Great lakes health
Real-time Flurometric Assay for Sewage Presence: A Cost-effective Method to Determine Potential Water Quality Threats to Swimmers and Ecosystem Health
Abstract
Bacteria from both animal and human sources contaminate bodies of water, but the health risk is higher from human sewage. In testing for human contamination, biological assays require at least 18 hours for cultures of E. coli or enterococci to develop. A faster method involves the utilization of chemical markers, such as fluorescent whitening agents (found in laundry detergents), which are part of sewage discharge. The detection and quantification of the FWAs have been investigated in laboratory-prepared solutions, Lake Michigan water samples using optimized fluorescence excitation-emission matrix characterizations and HPLC equipped with fluorescence detection. These water samples have also been analyzed for the presence of E coli using the membrane filtration method. The Lake Michigan water samples were collected from 3 different sites on the southern shore of the lake in Northwest Indiana: Burns Ditch and two nearby beach locations (Ogden Dunes Beach). The river drains three counties that include five wastewater treatment plants with occasional sewer overflows and a wide variety of non-point sources. E. coli levels in Burns Ditch are generally higher than the adjacent lake (beach). The Burns Ditch outfall directly impacts the Ogden Dunes Beach particularly during north to northeast winds. HPLC chromatograms of secondary treated wastewater effluent and lake water samples revealed the presence of FWA compounds.
Introduction
The majority of wastewater treatment plants (~70%) in the United States utilize primary and secondary treatment facilities to process wastewater, the minimum requirement established by the Clean Water Act.6 While an effective primary/secondary treatment plant can result in the removal of up to 90% of dissolved organics, the system relies on nature to remediate the remaining contaminants. In many cases the treated water is discharged into a natural body of water. A 2004 U.S. EPA report states that an estimated 850 billion gallons of sewage water from 770 treatment plants is released to natural bodies of water each year.7 The treated wastewater from most municipal facilities still contains, at minimum, 10% of the dissolved organics, with a significant amount originating from household laundry practices.
Fluorescent whitening agents (FWAs) are part of laundry detergent formulations. Commercial whitening agents are either of the diaminestilbene (DAS) or the distyrylbiphenyl (DSBP) variety. Several commercial variants of the two structures are used. An estimated 90% of all whitening agents contain either the DAS or DSBP structures.(ref)
The presence of these FWAs, in natural bodies of water, i.e. Lake Michigan, may be linked to human bacterial contamination originating predominantly from wastewater treatment plant discharge. While bacteria from both animal and human sources contaminate natural bodies of water, the health risk is substantially higher from human sources. In the typical testing process for bacterial contamination, biological assays are utilized and require at least 18 hours for cultures of Escherichia coli (E. coli) or enterococci to develop. A proposed, faster method for the detection of human contamination involves the utilization of chemical markers, such as fluorescent whitening agents, which are part of sewage discharge.
We have used a combination of high resolution fluorescence spectroscopy and HPLC with fluorescence detection to determine the magnitude and distribution of fluorescent whitening agents along a southern portion of the Lake Michigan shoreline.
Experimental
Water Samples: The Lake Michigan water samples were collected in dark glass containers at 3 different sites on the southern shore of the lake in Northwest Indiana (Figure 2): Burns Ditch and two nearby beach locations at Ogden Dunes, referred to as BD, OD1 and OD3, respectively. The OD3 site is furthest from the Burns Ditch outflow. Local secondary treated wastewater was obtained from the Portage Reclamation Facility in Portage, IN, which discharges into Burns Ditch.All lake water and treated wastewater samples were filtered using 0.2 μm filter paper and utilized within three hours of the collection. All water samples analyzed for fluorescent whitening agents were prepared and/or stored in the dark to avoid extraneous photochemical reactions.
Chemicals: Keyfluor White CBS-X was donated by Keystone Chemicals. 4,4’-Diamino-2,2’-stilbenedisulfonic acid (DAS) was purchased from Aldrich Chemicals at 85% purity. The Fluorescent Brightener 28 (FB28) was purchased from Sigma. Suwannee River Fulvic Acid Standard II and NOM were purchased from IHSS (International Humic Substances Society).
Solid Phase Extraction: Solid phase extraction disks, 47 mm ENVI-DiskTM (C18 bonded) were purchased from Supelco and used with a vacuum filtration assembly. The disks were utilized as stated in the supplied procedure: 5 mL of methanol were poured through the filter followed by 5 mL of water and then the solution of interest (usually 1L). Finally, 10 mL of methanol (in two successive 5 ml aliquots) was used to recover the membrane extracted solute. If we assume 100% extraction efficiency, the concentration of the extracted solute in methanol should be 100 times that of the original solution.
FWA Analyses: Fluorescence measurements were performed using a Jobin-Yvon fluorimeter (Fluormax4). In all fluorescence measurements, the scans were recorded in steps of 1nm, integration times of 0.1s and 2 nm slitwidths for both excitation and emission monochromators. Excitation for the fluorescence measurements was at 350 nm, while the emission was monitored from 390-600 nm. A deionized water sample (solvent blank) free of any fluorescent impurity was recorded prior to every ten sample measurements. All fluorescence spectra from the natural water samples shown have been corrected for the water blank. Methanol was used as the blank for the SPE extracted samples and was similarly subtracted. For the excitation-emission matrix (EEM) measurements, the excitation wavelengths were incremented from 230-400 nm in 5nm steps, while the emission was scanned from 300-575 nm in 1 nm increments. Slit widths and integration times were the same as in the fluorescence measurements. Fluorescence quantum yield measurements were made using quinine sulfate in 0.1 M H2SO4 as the standard.
A Beckman System Gold HPLC, equipped with a Supelco Discovery® C18 column (5 μm, 250 mm x 4.6 mm) and a JASCO FP 1520 fluorescence detector, was used to separate and quantify the fluorescent compounds. The excitation/emission wavelengths for all samples were 350 nm and 430 nm. The mobile phase for the chromatography utilized two solvents: methanol and 0.1M (NH4)3PO4 or 0.1M NH4CH3CO2. The initial solvent flow consisted of 90% methanol and 10% aqueous ammonium salt. The flow changed to 10% methanol and 90% ammonium salt by 25 minutes and returned to the original mixture by 35 minutes.
Total dissolved organic carbon was measured using a Shimadzu Total Organic Carbon Analyzer, model TOC-5050 equipped with an ASI-500A autosampler. All water samples were acidified and sparged gently with nitrogen to remove the inorganic carbon. Measurements were carried out in triplicate. The reported values are within the experimental error of 5%.
Results and Discussion
The presence of fluorescent whitening agents in filtered Lake Michigan water samples and secondary treated wastewater samples was verified using HPLC with fluorescence detection, while the quantification of the FWAs in these samples was determined by direct fluorimetry. The total fluorescence emitted from these samples was compared to laboratory standard solutions. The Lake Michigan water samples were taken at and near Burns Ditch, which is the outfall of the Little Calumet River into southern Lake Michigan. The Burns Ditch outfall directly impacts the Ogden Dunes Beach particularly during north to northeast winds.
Direct water samples and concentrated water samples from solid phase extraction (SPE) filters were both tested using both fluorescence and HPLC methods. HPLC chromatograms of secondary treated wastewater effluent and lake water samples revealed the presence of mainly two different FWA compounds, Tinopal CBS-X and 4,4’-Diamino-2,2’-stilbenedisulfonic acid (DAS) with occasional evidence for the presence of the DAS analog, FB28. Figure 1 shows representative chromatograms of a) a standard containing the two main FWA compounds; b) a representative water sample from Burns Ditch, filtered and tested directly and c) the same sample as in b, but concentrated using the SPE technique into 10 ml of methanol. The chromatograms obtained for the field water samples in most cases exhibit only two peaks, which correspond very well with that of the standards CBS-X and DAS. The chromatograms of the extracted samples show similar peaks corresponding to DAS and CBS-X.
The HPLC data as seen in Figure 1a suggest a sizable difference in fluorescence efficiencies between the CBS-X and DAS. We have carried out fluorescence quantum yield measurements using quinine sulfate as the standard. These results confirm that the fluorescence efficiency of CBS-X is about 100 times that of DAS.
Figure 1: HPLC of FWAs a) standard containing DAS and CBS-X; b) filtered water sample from Burns Ditch c) same filtered sample but extracted using SPE.
The observed fluorescence from the standard samples can be expressed as follows: IfDAS= fDAS CDAS and IfCBS= fCBS CCBS where f represents the quantum yield.
The quantum yield ratio of 0.01 between DAS and CBS-X can then be used to determine the expected theoretical intensity ratio between the two chromatogram peaks in standard samples: 
.
We have run chromatograms of a number of standard samples ranging from 50 to 10 ppb of total FWA containing 4 times more DAS than CBS-X (in ppb amounts) and in each case the observed ratios in integrated areas between the two peaks is between 0.04 and 0.05.
Figure 2 shows representative fluorescence spectra for samples collected from Burns Ditch and from the neighboring Ogden Dunes beach site labeled OD1, both filtered and examined directly. In each case, we observe an emission maximum at ~430 nm. This maximum corresponds closely to that observed in the fluorescence of the standard samples, also shown in Figure 2, of CBS-X and DAS at concentrations which fluoresce to a similar degree to the lake water samples. The observed emissions from the field samples match well with those of the two standards, except that the spectra in the former case are generally broader.
Figure 2: Fluorescence spectra of the two standards: DAS and CBS-X and field samples.
We have shown that measurement of the total fluorescence at 440 nm can provide reasonable estimates of the amount of FWAs in natural waters and their dispersion along the Little Calumet river into Lake Michigan. To enable quantitation, we have used the data from the chromatograms determined the fluorescence of the natural water samples collected over two days in August 2008 and quantified the amount of FWAs present in these samples using calibration standards made up of one part DAS1 to 20 parts CBS-X.
The results are shown in Table 1 along with the corresponding TOC content of the samples.
Keywords
Fluorescence, fluorescent whitening agents, HPLC.
Lay Summary
The detection and quantification of the FWAs have been investigated in laboratory-prepared solutions, Lake Michigan water samples using optimized fluorescence excitation-emission matrix characterizations and HPLC equipped with fluorescence detection. These water samples have also been analyzed for the presence of E coli using the membrane filtration method. The Lake Michigan water samples were collected from 3 different sites on the southern shore of the lake in Northwest Indiana: Burns Ditch and two nearby beach locations (Ogden Dunes Beach). The river drains three counties that include five wastewater treatment plants with occasional sewer overflows and a wide variety of non-point sources. E. coli levels in Burns Ditch are generally higher than the adjacent lake (beach). The Burns Ditch outfall directly impacts the Ogden Dunes Beach particularly during north to northeast winds. HPLC chromatograms of secondary treated wastewater effluent and lake water samples revealed the presence of FWA compounds.
Media Coverage
One newsaper article is enclosed
Publications
Manuscript in preparation:“Distribution of Fluorescent Whitening Agents and Relation to Bacteria Levels in Southern Lake Michigan”
Julie Peller, K. Vinodgopal, Rebecca Turpin, Aditya Shah , Keith Murphy
Murulee Byappanahalli, Richard Whitman
Conference Presentations:
International Conference on Water Scarcity, Global Changes, and Groundwater Management Responses (Poster Presentation 12-2 -08), University of California, Irvine. “Chemical Markers to Determine High Bacteria Counts in Lake Michigan”
LSAMP Student Poster Presentations (11-7-08), Fourth Annual Joint Research Conference, Purdue University, West Lafayette, IN. “Fluorescent Whitening Agents as Chemical Markers of Human Contamination”
Great Lakes Beaches Association Meeting, Poster Presentation (9-15-08), Porter, IN. “Fluorescent Whitening Agents as Chemical Markers of Human Contamination”
Undergraduate Student Participants
Kevin Murphy, B.S. in Chemistry expected May 2009
Rebecca Turpin, B.S. in Chemistry expected May 2009
Aditya Shah, B.S. in Chemistry expected May 2009
