Ž . Sensors and Actuators B 69 2000 336–341 www.elsevier.nlrlocatersensorb An electronic nose system for monitoring the quality of potable water Julian W. Gardner a, ) , Hyun Woo Shin a , Evor L. Hines a , Crawford S. Dow b a School of Engineering, UniÕersity of Warwick, Gibbet Hill Road, CoÕentry CV4 7AL, UK b Department of Biological Sciences, UniÕersity of Warwick, Gibbet Hill Road, CoÕentry CV4 7AL, UK Received 20 September 1999; accepted 22 February 2000 Abstract A measurement system has been developed for the testing of cyanobacteria in water, and it consists of three main stages: the odour Ž . sampling system, an electronic nose e-nose and a CellFacts instrument that analyses liquid samples. The e-nose system, which employs an array of six commercial odour sensors, has been used to monitor not only different strains but also the growth phase of cyanobacteria Ž . Ž . Ž . i.e. blue-green algae in water over a 40-day period. Principal components analysis PCA , multi-layer perceptron MLP , learning vector Ž . quantisation LVQ and Fuzzy ARTMAP were used to analyse the response of the sensors. The optimal MLP network was found to classify correctly 97.1% of the unknown nontoxic and 100% of the unknown toxic cyanobacteria. The optimal LVQ and Fuzzy ARTMAP algorithms were able to classify 100% of both strains of cyanobacteria samples. The accuracy of MLP, LVQ and Fuzzy ARTMAP in terms of predicting four different growth phases of toxic cyanobacteria was 92.3%, 95.1% and 92.3%, respectively. These results show the potential application of neural network based e-noses to test the quality of potable water as an alternative to instruments, such as liquid chromatography or optical microscopy. q 2000 Elsevier Science S.A. All rights reserved. Keywords: Cyanobacteria; Gas sensor array; Electronic nose system; Neural network; Fuzzy ARTMAP 1. Introduction The diversity of cyanobacteria species is becoming both a severe problem in the quality of potable water and a common source of odour pollution in freshwater reservoirs w x and local environmental water 1,2 . It has been associated with sewage effluent, industrial effluent, waste products from agriculture, and animal wastes from intensive farm- ing. Cyanobacteria, which are usually called blue-green algae, grow in lakes and reservoirs and can be a serious nuisance due to their unpleasant odour and taste, in the case of reservoirs. The cyanobacteria is the largest group of photosynthetic prokaryotes and contain chlorophyll that differs from the bacteriochlorophylls of the photosynthetic eubacteria. The main problem associated with certain cyanobacteria is that they can produce toxins that are poisonous to cattle, wildfowl, fish and people. Many species of cyanobacteria have been observed to produce these toxins. The toxins can be divided into three groups: peptide hepatotoxins, neurotoxins and lipopolysaccharides. ) Corresponding author. Tel.: q 44-24-76523695; fax: q 44-24- 76418922. Ž . E-mail address: j.w.gardner@warwick.ac.uk J.W. Gardner . Thus, appropriate methods to detect and quantify these toxins in natural waters are very important. Here, we Ž . wx report on the use of an electronic nose e-nose 3 , based Ž . on an array of six metal oxide semiconductor MOS sensors, to analyse cyanobacteria cultures grown in water, with the intent to provide a simple tool to test the quality of potable water as an alternative to analytical instruments that are based on liquid chromatography or optical mi- croscopy. Earlier work on microbial detection has been wx reported on Escherichia coli grown in blood medium 4 , wx wx biopharmaceutical process 5 and vagina infection 6 , and shows that an e-nose has potential application within the fields of bioprocess monitoring and medicine. 2. Experimental We have constructed a measurement system for the testing of the cyanobacteria over a period of 40 days. The system consisted of three main parts: the odour sampling Ž unit, the Warwick-modified Fox 2000 unit Alpha MOS, . Ž France and a CellFacts I instrument Microbial System, . Coventry . The number of cells and cell size are sensitive indicators of the physiological status of algal cells: they 0925-4005r00r$ - see front matter q 2000 Elsevier Science S.A. All rights reserved. Ž . PII: S0925-4005 00 00482-2