Chemical sensor array for multicomponent analysis of biological liquids

The development of new methods of on-line monitoring of biological liquids such as human blood, blood plasma or dialysis solutions for arti®cial kidney is still an urgent task. Ion-selective chemical sensors (ISEs) are convenient instruments for this purpose and are widely used in commercial blood analysers. However, they have signi®cant drawbacks. The promising way to overcome them is the use of array of sensors with cross-sensitivity together with pattern recognition methods. The present work has been concerned with the development of a multisensor system for simultaneous determination of Ca2‡ , Mg2‡ , Na ‡ , HCOÿ 3 , Cl ÿ , H ‡ and HPO2ÿ 4 ions in solution modelling human blood plasma. The sensor array included 30 sensors based on non-speci®c original PVC materials, both cation- and anion-sensitive. Sensing materials for the array were chosen on the basis of cross-sensitivity estimation results with the help of the method that we suggested earlier. Back-propagation arti®cial neural network was utilised for data processing. It was demonstrated that the sensor array approach allows to measure the content of all these species with an average precision of about 1±4% depending on the component in typical ranges for biological liquids, including Mg2+‡, HCO3- and H2PO4- determination and pH measurement without an oxide glass pH-electrode. # 1999 Elsevier Science B.V. All rights reserved.