There are times when water treatment seems to obey rules other than those currently understood by process engineers. By way of explanation, Nick Broomfield examines 'The H2O files'.

The technical director put down his cheese and pickle sandwich, leaned back in his swivel chair and gazed at the ceiling. "Then there was the case of the Loch Lomond monster."
He continued: "It didn't look like a problem job at first; 90m3/h flow and only a 0.1µS/cm conductivity - that's 10M½cm resistivity - specification. It should have been a doddle. The technology we used wasn't new. We put in a make-up deioniser - two bed reverse flow regenerated with a layered weak base/strong base anion because of the high organic content of the mains water supply - to produce better than 1M½cm.
"The deionised water went into a 400m3 tank and was pumped round a distribution ring main with a big mixed bed unit in it. It's exactly the same flow sheet that all the local power stations use to produce high purity boiler feed water and we expected to get much better than 10M½cm off the mixed bed."
"Now here's the strange bit." The technical director got up and started scribbling on the whiteboard.
"The make-up section performed much better than we'd anticipated, producing better than 5M½cm. The conductivity increased a bit in storage as you'd expect, but the really odd thing was that the mixed bed wouldn't get above 5M½cm." He poured a coffee and sat down again. "We were completely stumped.
"We undertook a series of on-site tests to establish the cause of the problem. Put our two best investigators on it - we could have done with those two from the telly, Sculder and Mully. They checked the plant thoroughly then told us what we already knew. Where is it?"
He opened a filing cabinet, pulled out a battered blue file and thumbed through it. "Yes, here it is: 'The fall in resistivity in the storage tank suggests pick-up of atmospheric carbon dioxide (which is normal in atmospheric storage tanks) and the subsequent increase in resistivity across the polishing mixed bed indicates that the mixed bed is removing the carbon dioxide. This indicates that the mixed bed is actually working as designed but that the impurities in the water which are causing the low resistivity cannot be removed by the mixed bed, that is they are not soluble species.
This is supported by the fact that filtration of the mixed bed polished water to 0.2µm improved the resistivity from 5 to 11M½.cm and ultrafiltration from 5 to 15M½.cm indicating that the problem is one of colloidal rather than ionic origin.' There, what do you think of that? Another coffee?"
He closed the file and returned it to the drawer then reached for the coffee pot.

Side stream trials

"Well, we set up a series of side stream trials so as not to compromise the water supply to production. First we tried granular activated carbon treatment of the raw water, but it gave no improvement in the treated water quality, indicating that the problem wasn't associated with organic matter. We didn't think it was, because the TOC was reduced by 80 per cent in the make-up plant - that's actually rather better than the 70 per cent reduction you'd normally expect if you had a separate organic scavenger unit.
"Chemical conditioning of the raw water with polyelectrolyte followed by multi-media filtration - a process widely used for colloid removal - did not give a significant improvement but conditioning with polyelectrolyte immediately prior to cation exchange achieved a treated water quality of better than 10M½.cm from the polishing mixed beds."
"The results of the on-site tests lead us to the conclusion that the low resistivity of the treated water was caused by a colloid. This, in itself, is most unusual. Most of the semiconductor factories in Scotland's 'Silicon Glen' have had problems at some time with colloids, usually clay (alumino-silicate based minerals), but these problems have been generally confined to blockage of filters or reverse osmosis pre-treatment plants. It is highly unusual for a colloid to contribute a measurable conductivity to water, even ultrapure water. In fact the mechanism by which a colloid, by nature an un-ionised particle, can give rise to conductivity is far from clear.

Some natural clay-like minerals such as zeolites have cation exchange properties, probably due to the replacement of silicon ions in the silicate structure by aluminium or ferric ions, and it is possible that some colloidal clays have the same property. Another possibility is the absorption into the colloid of organic matter containing ionisable carboxylic groups. The presence of ionisable groups on such colloids could result in free counter-ions in solution causing conductivity. The peculiar nature of this colloidal material may even have been the result of the lime which had been added to the Loch to counteract increasing acidity."
He looked a little dubious. "This, however, is speculative in the extreme," he added.

Nick Broomfield is Industrial Marketing Manager with USF.