent using a loose fibre packing in the OFM. It was felt best to use high voidage (98%) glass fibres to capture the biomass. In this use the filter needs regular regeneration, and this is done by violent oscillations in the fluid, with the filter in situ. This leads to a slurry of high concentration.

The process contraints are several:

• Reliable, durable and low pressure drop
• Effective solids capture and release
• Cope with varying feed concentrations
• Cannot use chemical additives
• Unattended running necessary

Cambridge looked at many filter configurations, and is still working on finding a satisfactory solution. One investigated is a pack of fibres between stainless steel honeycomb grids to give support. This works to a degree, but becomes less effective as the number of cycles increases. The filtration mechanism seems to be firstly deposition of the biomass, followed by breakthrough of the biomass as the loading becomes very high. Oscillations of the flow effectively strip off the biomass from the filter, as conceived, but in the next cycle the carry-over/breakthrough is greater. It is hypothesised that the fibres move and allow higher voidage regions, losing filter effect. The fusing of the fibres to give greater strength is now being studied.

Other ideas in discussion included using another fibre material, vibrating the elements, looking at reverse flow, using a Stirling engine-type mesh regenerator, or using ceramic foams (Porvair).

Scale up of a Continuous Process using Diazomethane: Lee Proctor of Phoenix Chemicals (Lproctor@phoenixchem.com ) described the development and scale up of a 4-stage continuos process requiring diazomethane, the output going into anti-aides drugs. The goal set was to get a plant of output>2000 tpa of chloroketone, which could handle unstable intermediates. Stage 1 involved a mixed anhydride, with a PFA tube mixer reactor capable of feeding 2000tpa to a centrifugal separator. Stage 2 involved diazomethane (DAM), which is explosive and a very small amount in a nitrogen atmosphere can explode at ambient conditions.

Lee described current methods of manufacture, and set the scene for the design aims – non flammable system; continuous process; a need to purify the DAM and determine the lel. Full scale explosion testing had to be done on the full scale final reactor design. There was also a need to develop methods for continuously monitoring DAM.

The process resulting was a reactor sitting on a bursting disc rated at 2.5 bar, surrounded by a quench liquid.

The process advantages were rapidity with >95% yield, no moving parts, no moving seals or dead spaces. A 1 litre pilot reactor was tested, then a 200 l full scale reactor. It was initially tested with a gas mimicking DAM, which is equivalent to 7% ethylene in air.

The third stage is the making of diaketone, carried out in a packed column followed by a mixer-settler and continuous evaporation.

Lee said that the timescale from go-ahead with the major investment to completing the plant was 1.5 years. The FDA (USA) and others to some extent influence the use of plants like this, as continuous processes help to control impurities, hence it is beneficial for making pharmaceutical intermediates.

Multiphase Reactor for Organic Nitration: The final formal talk was by John Burns of Newcastle University (J.R.Burns@ncl.ac.uk ). John talked about slug flow in micro-reactors. He is developing multiphase micro-reactors, geometrically scaled to produce short diffusion paths, the residence time being determined by length/velocity.

There were several methods for contacting two fluid phases – parallel flow; droplet emulsion; and slug flow, the latter being the focus of work at Newcastle, in reactors where channel sizes range from 50 to 500 microns. John described the factors affecting slug flow in glass channels and discussed examples, such as titration in a slug flow unit.

Organic nitration reactions were carried out at Newcastle. The reaction rate was found to increase as diameter decreased from 250 to 127 microns, and increased also with increasing velocity. Comparable performance is being obtained to those used in industry, but the micro-system is still not optimised. Scale up to 1 kg/day is equivalent to 10s of micro-litres/sec, and can be done in blocks with 1000 channels. The remaining main problem is manifolding all these channels. The ultimate aim is to integrate these units into intensified desktop units.

In discussion, it was pointed out that the heat effects would necessitate use of an excellent heat exchanger block, with alternating reaction/heat transfer channels. With regard to pressure drop, John said that one could engineer whatever required. The pressure drop was less than 1 bar across the block.

Following a buffet lunch overlooking the main Jesus College dining hall, and an opportunity to view posters of the work at Cambridge, the impromptu session of 5 minute talks began.

Prof. Gupta, (J.P.Gupta@lboro.ac.uk ) reported upon the survey he has been carrying out into Inherently Safer Design, (IHD). Questionnaires had been sent to companies and journals etc., and 67 replies were received. It was found that a few had practiced ISD for decades, and ISD had a favourable impact on the balance sheet. It can influence R&D activity and should be advocated by the IChemE as part of undergraduate courses.

Drummond Hislop (Hislop@msn.com ) introduced a Stirling engine eat exchanger concept which could potentially have other applications, such as heat exchanger-reactors. It is currently used as a gas cooler and has a low pressure drop. It comprises 6 annular elements and was initially made of chemically etched and diffusion bonded rings. A second variant comprised a series of grooved sleeves in an outer barrel which is water-cooled. Drummond is looking for other PI applications – contact him if you wish to follow this up.

Dawn Arda of Cambridge University (DRN20@cam.ac.uk ) discussed her work on gas-assisted extrusion of polymers. The problems are die swell and extrusion instabilities. Gas is used to lubricate the wall and reduces the wall shear stress. This results in much less die swell and 20% less pressure drop. The extrudate is cooled more rapidly, too.

Richard Bahu (Richard.Bahu@compuserve.com ) described the new Manufacturing Molecules Initiative (MMI), which is currently the only initiative under the DTI Chemicals Branch. The aims are to reduce the time to develop complex molecules, and training. Target molecules are drugs and speciality organics. Technical priorities include chiral formation, new catalyst technology, novel synthesis and enhanced fields. Support is available for scoping studies, pilot studies and demonstrations, with rapid turn-around of proposals promised. More data on: www.gov.uk/mmi

Andrew Green of BHR Solutions (Agreen@bhrsolutions.com ) introduced us to the new BHR project on a feasibility study for the development of multi-stage PI plant, supported under the low carbon technology initiative. This is looking at very fast reactions and potential partners are being sought. Contact Andrew if you would like further information.

Jeremy Matcham, who was until recently at Cranfield University, (Jeremy.Matcham@yahoo.co.uk ) introduced ideas on intensification of water electrolysis and reforming and storage of hydrogen. Jeremy ran an interesting hydrogen energy web site and has numerous contacts in the above area. Contact him for more data.

Finally, Henk van den Berg, (h.vandenberg@ct.utwente.nl ) then promised us a talk at the next meeting on the perspectives of PI for the next 50 years. This was a most appropriate promise on which to close the meeting, which was the last under EPSRC funding, for which we thank them.

Thanks to Adam Harvey and his colleagues for hosting the meeting, arranging the lab. visit, and providing other local assistance.

(Minutes written by David Reay, based upon his notes, 7 December 2001).