w a t e r r e s e a r c h 6 8 ( 2 0 1 5 ) 4 8 7 e4 9 7

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevier.com/locate/watres

Relationships between molecular weight and fluorescence properties for size-fractionated dissolved organic matter from fresh and aged sources C.W. Cuss a, C. Gueguen b,* a b

Environmental and Life Sciences Graduate Program, Trent University, ON, Canada Chemistry Department, Trent University, ON K9J 7B8, Canada

article info

abstract

Article history:

Relationships between the molecular weight (MW) and fluorescence properties of dissolved

Received 31 July 2014

organic matter (DOM) are important considerations for studies seeking to connect these

Received in revised form

properties to water treatment processes. Relationships between the size and fluorescence

2 October 2014

properties of nine allochthonous DOM sources (i.e. leaf leachates, grass, and headwaters)

Accepted 6 October 2014

were measured using asymmetrical flow field-flow fractionation (AF4) with on-line

Available online 18 October 2014

absorbance and fluorescence detectors. Correlations between optical properties and MW were readily apparent using parallel factor analysis (PARAFAC) coupled to self-organizing

Keywords:

maps (SOM): protein/polyphenol-like fluorescence (peaks B and T) was highest at lower

PARAFAC-EEM

molecular weights (1 kDa). Proportions of peaks B, T, and A þ C were significantly

(MWD)

correlated with MW (p < 0.001). The first principal component (PC1, 42% of variation in

Self-organizing maps (SOM)

fluorescence properties) was a significant predictor of sample MW (R2 ¼ 0.63, p < 0.05),

Asymmetrical flow field-flow frac-

while scores on PC2 (27% of total variance) traced a source-based gradient from deciduous

tionation (AF4)

leachates/headwaters through to coniferous leachates/headwaters. PC3 (13% of var.) was

Supramolecular assemblies

also correlated with MW (p < 0.005). A secondary peak in peak T fluorescence was associated with larger size fractions in aged sources, and scores on PC1 also traced a path from the leachates of fresher leaves, through more humified leaves, to headwaters. Findings are consistent with the hypothesis that the structure of aged DOM arises through supramolecular assembly. © 2014 Elsevier Ltd. All rights reserved.

1.

Introduction

Dissolved organic matter (DOM) is a complex, heterogeneous, and polymorphous mixture in all natural waters that

* Corresponding author. Tel.: þ1 705 748 1011; fax: þ1 705 748 1625. guen). E-mail address: [email protected] (C. Gue http://dx.doi.org/10.1016/j.watres.2014.10.013 0043-1354/© 2014 Elsevier Ltd. All rights reserved.

functions in several important roles, such as: a primary source of nutrients and energy for microorganisms (Kirchman, 2003), binding to heavy metals and pollutants to control their  guen and Dominik, 2003), and toxicity, transport, and fate (Gue generating carcinogenic disinfection by-products (DBP) during

488

w a t e r r e s e a r c h 6 8 ( 2 0 1 5 ) 4 8 7 e4 9 7

water treatment (Reckhow et al., 2004; Beggs and Summers, 2011; Hur et al., 2012). The application of chemometric methods such as parallel factor analysis (PARAFAC) (Stedmon et al., 2003) and selforganizing maps (SOM) (Kohonen, 2001; Bieroza et al., 2012a) to fluorescence excitation-emission matrices (EEM) (Coble, 1996) has allowed the connection of fluorescence properties to the effectiveness of DOM as: a source of nutrients and energy for microorganisms (Wickland et al., 2007; Fellman et al., guen, 2012a), a metal-binding agent (Cuss 2008; Cuss and Gue  guen, 2012a; Yamashita and Jaffe , 2008), a fingerand Gue print for DOM origins and end members (Stedmon and guen, 2013), a tracer for the Markager, 2005a; Cuss and Gue production and degradation of autochthonous DOM (Stedmon and Markager, 2005b), and a generator of DBP (Beggs and Summers, 2011; Bieroza et al., 2012a; Pifer and Fairey, 2012). The size properties of DOM have also been connected to its source (McElmurry et al., 2013), its effectiveness as a binding agent for metals (Benedetti et al., 2002; Wu et al., 2004; Chen et al., 2013), and its DBP formation potential (Hur et al., 2012). Thus, both the fluorescence and size properties of DOM are related to its functionality; however, little is known about relationships between its size and fluorescence properties. The existence of such relationships seems likely given that the absorbance properties of DOM are frequently used as a proxy for molecular weight, fluorescent DOM is a subset of absorbing DOM, and fluorescence is a more sensitive analytical method. Fluorescence also potentially offers a more finely-grained distinction of DOM constituents via the separation of fluorescence EEMs using PARAFAC; however, the definitive attribution of underlying chemical species to fluorescence regions remains a challenge (Aiken, 2014). Alternatively, the use of absorbance-based proxies for molecular weight is based on a greater proportion of underlying chemical species (i.e. all absorbing species) which potentially offers more information, but less resolving power compared to EEMs-PARAFAC. However, there are absorbance- and fluorescence-based proxies that are based on similar chemical properties that have been associated with molecular weight. For example, there is a correlation between molecular weight and absorption at longer wavelengths based on aromaticity (Chin et al., 1994; Peuravuori and Pihlaja, 1997), which has been related to molecular weight via the spectral slope (Helms guen and Cuss, 2011). At the same time, et al., 2008; Gue increasing signal in the 'humic-like' fluorescence region has been associated with increases in relatively aromatic, humic materials and shifts to longer wavelengths via the humification index (HIX; Zsolnay et al., 1999). Correlations between HIX and the spectral slope, and between HIX and both molecular weight and specific UV absorbance (SUVA) have also been noted, where SUVA is a strong indicator of aromaticity (Weishaar et al., 2003; Hur and Kim, 2009; Chen et al., 2011). Tacit indications of relationships between the size and fluorescence properties of DOM have been suggested by studies using ultrafiltration (UF) (Liu et al., 2007; Huguet et al., guen et al., 2013), size2010; Caron and Smith, 2011; Gue exclusion chromatography (SEC) (Her et al., 2003; Maie et al.,  guen and 2007; Romera-Castillo et al., 2014), and AF4 (Gue

 guen, 2012b, 2013). However, efforts Cuss, 2011; Cuss and Gue have focussed upon relatively few discrete fractions or samples, single fluorophores and wavelength pairs, or a narrow range of sizes. Assessments of relationships have also been primarily descriptive, owing in part to the lack of available chemometric methods. Consequently, knowledge about relationships between the size and fluorescence properties of DOM, and how these relationships may change during biogeochemical cycling, remains limited. Bioavailability has been positively correlated to the proportion of peak B and peak T fluorescence in DOM (Wickland guen, 2012a; et al., 2007; Fellman et al., 2008; Cuss and Gue Balcarczyk et al., 2009), and it has also been suggested that the size/structure of larger, more recalcitrant, humified, and aromatic DOM in part arises from molecular associations that follow the biodegradation of labile, relatively small constituents (Wershaw, 2004; Sutton and Sposito, 2005; Wickland et al., 2007; Hur et al., 2009; Cory and Kaplan, 2012). Indeed, the results of a recent study using size exclusion chromatography and four freshwater samples suggest that the structural and fluorescence characteristics of DOM may be controlled by such molecular assemblies, and emphasize the importance of assessing the optical properties of DOM over the size continuum for understanding its processing and structure (RomeraCastillo et al., 2014). The possibility of such controls on the structure of biologically processed DOM, with a corresponding relationship between size and optical properties, is of interest in drinking water treatment systems that incorporate biological processing. In particular, it has been shown that DOM leached from different leaf species has different DBP formation potentials that are modified by microbial processing (Reckhow et al., 2004; Chow et al., 2009; Pellerin et al., 2010; Beggs and Summers, 2011; Hur et al., 2012). In the present study, relationships between the size and fluorescence properties of nine DOM sources that span a range of humification were investigated using AF4 coupled to absorbance and fluorescence detectors: three headwaters, leachates from the senescent leaves of five tree species, and the partially humified leaves of one species. Results were effectively visualized by combining PARAFAC and SOM. Principal component analysis (PCA) was used to extract latent variables from the fluorescence composition of size fractions. Latent variables were in turn used to predict sample molecular weights using only the fluorescence properties thereof.

2.

Materials and methods

2.1.

DOM samples

A variety of leachates and freshwater samples were chosen in order to span the ranges of molecular weights, optical properties, and degree of humification typically encountered in headwaters (Table 1). Fresh senescent leaves or needles were gathered from the ground beneath, or shaken from the branches of three different tree species: A. rubrum (senescent red maple gathered from ground, source SRM), A. saccharinium (silver maple, shaken from branches; SSM), and P. glauca

w a t e r r e s e a r c h 6 8 ( 2 0 1 5 ) 4 8 7 e4 9 7

(white spruce, shaken from branches; SWS). Senescent and browned reed canarygrass (P. arundinacea; SRC) was also leached, as were overwintered leaves collected from beneath an A. saccharinium after being exposed to freezeethaw cycles over winter, and leached by spring melt (source OSM). Finally, a leaf sample (STL) was collected by stripping senescent needles from a Tamarack (L. laricina). The procedures used for storing and leaching leaves followed those outlined in earlier  guen, 2012a, 2013). studies (Cuss and Gue Headwater samples were collected from three locations just after the peak of spring freshet (May/June 2013): seasonal meltwater at the perimeter of a cattail marsh dominated by cattails, tussock sedge, and reed canarygrass (source WW; Peterborough, ON., 44.34 N, 78.36 W), and two vernal headwaters with sub-catchments dominated by deciduous (DHW) and coniferous (CHW) vegetation (Dorset, ON., 45.14 N, 79.092 W and 45.14 N, 79.090 W, respectively) (Figure S1). The coniferous site was heavily influenced by grass immediately upstream of the sampling point (site photographs are shown in Figure S1). Headwater samples were taken just after the peak of spring freshet, so that the primary source was snowmelt running across the surface of the soil. The DOM content of leachates and headwater samples was isolated by vacuum filtration using combusted 0.45-mm glassfibre filters, and filter-sterilized using rinsed, 0.2-mm polycarbonate filters (Millipore). The resulting DOM samples were refrigerated at 4  C in combusted (450  C for 5 h) amber-glass vials prior to analysis, which was completed within seven days of leaching/collection.

2.2.

AF4-DAD-EEM

The methods and apparatus used for on-line size fractionation with UVeVisible and fluorescence detection are outlined guen and Cuss, 2011). Two in detail in an earlier study (Gue millilitres of each sample was injected into the 0.3-mL sample loop of an asymmetrical flow field-flow fractionator (AF2000 Focus; Postnova Analytics), equipped with a polyethersulfone membrane (300-Da molecular weight cut-off; Postnova), and

Table 1 e Peak-maximum molecular weights (Mp) and origins of DOM sources. Source

Origin

Senescent white spruce (SWS) Tree branches Senescent red maple (SRM) Ground under tree (in autumn) Senescent silver maple (SSM) Shaken from tree Senescent Tree branches tamarack/larch (STL) Over-wintered silver Ground under tree maple (OSM) (in spring) Wetland (WET) Spring headwater Deciduous-dominated Spring headwater headwater (DHW) Coniferous-dominated Spring headwater headwater (CHW) Senescent reed Senescent/dying canarygrass (SRC) grass (in autumn)

Mean Mp ± 95% CI (kDa) 0.42 ± 0.03 0.50 ± 0.04 0.48 ± 0.03 0.43 ± 0.03 1.05 ± 0.04 1.12 ± 0.04 1.48 ± 0.13 1.71 ± 0.06 2.06 ± 0.12

489

coupled to on-line diode array and fluorescence detectors (Shimadzu SPD-M20A and Agilent 1200 series FLD model G1321A, respectively). In this study a lower cross flow rate was applied to achieve higher resolution in the low molecular weight range, necessitating changes in other flow rates to optimize focussing position and maintain channel pressure at 1 kDa, as assessed by ultrafiltration (Osborne et al., 2007). The MWD of leachates with higher Mp were also skewed further towards the higher MW range, indicating that a greater proportion of absorbing DOM was in larger size fractions. Specific excitation/emission wavelength maxima and full EEMs of the six components detected by PARAFAC are shown in Table 2 and Figure S5, respectively. Six EEMs measured at

w a t e r r e s e a r c h 6 8 ( 2 0 1 5 ) 4 8 7 e4 9 7

Table 2 e Excitation and emission wavelengths (in nm) of primary (and secondary) maxima of PARAFAC components. Component C1 C2 C3 C4 C5 C6 C7

Excitation

Emission

Peak