© by PSP Volume 21 – No 12b. 2012 Fresenius Environmental Bulletin 4030 APPLICATION OF FLUORESCENCE EXCITATION– EMISSION MATRICES AND PARAFAC ANALYSIS FOR INDICATING THE ORGANIC MATTER REMOVAL FROM MICRO-POLLUTED RAW WATER IN WATER TREATMENT PLANT Erdeng Du 1, 2 , Pengrui Cao 2 , Yue Sun 2 , Naiyun Gao 1, * and Liping Wang 2 1 College of Environmental Science and Engineering, Tongji University, Shanghai, 200092, China 2 School of Environmental and Safety Engineering, Changzhou University, Changzhou, 213164, China ABSTRACT Fluorescence excitation and emission matrices (EEMs) and parallel factor analysis (PARAFAC) were used to explore the organic matter removal from micro-polluted raw water in Zhongqiao water treatment plant. Based on PARAFAC analysis, four fluorescent components were extracted, including tyrosine-like, tryptophan-like, UV humic-like and visible humic-like substances. The water samples were dominated by protein-like substances, in- cluding tryptophan-like and tyrosine-like substances. The fluorescent components have the spectral features similar to those previously identified from EEMs of diverse aquatic environmental samples. The changes of maximum fluores- cence intensities (F max ) illustrated that the fluorescent com- ponents could be effectively removed during the water treatment process. Ozonation and BAC filtration played an important role in the removal of organic components, with the reduction of F max by 76.4% for tyrosine-like, 74.2% for tryptophan-like, 60.1% for UV humic-like, and 46.7% for visible humic-like. Protein-like substances were more readily eliminated than humic-like substances during BAC (biological activated carbon) filtration. PARAFAC analysis and F max could be regarded as an effective tool and indicator for water quality analysis and water treat- ment plant evaluation. KEYWORDS: organic matter, water treatment plant, fluorescence excitation– emission matrices (EEMs), parallel factor analysis (PARAFAC) 1 INTRODUCTION The conventional water treatment process is mainly concerned about the removal of colloidal particles and bac- * Corresponding author teria in water. With the rapid development of economy and society, in both developing and industrialized coun- tries, a growing number of contaminants are entering water suppliers from human activity: from traditional compounds to emerging micro-pollutant, such as endocrine disrupters, pharmaceuticals and nitrosamine [1]. Therefore, more re- searchers focus on the study of the conventional and ad- vanced water treatment process for the removal of the organic contaminants in water [2-7]. Fluorescence, which is revealed to be the more sensi- tive tool compared with UV-vis absorbance, has been used in many scientific fields, such as chemistry, medicine, en- vironmental and food science [8]. Various fluorescence spectroscopy techniques have been used to characterize water samples, including emission scan at fixed excitation wavelengths, synchronous scan at a constant offset wave- length between excitation and emission wavelengths, and excitation-emission matrices (EEMs). Recently, EEMs spec- troscopy has successfully applied for the investigation of dissolved organic matter (DOM) in different type of soils [9-11] and water bodies, including sea [12], lake [13], river [14], wastewater [15, 16], etc. EEMs spectroscopy contains a substantial amount of information on DOM composition and structure. Each EEM demonstrates a specific combi- nation of fluorescence intensities over a range of excitation and emission wavelengths. The traditional peak-picking technique, as a kind of EEMs analysis tool, was usually used to identify particular fluorophores independently. Coble [17] classified fluorescence regions into peak A (UV humic-like), peak C (visible humic-like), peak B (tyrosine- like), and peak T (tryptophan-like). Chen [18] divided EEMs spectra into five excitation-emission regions based on fluo- rescence of model compounds, such as aromatic protein- and humic-like substances, and bacterial- soluble microbial products. However, EEMs spectroscopy commonly reflects the complex mixtures rather than simple purified compounds in water samples, which results in the overlap of individ- ual fluorophores [19]. Therefore, the common peak-picking