USE OF MATHEMATICAL-MODELS TO DESCRIBE DYNAMIC BEHAVIOR OF POTENTIOMETRIC BIOSENSORS - COMPARISON OF DETERMINISTIC AND EMPIRICAL APPROACHESTO AN UREA FLOW-THROUGH BIOSENSOR

The response of a flow-through biosensor is the consequence of complex interrelations between the dilution of the sample in a continuous str eam (governed by physical variables) and the phenomena involved in the reactions between enzyme and substrate (governed by diffusion and kin etic processes). These interrelations make it difficult to know in adv ance the influence of manifold parameters in the biosensor response. I f a reliable model of these processes is attainable, it will be possib le to test rapidly, by computer, the effect of the variation of differ ent operating variables on the response of the biosensor. This paper p resents the results obtained when two models are applied to describe d ynamic behaviour of a potentiometric flow-through urea biosensor. One of these models uses an empirical approach based on neural networks, w hile the other proposes a deterministic approach that takes into accou nt reaction-diffusion equations.