Chips with six appeal

In 1994, Yorkshire Water's R&D department commissioned several projects to develop miniaturised, cost-effective technologies to monitor water quality in distribution mains. This was a very challenging target, but the managers were confident that it was justified; the technological capabilities it represented would provide operational benefits and, more importantly, would allow the company to keep ahead of advancing regulations. The fruits of this project are generically known within Yorkshire Water as the Censar family of products.

YW joined forces with Siemens Environmental Services, amongst others, to develop the technology, and succeeded in obtaining government funding from the DTI to the tune of around £3/4million. The consortium involved 13 organisations, including Biwater.

Early laboratory prototypes, though proving the concept, were not robust enough for field application. The finished article is another matter, however. Based on a ceramic substrate, the sensors use materials such as noble metals (such as gold and platinum), solvent-resistant polymers and ceramic frits in place of the membranes, glass bulbs and buffer solutions traditionally used in water sensors.

The active sensor components are made using film printing techniques which were developed for the electronics industry and are widely used in gas sensors. The sensors are generally made by screen-printing thick films of conductive dielectric inks, while the electrodes that connect the sensors to the analysing microcircuitry are made using thin-film techniques.

The Censar instrument uses several different types of sensor, all of which share a common reference electrode. Housed within the sensor head, this uses silver and silver chloride.

Chlorine measurement is handled by a two-part sensor which incorporates a pH-modifying electrolysis proton generator. This reduces the pH of the sample to below 5, allowing the thick-film chlorine sensing electrode to measure both molecular chlorine and HOCl.

The oxygen sensor also relies on this proton generator, although in this case to acidify the surface of the sensor to reduce fouling. The sensor itself relies on electrochemical voltammetry - the current flowing into the working electrode is proportional to the concentration of oxygen dissolved in the water. The sensor's geometry is designed to reduce the instrument's sensitivity to flow rates; this, along with the electrochemical antifouling measures, mean that the sensor doesn't require the membrane usually used in dissolved oxygen sensors and is far more robust.

The other two sensors, a pH electrode and redox sensor, both use a potentiometric technique, comparing the current in a working electrode with that of the reference electrode. The pH sensor uses a metal oxide deposited onto a printed conductive base material, while the redox sensor uses platinum ink, which allows it to respond to all the ions in the solution to give a measurement of overall redox potential.

The Censar chip is 25mm square and has an operational life of around six months, depending on the measurement media. Replacement is simple, as the chip connects to spring-loaded terminals within the sensor housing. Because the single device is manufactured using established mass-production techniques Yorkshire Water estimates that the total cost of ownership for measurement of all six parameters `equates to a third of that which we would expect from any one single parameter.'

This performance spurred the consortium on to develop further low-cost, versatile instruments. Optical research at Dublin City University proved to be the spur for another device: the SpectraCense turbidity and colour sensor. Previously, these properties each needed their own dedicated sensor, so once again, combining the two should lead to substantial cost savings.

The consortium commissioned the Dublin City researchers to develop their technology into a commercial device for water quality application in 1997, on a `fast-track' scheme. It took two years to come to fruition. Using low-cost LEDs, the device measures turbidity by looking at the intensity of light scattered at 90 degrees to a 840-920nm transmitted beam. This wavelength was chosen to reduce the effect of sample colour, which, in turn, is measured by checking the direct absorbance of a 400-460nm light beam. In this case, the turbidity sensor provides a correction factor to account for the effects of suspended particles.

As befits the codeveloper of the technology, Yorkshire Water has been quick to deploy Censar - it currently has 75 devices in use. The main reason for this is the cost benefit. Team leader Issy Caffour estimates that the total cost of ownership for all eight measurement parameters of Censar is £2740 per year, compared with £12 626 for traditional sensors. `The chlorine chip alone provides a 40 per cent saving on current Deplox chlorine analysers,' he says. To put this into context, Yorkshire Water's current cost of ownership for Deplox analysers is over £1.1million. Overall, Yorkshire Water estimates that Censar could save water companies in excess of £500million per year. PE