| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Civil and Environmental Engineering, Arizona State University, Box 5306, Tempe, AZ
* Corresponding author (p.westerhoff{at}asu.edu)
Received for publication July 2, 1999.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: DOC, dissolved organic carbon • EEM, excitation–emission matrix • MWW, unfractionated dissolved organic carbon sample from a tertiary wastewater treatment plant • SRF, Suwannee River fulvic acid sample
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Fluorescence spectroscopy has been used to characterize aquatic and soil humic and fulvic acids, and quantify metal ion complexation by DOC (Senesi et al., 1989; Visser, 1983; Ewald et al., 1983; Lombardi and Jardim, 1999; Ryan et al., 1996; Seitz, 1981a; Miano et al., 1988; Ghosh and Schnitzer, 1981; Blaser and Sposito, 1987). Although a considerable body of research is available on DOC fluorescence in estuary and marine waters (Coble, 1996), less information is available on fluorescence analysis of DOC in general bulk water matrices (e.g., lake, stream, and ground waters), and treated wastewater effluents in particular (Donahue et al., 1998; McKnight et al., 2001; Smith and Kramer, 1998; Cabaniss and Shuman, 1987). The wastewater treatment industry has used various fluorescence techniques, with limited sample preparation, to detect specific compounds such as polyaromatic hydrocarbons and aliphatic and aromatic amines, and the presence of genotoxic antibiotics (Miege et al., 1998; Djozan and Faraj-Zadeh, 1998; Hartmann et al., 1998). Various on-line fluorescence approaches at wastewater treatment plants have been proposed for monitoring biodegradation of easy-to-degrade organics (e.g., DOC that exerts a biological oxygen demand) and other substrates (Isaacs et al., 1998; Reynolds and Ahmad, 1997; Tartakovsky et al., 1996). Synchronous fluorescence has also been used to track the relative percentage of wastewater present in river water downstream of wastewater treatment plant discharges (Ahmad and Reynolds, 1995; Galapate et al., 1998; Cabaniss and Shuman, 1987). Further work is needed to understand fluorescence property differences between well-characterized, isolated DOC samples and treated wastewater samples. Work is also needed to determine which fluorescence techniques should be used to readily characterize DOC structure in wastewater.
This article compares three different fluorescence measurements (emission scan, synchronous scan, and excitation–emission matrix) for characterizing DOC of filtered wastewater matrices and a well-characterized aquatic fulvic acid. Considerations for the effect of DOC concentration and pH are discussed in the context of selecting appropriate conditions for fluorescence analysis of treated wastewaters. The application of fluorescence for differentiating treated wastewater from other DOC sources is also discussed. This work differs from previous studies since it compares two DOC samples with distinctly different origins and structural properties: (i) a tertiary treated denitrified wastewater effluent and (ii) a standard fulvic acid. Recommendations are made for future research that would aid in the application of fluorescence to characterize DOC in bulk wastewater matrices.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The MWW and SRF samples had initial DOC concentrations of 5.93 and 40.0 mg/L, respectively. Samples were diluted with 0.01 M KCl to achieve a range of DOC concentrations. Sample pH was adjusted using 2 M HCl. Most results are presented for a DOC concentration of 1 mg/L and a pH of 3. These conditions are described, and recommended, herein for analysis of wastewater.
Analysis
Dissolved organic carbon (DOC) was analyzed with a combustion analyzer (Shimadzu [Kyoto, Japan] TOC 5050). The detection limit was 0.4 mg/L. Ultraviolet and visible (UV/Vis) spectrometric measurements (200 to 660 nm) were recorded on a Shimadzu 1601 instrument using a 1-cm pathlength quartz cell at a DOC of 1 mg/L. Samples were diluted with 0.01 M KCl. The percentage of hydrophobic acid in the sample was determined by XAD8 resin chromatography, based upon standard methodologies (Aiken et al., 1992).
Fluorescence measurements were conducted with a Shimadzu LSB-50 fluorescence instrument and controlled and analyzed with Shimadzu Winlab software. A 1-cm pathlength quartz cell was used. A scan speed of 500 nm/min was used with a slit width opening of 10 nm. Opening the slit wider allows more light energy from the xenon light source to excite the DOC molecules in the sample. A larger slit width applies more excitation light energy to a sample, and a less–well defined spectral purity of the excitation and emission bands (e.g., broader emission bands rather than sharper bands) results.
Three types of fluorescence analysis were conducted: (i) excitation–emission matrix (EEM) profile, (ii) emission spectra, and (iii) synchronous spectra. All EEM profiles and emission scans were performed over a range of excitation (
EX) and emission (EM) wavelengths. Emission spectra were conducted at an excitation of 370 nm, which has been used in previous studies of aquatic DOC (Donahue et al., 1998). Synchronous spectra quantifies the emission fluorescence intensity at a constant wavelength offset 
from a set of excitation wavelengths. Studies in the literature have used 
values between 2 and 100 nm to characterize DOC (Cabaniss and Shuman, 1987; Galapate et al., 1998; Miano and Senesi, 1992). The theoretical selection of for a particular fluorophor is based on the Stokes shift of a compound. Compounds with a ridged molecular structure normally show a shorter Stokes shift (shorter ) while compounds that undergo reactions in the excited state (e.g., proton transfer, twisted intramolecular charge transfer reactions, etc.) show a larger Stokes shift (detected by a longer ). The slit width cannot be set to a value greater than half the value, so a value of 10 nm was used for all synchronous spectra. A 290-nm emission cutoff filter was used in most EEM analyses.
Quinone sulfate (QS) solution (1 µg QS/L in 0.1 M H2SO4) was used during the study as a standard solution to monitor the relative energy emitted by the xenon lamp in the fluorimeter (Seitz, 1981b; McKnight et al., 2001; Scully and Lean, 1994). During this study no change in quinone sulfate fluorescence was observed, and consequently the standard was used as a sensitivity check rather than an as an absolute calibration tool.
Data Handling
To account for Raleigh scattering, the fluorimeter instrument response of a 0.01 M KCl solution was subtracted from the fluorescence spectra recorded for samples containing DOC. The 0.01 M KCl solution was prepared from Super-Q water (Millipore), and contained a DOC < 0.2 mg/L.
To account for the absorbance by the DOC of light from the lamp and emitted light, an inner-filter correction (Mobed et al., 1996) was conducted using UV/Vis absorptivity data reported herein. We assumed that the emission spectrometer views only a small illuminated volume in the center of the cell, and that the effective pathlength of excitation light is 0.5 cm and the pathlength of the fluorescence light going through the sample to the emission monochromator is also 0.5 cm. The absorbance of excitation light (AEX) was calculated as the measured absorptivity at a given wavelength multiplied by the DOC of the solution and by 0.5 cm; similarly, the absorbance of emitted light (AEM) was calculated. The corrected fluorescence intensity (ICOR) reported in arbitrary units (AU) was calculated from the observed fluorescence intensity (IOBS) by the following relationship:
 | [1] |
Samples were not corrected for instrument responses (excitation lamp source or emission spectra detection). All data were collected on a single instrument and therefore the results were internally comparable.
Excitation–emission matrix spectra were normalized to 1.0 by dividing all fluorescence intensity at each wavelength pair by the highest fluorescence intensity recorded for the sample. The increment of contour lines was 0.01, resulting in 100 contours per EEM graph. Data handling was conducted with Matlab software (The Mathworks, 2001).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
0.75 mg N/L) or other anions. For all samples, absorbance between 400 and 660 nm was less than 0.003 cm-1. The effect of pH on absorbance in MWW was negligible. Absorbance at pH 3 in the SRF sample was approximately 25% lower than the absorbance at pH 7 over the wavelength range of 200 nm to 350 nm, and then slightly greater at longer wavelengths.
|
|
The SRF sample exhibited two broad-shaped EEM peaks located near 210 nm:440 nm and 325 nm:440 nm. The EEM peak near 325 nm:440 nm in SRF was similar to the EEM peak detected in MWW. The EEM peaks represent the presence of classes of DOC (discussed later). However, reading fluorescence values from EEM contours can be difficult. Emission spectra represent a "slice" through an EEM profile at a single excitation wavelength. A portion of the broad-shaped EEM peak described above for SRF, and for MWW, is presented in Fig. 3
for DOC concentrations of 1 mg/L (pH = 3) and a fixed EX of 370 nm. Data shown in Fig. 3 were absorbance corrected. However, at a DOC concentration of 1 mg/L and the longer excitation and emission wavelengths, the correction factor was constant over all the wavelength and just a function of the cell path length of 0.5 cm (i.e., absorbance correction had no effect on the peak shape). Therefore, the correction had minimal effect on the spectra shape. The SRF sample exhibited a broader-shaped peak than MWW, and had a peak fluorescence 
located at a longer wavelength than MWW 
.
|
). Therefore, synchronous spectra represents multiple EEM peaks, or a "slice" through an EEM profile with a slope equal to . Synchronous spectra for MWW and SRF for values of 20, 40, and 60 are presented in Fig. 4
. Increasing from 20 nm to 40 nm and 60 nm resulted in higher fluorescence intensity and changed the spectra shape for the MWW and SRF samples. The MWW sample generally exhibited greater fluorescence compared against the SRF sample. Improved peak separation was observed for the MWW sample at longer values.
|
. Absorbance-corrected fluorescence intensities at three excitation wavelengths (275, 327, and 380 nm) were plotted as a function of DOC (Fig. 5)
. The corrected fluorescence data is nearly linear 
, although a logarithmic data fit provided a slightly higher R2 value but did not visually fit the data as well as the linear fit. The nonlinear response occurred in MWW with only small dilutions (e.g., DOC > 3 mg/L). The MWW samples had more nitrate than solutions with greater dilutions (i.e., lower DOC), which may have affected the light energy reaching the molecule or detected by the instrument due to the assumption implicit to the inner-filter absorbance correction method. Similar analysis was conducted using the fluorescence intensity maxima from an emission spectra 
, and a linear response of absorbance-corrected fluorescence versus DOC was observed (not shown). Comparable analysis was conducted using SRF over a larger DOC range (0.5 to 40 mg/L), and the absorbance-corrected fluorescence response was linear from 0.5 to 30 mg/L. Above 30 mg/L, a logarithmic response was observed.
|
conducted at pH 3 and 7, with and without absorbance correction, are presented in Fig. 2. Minor differences occurred between pH 3 and 7 relative to the location of EEM peaks for SRF. The effect of absorbance correction on the location of EEM peaks at pH 3 and 7 was minor; a slight blue shift in EEM peaks to shorter wavelengths. Comparable results were obtained for MWW. Lowering the pH from 7 to 3 in both SRF and MWW resulted in approximately a 30 to 40% decrease in maxima fluorescence intensities at most EEM peaks and over most of the EEM range illustrated in Fig. 2. However, the EEM peak located near 250 nm:320 nm was more sensitive to pH changes than other EEM peaks, with higher fluorescence intensity at pH 7 than pH 3. In SRF this EEM peak is almost undetectable at pH 3 (Fig. 2C), but quite pronounced at pH 7 (Fig. 2E). The fluorescence intensity of the same EEM peak in MWW differed by approximately 30% in MWW.
Effect of Detergent Whiteners
The three aromatic disulfonic acid whitening agents all exhibited high fluorescence efficiencies with measurable fluorescence occurring in the presence of 1 µg/L of the whitening agent. The whitening agents exhibited pH dependencies (pH 3 to 7) as functional groups were protonated or deprotonated. The largest pH effects were observed for the CBS-X (distyrylbiphenyl) whitening agent. The locations of EEM peaks for the whiteners were near 260 nm:430 nm, 260 nm:540 nm, and 400 nm:460 nm.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Energy absorbed by different classes of organic compounds is emitted as light energy (fluoresce) at various efficiencies and wavelengths. The MWW sample had lower UV/Vis absorbance than SRF above 240 nm (Fig. 1). However, MWW had higher fluorescence intensities over a large range of excitation and emission pairs (Fig. 4), especially over the range of excitation wavelengths shorter than 400 nm, and emission wavelengths shorter than 450 nm. At longer wavelengths SRF tended to have a higher fluorescence intensity than MWW (Fig. 3).
The SRF sample was characterized by two dominant EEM peaks located near an emission wavelength of 450 nm. Other researchers have also reported the presence of two EEM peaks for SRF (Goldberg and Weiner, 1989; Coble, 1996). Based upon fluorophor lifetime measurements, it has been suggested that the two EEM peaks are due to the presence of two different fluorophors in SRF (Goldberg and Weiner, 1989).
The two fluorophors, or EEM peaks, described above showed minor pH dependencies in comparison with the EEM peak located near 250 nm:330 nm that was detected in SRF at pH 7, but not pH 3 (Fig. 2C and 2E). It is difficult to assess whether this is a unique fluorophor, or if the peak represents a first or second excited state of another EEM peak. In addition, it could be related to the ability of light energy to penetrate and excite a "coiled" (pH 3) versus "uncoiled" (pH 7) confirmation of the molecule. It is also possible that the observed EEM peak is related to functional groups present on the DOC molecule. Different functional groups can cause a large decrease (e.g., carboxyl) or increase (e.g., amine, hydroxyl) in fluorescence intensity due to the electron withdrawing or donating properties of the functional groups, respectively (Seitz, 1981b; Da Silva et al., 1998). The EEM peak detected at pH 7 (250 nm:330 nm) may contain either amine or hydroxyl functional groups. Since SRF is from an allochthonous source with lignin-derived DOC, which is typically low in organic nitrogen, the EEM peak detected at pH 7 may be dependent upon the protonation state of hydroxyl functional groups.
The dominant fluorophors present in MWW occurred at shorter wavelengths (blue-shifted) than the SRF sample. Peaks in the MWW sample may have been associated with protein-like compounds, since aromatic amino acids have been suggested to fluoresce in that EEM region (Coble, 1996; Wolfbeis, 1985). The MWW sample contained only 26% of the DOC as hydrophobic acids. The remaining 74% would be characterized as hydrophilic acids or hydrophobic–hydrophilic neutrals and bases. Marhaba (2000) reported regions of fluorescence for six fractions of DOC (hydrophobic–hydrophilic acids, neutrals, and bases) from the influent for a water treatment plant. The fractions, analyzed separately, had peaks at excitation wavelengths ranging from 225 to 250 nm. The hydrophobic acid fraction had a peak at 249 nm:429 nm, similar to the SRF sample (Fig. 2). The base and neutral fractions had peaks at emission wavelengths of 300 to 380 nm. Peaks on EEM spectra for MWW (Fig. 2) at shorter wavelengths may be representative of base and neutral fractions. Such molecules frequently include organic nitrogen–containing functional groups. This would be consistent with wastewater effluent, which contains soluble microbial products (e.g., amino acids) from the bacterial treatment processes (e.g., activated sludge treatment).
Fluorophors, or peaks on the spectra, for the whitening agents were located away from the dominant peaks in the MWW sample. Despite the highly fluorescent nature of the detergent whiteners potentially present in wastewater effluents, it was likely that the whitners were present at very low concentrations and did not affect the spectra of the bulk MWW sample. It is possible that compounds similar to the whitening agents used in our study entered the Mesa Northwest Wastewater Reclamation Plant (NWWRP), but underwent biochemical transformations that altered the molecular structure and hence altered its fluorescence properties (e.g., shifted its EEM peak). Likewise it was assumed that although the fluorescence spectra for the detergent whiteners exhibited a strong pH dependency, they were not responsible for the observed pH effects in the MWW sample.
Wastewater-Related Data Analysis and Applications of Fluorescence
Quantifiable Parameters to Assess Different Sources of Dissolved Organic Carbon
Excitation–emission matrix peaks are broad shaped (Fig. 3), and have been fit by some researchers with a Gaussian function to represent a mixture of heterogeneous compounds with similar properties (Pelikan et al., 1994; Korshin et al., 1999). A Gaussian function (Eq. [2]) may be a convenient way to quantify differences in spectral bands among samples from different sources:
 | [2] |
In Eq. [2], the emission intensity, I(E), is a function of emitted light energy, E (eV), and EM, which is the width of the emission band measured at 50% of the maximum intensity, IMAX. The value of emitted light energy (E) was computed as a function of the wavelength (, nm) by the equation: E 
= 1240/. Computed parameter values for SRF and MWW samples analyzed in triplicate are presented in Table 1, and fitted lines using the parameters are illustrated with the observed data in Fig. 3. The fitted parameters are statistically significant at the 95% confidence level based upon a Student t test statistical analysis. Compared against the SRF sample, the MWW sample exhibits a higher maximum intensity (EMAX) and a broader emission band at 50% of the maximum intensity (EM). Such differences may be indicators for DOC of different origins (e.g., wastewater vs. natural fulvic substances).
|
max and/or a fluorescence ratio (FR) of emission intensities at 450 to 500 nm 
was found to be suggestive of either terrestrial-based (i.e., allochthonous) or algae- and bacteria-based (i.e., autochthonous) origins for the DOC (McKnight et al., 2001; Donahue et al., 1998). A shorter maximum wavelength (max) value (<450 nm) or high FR value (>1.8) was correlated with DOC having algae- or bacteria-based sources. Longer max values (>450) and low FR values (<1.5) were more indicative of DOC derived from a plant- or soil-based origin. Data for the SRF and MWW samples are presented in Table 1. Interpretation of the max and FR values in Table 1 would suggest that fluorophors in the SRF sample were derived from soils or plants (i.e., allochthonous) while the MWW samples were derived from alga or bacteria. The inferred sources based upon max and FR values were consistent with the known origins of the samples. The SRF sample was isolated from a river draining the Okefenokee Swamp, which contains large amounts of lignin-based carbon sources. The MWW sample was obtained from the effluent of a biological (i.e., bacteria) wastewater treatment plant. Assessment of the spectra by either a Gaussian function or arithmetic analysis provided quantitative measures for the distinct characteristics of DOC from different sources.
Applying Fluorescence to Estimate the Percentage of Dissolved Organic Carbon from Wastewater Sources
Dissolved organic carbon from different sources contains different dominant fluorophors and exhibits different fluorescence "signatures" (i.e., different EEM peak locations and intensities). Several methods for estimating the percentage contribution of one water source (e.g., a wastewater effluent) after being mixed with another water source (e.g., river sample located downstream of a wastewater treatment plant) have been reported (Cabaniss and Shuman, 1987; Seitz, 1981a). To demonstrate the potential of fluorescence spectroscopy for this application, various blends of MWW and SRF samples were prepared and analyzed by synchronous fluorescence . Fluorescence intensities at four excitation wavelengths 
were selected for analysis because they generally represented localized maxima on the synchronous spectra. All samples were prepared at an equivalent DOC concentration (1 mg/L) for direct comparison of the data. The fluorescence intensities at several wavelengths were then plotted against the known percentage of MWW relative to SRF (Fig. 6)
. At all wavelengths an R2 goodness of fit value greater than 0.99 was obtained. The relationship at 228 nm had the steepest slope, and would therefore provide the accurate prediction for the percentage of MWW in an unknown mixture of MWW and SRF.
|
Recommended Sample Conditions for Fluorescence Analysis of Wastewater
Organic molecule efficiency to fluorescence can vary as a function of pH (Ghosh and Schnitzer, 1980; Balkas et al., 1983). Deprotonated acids can exhibit a reduction in fluorescence at one wavelength and an increased fluorescence at another. Carboxyl, hydroxyl, and amine function groups are the most common functionality on the carbon backbone for DOC in environmental applications. At high (>11) and low (<3) pH levels the shape and confirmation of fulvic acids can change as a result of stripping protons that break hydrogen bonds or force the addition of protons onto oxygen-containing functional groups (Goldberg and Weiner, 1989; Lapen and Seitz, 1982). Another important pH dependency factor for fluorescence is complexation of DOC with metals (Blaser and Sposito, 1987; Bidoglio et al., 1997; Grimm et al., 1991). Fluorescence changes, either enhancement or quenching, during metal titrations can be observed and used to fit complexation models to determine DOC–metal conditional stability constants (Ryan and Weber, 1982; Smith and Kramer, 1998). Metal–DOC complexation is pH dependent, with less interaction occurring at low (<4) pH levels. Many wastewaters contain metals (e.g., calcium, aluminum, iron, magnesium) that can complex DOC and quench or enhance fluorescence.
The authors recommend that all wastewater samples be acidified to pH 3 prior to fluorescence analysis. At pH 3 metal complexation of DOC would be minimized, precipitation of hydrophobic acids should not occur, and the UV/Vis absorbance is only moderately affected from neutral pH conditions.
As illustrated in Fig. 2, when a wastewater sample or model solution (i.e., SRF) is diluted with 0.01 M KCl to a final DOC concentration of 1 mg/L, application of an inner-filter correction has minimal effects on the location of EEM peaks. Furthermore, at a DOC concentration of 1 mg/L, the correction factor associated with an inner-filter correction is almost constant over wavelengths above 340 nm. At shorter wavelengths the amount of UV-adsorbing DOC material is low, and has minor effects on the shapes of EEM peaks. Most treated wastewaters have DOC ranging from 5 to 25 mg/L and nitrate ranging from 1 to 10 mg N/L.
The authors recommend dilution of wastewater, containing inorganics that can adsorb light (e.g., nitrate, sulfite), to a DOC of 1 mg/L with 0.01 M KCl. This would reduce the absorbance interference from inorganic anions and the need for time-consuming inner-filter corrections. The drawback to not conducting inner-filter corrections is a potential to overestimate the fluorescence intensity.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Excitation–emission matrix profiles can be time consuming to collect and analyze. More rapid analysis of large numbers of wastewater samples could be accomplished by a single emission spectra or synchronous spectra. The analysis herein supports the recommendation of the following spectra analysis (slit width = 10 nm) for wastewaters: (i) emission spectra 
with EX = 370 nm to detect organic acids or with EX = 280 nm to detect organic bases and neutrals that could represent soluble microbial products (e.g., amino acids); or (ii) synchronous spectra with = 60 nm from EX = 250 to 600 nm. It is recommended that all samples be filtered, diluted to a DOC of 1 mg/L with 0.01 M KCl, and acidified to pH 3. This approach would prevent absorbance interferences from particulates and anions (e.g., nitrate), reduce the interaction between metals and DOC, which might quench or enhance fluorescence, reduce absorbance interferences from DOC molecules, and eliminate the need for applying an inner-filter correction.
Data analysis of fluorescence EEM profiles and spectra can be conducted qualitatively or quantitatively. The location of EEM peaks provides a qualitative indication of types of DOC molecules present in the wastewater. Gaussian parameters (EMAX and EM) or simple arithmetic parameters (e.g., fluorescence ratio) obtained from emission spectra provide quantitative data that appear to indicate potential origins of DOC. An additional quantitative application of fluorescence spectra is the use of fluorescence intensities as "fingerprints" for determining the percentage of DOC from one source (e.g., wastewater effluent) or another source (e.g., surface water receiving wastewater effluent).
The authors recommend synchronous fluorescence for monitoring the presence of wastewater in source water. Excitation–emission matrix analysis is recommended for identifying major differences in fluorescence between sources. For analysis of a large number of samples, water managers must balance time requirements for analyzing a large number of samples against the information that can be elucidated from the different analytical approaches.
Further research is needed to address the following research questions:
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. Ohno and R. Bro Dissolved Organic Matter Characterization Using Multiway Spectral Decomposition of Fluorescence Landscapes Soil Sci. Soc. Am. J., October 27, 2006; 70(6): 2028 - 2037. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Vadose Zone Journal | Journal of Plant Registrations | ||||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal |
= 1 mg/L; pH = 3; 
at several wavelengths (275, 327, and 380 nm) for the unfractionated dissolved organic carbon sample from a tertiary wastewater treatment plant (MWW) (pH = 3; inner-filter correction applied).