FLUORESCENCE TRACING OF DIFFUSE LANDFILL LEACHATE CONTAMINATION IN RIVERS ANDY BAKER
School of Geography, Earth and Environmental Sciences, The University of Birmingham, Edgbaston, Birmingham, B15 2TT,UK (t) +44 121 415 8133 (f) +44 121 414 5528 (e)
[email protected]
1
ABSTRACT
Landfill leachates are composed of a complex mixture of degradation products which include a wide range of potentially fluorescent organic molecules and compounds. Here we investigate the use of fluorescence excitation-emission matrix (EEM) analysis in detecting diffuse landfill leachate contamination in rivers. Landfill leachates from three unlined landfill sites adjacent to our study river are characterised by intense fluorescence at excitation wavelength 220-230 nm, and emission wavelength 340-370 nm, which derives from fluorescent components of the xenobiotic organic matter fraction. Seven surface water sample sites on an adjacent polluted river system were analysed for fluorescence and water quality properties. The 220-230 nm excitation wavelength, 340-370 nm emission wavelength fluorescent centre was also detected in this river system at the sample locations downstream of the landfills, but not at upstream control sites, demonstrating its use as a tracer of landfill leachate contamination. Negative correlations are observed between this fluorescence centre and dissolved oxygen in the river water samples, demonstrating the water quality implications of leachate contamination at this study site. The fluorescence intensity at the 220-230 nm excitation wavelength, 340-370 nm emission wavelength fluorescent centre in landfill leachates is such that it remains detectable at dilutions of 102-103, and the fluorescence EEM technique is rapid and cost-effective for use by river managers and water quality regulators.
2
Keywords:
fluorescence, excitation-emission matrix, landfill leachate, river
pollution
3
INTRODUCTION
Landfill leachate is formed by excess rainfall percolating through waste layers in a landfill site. Leachate characteristics are a function of the quality of the landfill input, which determines the relative importance of dissolved organic matter (DOM), xenobiotic organic matter (XOM), inorganic macrocomponents, and heavy metals, as well as the landfilling technology used and the age of the leachate (Christensen et al., 2001). Within the organic matter fraction, Kang et al. (2002) demonstrate an increase in leachate humification with age and Calace et al. (2001) demonstrate that old leachate has a wider range of molecular weight and with a high molecular weight fraction being present. Schwarzbauer et al. (2002) identified plant material derived compounds, degradation products of peptides, carbohydrates and lignin, and numerous XOMs from pharmaceuticals, plasticisers, pesticides and chlorinated aromatics in contained and leaked leachate.
Landfill leachate can be a significant environmental contaminant if it leaves the landfill site, which can occur if the site is designed as a ‘dilute and disperse’ site, if the landfill is unlined (for example, at landfills in relatively impermeable clay soils, or in developing countries where unlined landfilling is still regular practice) or if the landfill is lined but the lining fails. Typically, leachate that leaves a landfill site will enter the ground water as a contaminant plume, where it is then diluted and also transformed by a variety of processes such as sorption, chemical precipitation and microbial degradation. Rivers may also be impacted by landfill leachate; this may occur due to runoff directly from uncovered landfills (Zafar and Alappat, 2004), due 4
to contaminated groundwater inputs into the river banks or bed (Atekwana and Krishnamurthy, in press), or erosion of bankside landfills (Kao et al., 2003). The pollution impacts of landfill leachate are varied, depending on the nature of the fill material, the age of the landfill, (if in a plume then) the extent of any attenuation processes, and the concentration of leachate. Many researchers have suggested that the most significant long-term pollutant from landfills is ammonia (Christensen et al., 2001; Barlez et al., 2002), which can lead to toxicity in surface water at concentrations 102-105 lower than in landfill leachate. Eutrophication from phosphorus rich run-off from exposed landfill surfaces has also been suggested to be a surface water quality issue (Hodgkiss and Lu, 2004). The XOM fraction in the leachate may also reach toxic levels, with BTEX compounds, PAHs, and pesticides commonly reported in leachates (Baun et al., 2003; Schwarzbauer et al., 2002) as well as compounds with xeno-estrogenic properties (Wintgens et al., 2003; Coors et al., 2003), although the toxicity of XOM in leachate has received little attention (Christensen et al., 2001). Landfill leachate induced toxicity has recently been reported in mice (Bakare et al., 2003). Given the potential water quality issues arising from landfill leachate contamination, the identification of a suitable tracer for leaked landfill leachate has been the focus of many research studies. These studies have in general focused on ground water contaminant plumes. For example, using GC/MS Schwarzbauer et al (2002) identified 184 organic compounds (both natural and xenobiotic) from two landfill leachates in landfills, but only 92 organic compounds in leakage water. Of the latter, several XOMs appeared both persistent and specific to the landfill leachates (e. g.
N,N-diethyloluamide
(DEET),
N-butylbenzenesulfonamide
(NBBS)
and
propyphenazone) and therefore could be used as tracers of landfill leachate using 5
GC/MS. Van Breukelen et al (2003) used a combination of biogeochemical and isotope tracers to both identify the ground water leachate plume extent as well as interpret the redox conditions within it. Vilomet et al (2001; 2003) demonstrated that leachate contaminant plumes can be detected through both stable lead (206Pb/207Pb) and strontium (87Sr/86Sr) isotope ratios. In river systems, Atekwana and Kirshnamurthy (in press) investigated the use of
13
C isotopes as a tracer of leachate
contaminated groundwater seeping into a headwater stream, but although leachate 13C was elevated (δ13C of -2.3 to +5.7 per mil in seepage groundwaters) with respect to natural
13
C (δ13C of -16.9 to -10.0 per mil in upstream river samples), grab samples
failed to detect the presence of leachate pollution in the stream itself due to dilution effects. Ammonia might also be considered to be a useful tracer of landfill leachate in surface waters, with typical quoted leachate concentrations (50-2200 mg/l; Chistensen et al., 2001) more likely to be detectable after dilution, but oxidation and the presence of multiple ammonia contaminant sources in many river environments limits its use as a unique fingerprint.
Most recently, fluorescence has been used to fingerprint landfill leachates within three landfill sites (Baker and Curry, 2004). Rapid fluorescence analysis is now possible using fluorescence spectrophotometers such that an excitation emission matrix (or EEM) can be generated in approximately 1 minute, with analysis possible in situ with portable spectrometers (Baker et al., 2004). The technique is now widely used in wastewater characterisation within the wastewater treatment process (Reynolds and Ahmad, 1997; Westerhoff et al., 2001; Vasel and Praet, 2002), for DOM characterisation in marine and estuarine waters (Coble et al., 1990; Mopper and Schultz, 1993; Mayer et al., 1999; Parlanti et al., 2001; Stedmon et al., 2003), and
6
riverine DOM fingerprinting (Baker, 2001, 2002a,b,c; Baker et a;. 2003; McKnight et al., (2001, 2003); Thoss et al., (2000); Newson et al., (2001), Yan et al. (2000) and Stedmon et al., (2003)). Baker and Curry (2004) demonstrated that landfill fluorescence properties are all characterized by intense fluorescence at λex = 220-230 nm, and λem = 340-370 nm that was suggested to derive from fluorescent components of the XOM fraction such as naphthalene (the “XOM peak”). A second fluorescence centre, at λex = 320-360 nm, and λem = 400-470 nm, derives from a higher molecular weight fulvic-like fraction, and a third centre at λex = 270-280 nm, and λem = 340-360 nm to tryptophan-like fluorescence. Baker and Curry (2004) demonstrated that leachates from different landfill sites could be statistically discriminated using a combination of fluorescence and basic geochemical parameters. For the XOM fluorescence peak, fluorescence could be detected at ~ 103 dilution, suggesting that this technique could be appropriate for leachate detection in river systems.
Previous
research
therefore
suggests
that
fluorescence
EEM
spectrophotometry is able to fingerprint landfill leachate contamination in surface waters. However, to date no investigations have been undertaken to test this hypothesis on rivers that suffer from landfill leachate contamination. Therefore, results are presented of the analysis of fluorescence EEM spectrophotometric analysis of landfill leachate samples from potentially polluting landfill sites, as well as river water samples from an adjacent contaminated river in NE England.
7
MATERIALS AND METHODS
A site was investigated where it was thought that landfill leachate contamination was having an impact on a local watercourse. Therefore both the potential source material (the fluorescence properties of landfill leachates) as well as river water adjacent to the landfills was sampled. The site, including sampling locations for both leachate samples and river waters is presented in Figure 1.
Three unlined landfills are situated in former clay pits adjacent to the Rowletch Burn, a tributary of the River Team, part of the Tyne catchment in NE England (lat: 54° 53.37N, long: 1° 35:18W). The burn is a small ungauged stream of ~ 4 km length, heavily channelised and in places culverted, of channel width 1-2 m, which drains a mixed land use of agriculture, industry and housing. Water quality of the Rowletch Burn monitored monthly over the period 1990 to 2002 at site 5 (Figure 1) by the England and Wales water quality regulator the Environment Agency demonstrates that the quality is poor, with biochemical oxygen demand averaging 12.3±18.7 mg/l, ammonia 7.4±8.6 mg/l and dissolved oxygen 47±15% saturation (1 sd; n=132). Given the high ammonia concentrations observed in the burn, the adjacent landfills had been identified as probable sources of the poor water quality, but Environment Agency surveys failed to detect point pollution sources to the river. Diffuse source contamination through the stream banks or bed was therefore assumed to be occurring, but it was not known which of the three landfill sites was the possible source.
8
The Rowletch Burn passes over a geology that comprises thick laminated glacial clays that were deposited in a buried valley cut into the underlying Coal Measures. The laminated clays are overlain locally by more variable silty clays, which contain sands in places, as well as areas of made ground. The laminated clays, which are extensive both laterally and vertically, are of low permeability and were viewed to form a non-aquifer and therefore suitable for landfilling. Two landfills, Ouston Quarry and North Quarry (Figure 2), are disused and details about their history are sparse. Single boreholes are present in each landfill, permitting leachate sampling although not a spatial coverage of leachate quality variations within the former landfills. Four leachate samples were collected from each borehole over the sampling period. A third landfill, St Bede’s, had just ceased operation, and samples were collected just prior to closure. Due to its commencement age and its location in thick clay deposits, the site is also unlined. The landfill had an initial void space of 2,223,607m3; between 1970 and 1975 the northern half of the landfill was filled with a high proportion of incinerator residues together with some domestic and industrial wastes. The southern half of the void has been landfilled with domestic, commercial and industrial waste, incorporating a proportion of special wastes between 1993 and 2003. The northern area has also received additional waste over the older landfill to produce a domed restoration landform. Landfilling operations ceased during March 2003 and leachate from the site is discharged to sewer. Nine samples covering both old and active sectors of the landfill were sampled once in September 2002.
River water samples were collected on the Rowletch Burn at seven sampling locations, with sample locations limited by public access points (Figure 1). Sampling occurred in May and June 2002; initial weekly sampling (May) was followed by
9
biweekly sampling (June). Six samples at each site were collected over the study period. Sites 1 and 6 are of tributaries to the Rowletch Burn that are upstream of any landfill sites. Sites 2, 3, 4, 5 and 7 are progressively downstream on the Rowletch Burn. Site 2 is at the upstream corner of Ouston Quarry, site 3 adjacent the mid point of Ouston Quarry, and site 4 downstream of Ouston Quarry and upstream of St Bede’s. Site 5 is downstream of St Bede’s and upstream of North Quarry and site 7 is downstream of North Quarry. Over the length of the river sampled, no other tributary inputs were present, although several surface-water drains are present as well as combined sewage overflows.
Water samples were collected in 50 ml plastic containers that had been previously cleaned by soaking in 10% HCl for 24 hours and then triple rinsed in distilled, deionised water. Samples were returned to laboratory and analysed within 24 hours for spectrophotometric and geochemical properties. Samples were filtered using Whatman GF/C glass microfibre filter papers that had been previously heated to 450 Celsius to remove any possible organic contamination. Fluorescence EEMs were generated using a Perkin Elmer LS50B Luminescence Spectrophotometer as described elsewhere (Baker, 2001). The only modification was to scan a wider range of wavelengths with excitation from 200 to 370 nm and emission detected from 250 to 500 nm. The majority of samples were analysed at least x10 dilution due to their high fluorescence intensity. The Raman peak intensity of water at 348 nm over the analysis period averaged 23.6±0.9 arbitrary units (n=30); this value can be used to permit inter-laboratory comparison. Absorbance was measured using a WPA Lightwave UV/VIS spectrophotometer; absorbance was measured at 254, 340 and 410 nm. As well as providing water quality information, the measurement of absorbance permits
10
the inner-filtering correction of fluorescence intensities, although this was not required as our diluted samples all fell below the necessary absorbance (Ohno, 2002). Ammonia was measured using a Hanna Instruments Colorimeter and chloride concentration by silver nitrate titration. Total organic carbon was measured using a Shimadzu TOC analyser; measurements were made in triplicate and the mean value recorded. pH, dissolved oxygen and oxidation-reduction potential were measured in the field.
RESULTS AND DISCUSSION
Results are presented in Table 1 for both river water and leachate samples, and typical fluorescence EEMs are presented in Figure 2. Table 1 demonstrates that landfill leachates are characterised by high absorbance, high ammonia concentration, high total organic carbon concentrations, low dissolved oxygen and highly reduced condition as would be expected from such samples and described elsewhere (Christensen et al., 2001). River water samples downstream of the landfills also show high ammonia and low dissolved oxygen, indicative of poor water quality. In contrast, upstream sites 1 and 6 have low ammonia and high dissolved oxygen and are of good water quality. Other geochemical (pH, chloride, organic carbon) and absorbance data at the river sample sites show no significant differences between upstream and downstream sites.
Typical fluorescence EEM results for leachate samples (Figure 2a,b) demonstrate the distinctive and intense fluorescence peak at λex = 220-230 nm, and
11
λem = 340-370 nm, which we will call here the XOM peak, which is identical in location to that observed previously by Baker and Curry (2004). Other peaks clearly present include one at λex = 230-250 nm and λem = 400-440 nm that is attributed to humic material (Coble, 1996; Yan et al., 2000), and at λex = 320-340 nm and λem = 400-440 nm, a peak that is attributed to aromatic and aliphatic groups in the fulviclike fraction. A final fluorescence centre that is weakly visible as a shoulder in the other fluoresce peaks and with variable intensity is at λex = 275-280 nm and λem = 350-360 nm, attributed to the protein tryptophan, and widely observed in polluted river waters (Baker, 2001; 2002b). At the sites where replicate samples were taken (Ouston and North Quarries), the fluorescence EEMs at each site are similar in terms of the presence and intensity of the fluorescence centres present (including the presence of the XOM peak; Table 1). At St Bede’s landfill, where spatial sampling was possible, a wider range of variability of fluorescence properties is visible, suggesting that a similar variability might be expected at Ouston and North Quarry landfills had sampling of spatial variability within each landfill been possible. Differences also occur in the relative intensity of the fluorescence centres between the sites, probably relating to differences in leachate age and quality as demonstrated by Baker and Curry (2004).
Figure 2(c,d) presents the typical fluorescence EEMs of the diluted river water samples, both upstream and downstream of the landfill sites. The downstream sample (Figure 2c) is dominated by the XOM peak, although at a lower concentration than in the leachates. The upstream sample (Figure 2d) is similar to fluorescence EEMs of clean river waters, with only the presence of fluorescence centres at λex = 230-250 nm and λem = 400-440 nm, and at λex = 340 nm and λem = 400-440 nm, indicative of 12
natural dissolved organic matter. Total organic carbon data (Table 1) demonstrates no difference between TOC in upstream and downstream river water samples, irrespective of the presence of fluorescence XOM peak downstream sites, confirming that this fluorescence centre derives from a low concentration, high fluorescence efficiency source.
Visual inspection of Table 1 reveals differences between sample sites upstream and downstream of the landfills. Correlations between the fluorescence and geochemical results for all the river water samples (n= 40 to 42) are presented in Table 2. Strong correlations exists between the intensity of all the fluorescence centres (0.65
