Aquatic Organic Matter Fluorescence
109
3.5.2 Drinking Water Fluorescence
In comparison to wastewaters, the application of fluorescence spectroscopy for the detection and monitoring of organic matter (OM) in drinking water sources and water distribution and water systems is an emerging discipline. Organic matter, both dissolved
(DOM) and natural (NOM), is ubiquitous in all waters that are used to supply drinking
water systems (Matilainen et al., 2011). Even so, the recent advances of fluorescencebased techniques are sufficiently mature so as to be considered an alternative approach
to more conventional methods for the characterization of DOM and/or NOM in drinking
water systems. In a recent review by Matilainen et al. (2011), a comprehensive overview of current methods that are used to characterize NOM in relation to drinking water
treatment, including fluorescence, is provided. Early work (Rosario-Ortiz et al., 2007)
applied the use of EEMs as a tool for characterizing OM (inclusive of DOM or NOM)
in drinking water sources. DOM characterization in drinking water sources is important, as it is known that DOM contributes to the formation of disinfection by-products
(DBP), and therefore affects how water treatment facilities are optimized. The potential
for DOM in drinking water sources to generate DBPs was investigated by Marhaba et al.
(2009). Researchers were able to predict trihalomethanes formation potential by applying
a principal component regression model to dimensionally reduced spectral fluorescent
signatures. Beggs et al. (2009) studied the relationships between fluorescence intensities
(total fluorescence intensities and fluorescence indexes), redox index, chlorine demand,
and DBP formation during chlorination. This study used a PARAFAC model to extract 13
components (fluorophores or groups of fluorophores) from fluorescence EEMs. Quinonelike components were found to be strongly correlated to DBP formation. Johnstone and
Miller (2009) investigated the correlation of water quality characteristics of Iowa River
water and associated isolated fractions to the formation of DBP, specifically trihalomethanes and haloacetic acids, subsequent to chlorination using multifactor linear regression.
In this work, defined regions within fluorescence EEMs were identified using fluorescence
regional integration (Chen et al., 2003) and the changes in the fluorescence intensities
of these identified regions, in conjunction with chlorine consumption, were reported to
correlate to the formation of specific DBPs. This work was further developed in a study
in which a three component PARAFAC model was used to assess drinking water DBP
formation (Johnstone et al., 2009). This PARAFC model correlated with DOC, chlorine
consumption, and individual DBP formation potential. Interestingly, the multifactor linear
regression of selected component scores showed linear relationships to individual DBPs.
According to the researchers, the specificity of this approach makes the prediction of DBP
formation (DBP formation potential) possible.
A recent review by Henderson et al. (2009) concludes that the sensitive detection of
contamination events in recycled water systems may be achieved by monitoring peak T
and/or peak C fluorescence. Hambly et al. (2010) also examined the application of fluorescence spectroscopy as a monitoring tool in recycled water treatment plants and dual (recycled and drinking water) distribution systems. This work detected a 10-fold difference in
109
3.5.2 Drinking Water Fluorescence
In comparison to wastewaters, the application of fluorescence spectroscopy for the detection and monitoring of organic matter (OM) in drinking water sources and water distribution and water systems is an emerging discipline. Organic matter, both dissolved
(DOM) and natural (NOM), is ubiquitous in all waters that are used to supply drinking
water systems (Matilainen et al., 2011). Even so, the recent advances of fluorescencebased techniques are sufficiently mature so as to be considered an alternative approach
to more conventional methods for the characterization of DOM and/or NOM in drinking
water systems. In a recent review by Matilainen et al. (2011), a comprehensive overview of current methods that are used to characterize NOM in relation to drinking water
treatment, including fluorescence, is provided. Early work (Rosario-Ortiz et al., 2007)
applied the use of EEMs as a tool for characterizing OM (inclusive of DOM or NOM)
in drinking water sources. DOM characterization in drinking water sources is important, as it is known that DOM contributes to the formation of disinfection by-products
(DBP), and therefore affects how water treatment facilities are optimized. The potential
for DOM in drinking water sources to generate DBPs was investigated by Marhaba et al.
(2009). Researchers were able to predict trihalomethanes formation potential by applying
a principal component regression model to dimensionally reduced spectral fluorescent
signatures. Beggs et al. (2009) studied the relationships between fluorescence intensities
(total fluorescence intensities and fluorescence indexes), redox index, chlorine demand,
and DBP formation during chlorination. This study used a PARAFAC model to extract 13
components (fluorophores or groups of fluorophores) from fluorescence EEMs. Quinonelike components were found to be strongly correlated to DBP formation. Johnstone and
Miller (2009) investigated the correlation of water quality characteristics of Iowa River
water and associated isolated fractions to the formation of DBP, specifically trihalomethanes and haloacetic acids, subsequent to chlorination using multifactor linear regression.
In this work, defined regions within fluorescence EEMs were identified using fluorescence
regional integration (Chen et al., 2003) and the changes in the fluorescence intensities
of these identified regions, in conjunction with chlorine consumption, were reported to
correlate to the formation of specific DBPs. This work was further developed in a study
in which a three component PARAFAC model was used to assess drinking water DBP
formation (Johnstone et al., 2009). This PARAFC model correlated with DOC, chlorine
consumption, and individual DBP formation potential. Interestingly, the multifactor linear
regression of selected component scores showed linear relationships to individual DBPs.
According to the researchers, the specificity of this approach makes the prediction of DBP
formation (DBP formation potential) possible.
A recent review by Henderson et al. (2009) concludes that the sensitive detection of
contamination events in recycled water systems may be achieved by monitoring peak T
and/or peak C fluorescence. Hambly et al. (2010) also examined the application of fluorescence spectroscopy as a monitoring tool in recycled water treatment plants and dual (recycled and drinking water) distribution systems. This work detected a 10-fold difference in