PII: S0043-1354(00)00222-0

Wat. Res. Vol. 35, No. 1, pp. 327332, 2001 7 2000 Elsevier Science Ltd. All rights reserved Printed in Great Britain 0043-1354/00/$ - see front matter

www.elsevier.com/locate/watres

TECHNICAL NOTE EVALUATION OF A SECOND DERIVATIVE UV/VISIBLE SPECTROSCOPY TECHNIQUE FOR NITRATE AND TOTAL NITROGEN ANALYSIS OF WASTEWATER SAMPLES
MICHELLE A. FERREE1 and ROBERT D. SHANNON2*
249 Agricultural Engineering Building, The Pennsylvania State University, University Park, PA 16802, USA and 2Agricultural and Biological Engineering Department, 233 Agricultural Engineering Building, University Park, PA 16802, USA (First received 1 August 1999; accepted in revised form 17 March 2000) AbstractA method for nitrate analysis based on second derivative UV/Visible spectroscopy was developed by Simal et al. (1985: Simal J., Lage M. A., and Iglesias I. (1985) Second derivative ultraviolet spectroscopy and sulfamic acid method for determination of nitrates in water. J. Assoc. Analyt. Chem. 68, 962964) and Suzuki and Kuroda (1987: Suzuki, N. and Kuroda R. (1987) Direct simultaneous determination of nitrate and nitrite by ultraviolet second-derivative spectrophotometry. Analyst 112, 10771079), and later modied for the analysis of total nitrogen in aqueous samples of varying nitrate:organic nitrogen ratios (Crumpton et al., 1992: Crumpton W. G., Isenhart T. M. and Mitchell P. D. (1992) Nitrate and organic N analyses with second-derivative spectroscopy. Limnol. Oceanogr. 37, 907913). The procedure uses the second derivative of the absorption spectrum for nitrate (NO), which has a peak at 0224 nm that is proportional to the NO concentration. Samples 3 3 for total N analysis are rst oxidized to NO by persulfate digestion. The objectives of this study were 3 to: (1) test the accuracy and precision of the second derivative method through the use of NISTtraceable wastewater check samples; (2) determine whether the second derivative method for nitrate analysis can be used for wastewater samples and whether the method compares favorably with other currently used nitrate analysis methods; and (3) use the method to analyze wastewater samples containing a range of nitrate and total nitrogen concentrations. Our results indicated that the method needed to be modied to include a longer digestion time (60 min) and dilution of samples prior to digestion (if needed). With the modied method, nitrogen recoveries were not signicantly dierent (P r 0.05) from samples with known N concentrations. In addition, nitrate concentrations in constructed wetland and wastewater samples analyzed by both second derivative spectroscopy and ion chromatography were not signicantly dierent. Total nitrogen concentrations in wastewater samples also compared favorably to the same samples analyzed by Kjeldahl digestion. The method is faster, simpler, requires smaller sample volumes, and generates less waste than many EPA-approved methods of N analysis, and may oer a suitable alternative to current methods for analysis of nitrate and total N in wastewater samples. 7 2000 Elsevier Science Ltd. All rights reserved Key wordsnitrate, organic nitrogen, ammonium, spectroscopy
1

INTRODUCTION

Nitrogen is an essential nutrient for autotrophic and heterotrophic organisms and is often limited in aquatic ecosystems. Conversely, large inputs of nitrogen can cause excessive phytoplankton and macrophyte production, and the subsequent death and decay of these organisms can result in a depletion of dissolved oxygen. Because nitrogen compounds are biochemically transformable through

*Author to whom all correspondence should be addressed. Tel.: +1-814-865-7153; fax: +1-814-863-1031; e-mail: rds13@psu.edu 327

nitrication, denitrication, and mineralization, wastewater treatment systems are frequently designed to convert nitrogen compounds to more environmentally benign forms (i.e., N2 gas) that will not cause problems associated with nutrient enrichment and oxygen depletion. Nitrogen occurs in various inorganic and organic forms, including nitrite (NO), nitrate (NO), am2 3 monium (NH+) and organic nitrogen (organic N). 4 Nitrogen compounds must be monitored in wastewater treatment systems in order to understand the transformations that occur, and to develop mass balances on nitrogen compounds removed through

328

Michelle A. Ferree and Robert D. Shannon

treatment processes. EPA-approved analytical methods exist for the analysis of each of these forms, but many of these methods are expensive, complicated, or time-consuming. Ion chromatography may be one of the safer analysis methods, because it eliminates the need to use hazardous reagents. However, ion chromatography is expensive and the instrument requires frequent maintenance to function properly. This analytical method is also time-consuming, with analysis times as high as 15 min for one sample. An advantage of ion chromatography is that other ionic constituents can be analyzed simultaneously. The cadmium reduction technique is another common method for analysis of NO in water. This technique reduces NO to 3 3 NO by passing the sample through a column of 2 copperized cadmium metal lings. In addition to NO reduction, the cadmium reduction method 3 involves a colorimetric reaction of NO and a sub2 sequent absorbance measurement of the reacted solution. The cadmium-reduction technique requires specialized glass columns for ecient analysis, as well as a spectrophotometer or lter photometer. The technique also involves the handling and disposal of hazardous cadmium and phenol waste. Ion electrodes can measure both NO and NO, are 2 3 cheaper, and have a larger applicable range than other analysis methods. However, precision of this method is somewhat lower than that of other methods. Additionally, this method is subject to interference from several ions that occur in natural waters. The major EPA-approved method for the analysis of organic nitrogen is the Kjeldahl method (USEPA, 1983). The macro-Kjeldahl method is used for samples of varying N concentrations, but requires large sample volumes (up to 500 ml) for analysis of low N samples. A modication of the Kjeldahl method, the semi-micro-Kjeldahl method, allows smaller sample volumes, but the sensitivity limits often exceed the low N concentrations found in natural waters (APHA, 1998). Both Kjeldahl methods require time-consuming digestions, special safety equipment, and the use of toxic reagents like mercuric oxide. In addition, the Kjeldahl method is not suitable for the analysis of high NO/low or3 ganic N samples commonly found in agricultural and nitried wastewaters (Crumpton et al., 1992; Bachmann and Caneld, 1996). Because of the various disadvantages associated with current nitrogen analysis techniques, investigators have been seeking less time-consuming and more reliable methods for the analysis of nitrogen in water samples. One procedure that shows promise for freshwater samples is based on second derivative spectroscopy. This method has been investigated for the analysis of NO in fresh surface 3 waters (Simal et al., 1985; Suzuki and Kuroda, 1987; Wollin, 1987). More recently, the second derivative procedure has been adapted to accommo-

date aqueous sample digestion and analysis for organic N (Crumpton et al., 1992; Bachmann and Caneld, 1996). The second derivative spectroscopy method requires the use of a UV/Visible scanning spectrophotometer. This instrument has a large upfront cost, but requires minimal future maintenance. The method is simpler and much faster than many other methods, as spectrophotometric analysis takes approximately 1 min per sample. The purpose of this study was to examine the accuracy and precision of the method for the analysis of wastewater samples collected from a constructed wetland and conventional secondary wastewater treatment system. This was achieved by the use of quality control (QC) check samples of known N concentrations and through comparison of sample concentrations obtained with the second derivative method with currently accepted methods of N analysis.

MATERIALS AND METHODS

The procedure is based on the second derivative of the absorption spectrum for NO, which has a peak at 3 0224 nm that is proportional to the NO concentration 3 (Fig. 1). The second derivative peak, unlike the absorption peak at 203 nm, is not inuenced by other common constituents in fresh waters (i.e., organic matter). The principle of the modication to analyze for total N is to digest and oxidize all nitrogenous compounds to NO by auto3 claving. The NON concentration of the digested sample 3 is then equivalent to total N. Organic N can be determined by analyzing the sample for NO, NO, and NH+ 3 2 4 and subtracting these values from total N. Analysis of NO 3 NON standards of concentrations ranging from 0.1 3 to 3.0 mg l1 were prepared by dilution from a 1000 mg 1 l , NIST-traceable stock solution of NO. Ultrapure 3 water blanks, standards and samples were scanned on a PerkinElmer Lambda-40 UV/VIS spectrophotometer interfaced to a microcomputer, with ultrapure water used as a blank in the spectrophotometer reference cell. Standards and samples were scanned between 190 and 250 nm in quartz cells with a 10-mm optical path length (Crumpton et al., 1992). Scans were conducted at 120 nm/min, and the second derivative value of the NO peak at 3 224 nm was measured and recorded. A standard curve was developed by linear regression of the second derivative peak vs concentration of the standards. The standard curve was then used to quantify all samples. This method allows analysis of N concentrations up to 3 mg l1 without dilution, as the N curve is nonlinear beyond this range. Samples containing more than 3 mg l1 N were diluted to a concentration within the linear range. Analysis of total N/organic N Organic-N standards of concentrations ranging from 0.1 to 3.0 mg l1 were prepared from reagent-grade urea. Standards or samples (initially 10 ml) were placed in Teon-lined screwcap vials for digestion by autoclave. Oxidizing reagent was prepared by dissolving 6.0 g of low N potassium persulfate in 100 ml of 1.5 N sodium hydroxide solution. The oxidizing reagent (1.5 ml) was added to each vial with sample or standard, and the vials were tightly capped and shaken. After autoclaving, samples were cooled and acidied to below pH 2 with concentrated sulfuric acid. Samples were then analyzed for NO as 3

Evaluation of a nitrogen analysis technique described above, with a digested ultrapure water blank in the reference cell. Method validation for total N The standard protocol for sample digestion and total N analysis requires samples to be autoclaved at 1218C and 15 psi for 30 min (Crumpton et al., 1992). However, N was not fully recovered with this procedure, based on low values obtained from the QC check samples. The low recoveries indicated that the samples were not being completely digested and the experimental protocol needed to be modied. We performed a three-factor analysis of variance (ANOVA) experiment using the QC check samples with known N concentrations to simultaneously determine the eects of three factors on the amount of N recovered. The three factors were digestion time (30 vs 60 min), sample volume (5 ml vs 10 ml), and timing of dilution (prior to autoclaving vs after autoclaving). Several QC check samples were analyzed to assess the accuracy of NO recovery in undigested samples and for 3 the various treatments in the three-factor ANOVA experiments with digested samples. One check sample contained only inorganic N as NON (9.35 2 0.09 mg l1) and 3 NH+N (9.43 2 0.09 mg l1) and was analyzed without 4 persulfate digestion to determine the NO concentration 3 in the sample. The same QC sample was digested and analyzed for total inorganic N (NON+NH+N=18.78 mg 3 4 l1). The second QC sample, which contained organic N 1 as glycine (14.0020.14 mg l ) was also digested and analyzed for total N. In addition, Standard Methods (APHA, 1998) recommends the use of nicotinic acid in testing the Kjeldahl N procedure, so we also determined the recovery of a 3 mg N l1 nicotinic acid standard prepared from reagent grade nicotinic acid. Wastewater samples Water samples were collected from a constructed wetland system treating domestic wastewater to secondary standards and from a conventional secondary wastewater treatment system in order to obtain samples of varying N concentrations. Samples were placed in plastic vials and were stored at 48C or frozen until analyzed. All sample ltration was conducted in the eld or immediately upon return to the laboratory. For analysis by several methods, samples were shaken and split into separate vials immediately before analysis. All sample and standard concentrations were reported as NON, NH+N, and organic3 4 N. Dissolved organic carbon (DOC) concentrations were

329

measured on a subset of the wastewater samples with a Shimadzu total organic carbon analyzer. Statistical analyses All data were statistically analyzed using Minitab (1996). Three-factor ANOVA was used to determine factors having a signicant inuence on the recovery of N in the digestion procedure for total N. Paired sample t-tests were used to determine whether N concentrations recovered from the QC samples were signicantly dierent from certied values. Simple linear regression analysis was performed to test the comparability of the second derivative spectroscopy method for NO and total N analysis with 3 accepted methods for NO and total inorganic N analysis. 3 Signicance of all tests was determined with a=0.05.

RESULTS AND DISCUSSION

NO 3 raphy

recovery and comparison with ion chromatog-

We consistently obtained excellent recovery of NO in the undigested QC check sample. The mean 3 (295% condence interval) for nine separate analyses of the check sample (conducted on nine dierent days with new calibration standards) was 9.37 (20.15) mg l1, indicating that our recovered values were not signicantly dierent from the certied value of 9.35 mg l1 (0.2% higher) and that the precision of the method was also very good. To further assess the second derivative method, wastewater samples were analyzed for NO by sec3 ond derivative spectroscopy and by ion chromatography, a commonly used NO analysis technique. 3 Analysis by simple linear regression showed a very 2 strong relationship (r =0.99) between the two methods (Fig. 2). The slope of the best-t regression line (295% condence interval) was 1.00 (20.03) and the y-intercept (295% condence interval) was 0.17 (20.27) mg l1, demonstrating that there was no signicant dierence between NON concen3 trations obtained through ion chromatography and the second derivative method. The DOC of the

Fig. 1. Absorbance, rst derivative, and second derivative spectra for a 2.0-mg N l1 nitrate standard.

330

Michelle A. Ferree and Robert D. Shannon

wastewater samples ranged between 9 and 77 mg l1, but there were no signicant interferences from organic matter in the samples. However, other wastewaters containing high DOC concentrations may absorb within the range of NON absor3 bance and should be analyzed with caution. Our results indicate that the second derivative spectroscopy method is an accurate and precise technique for the measurement of NO in similar 3 domestic wastewater samples of varying NO con3 centrations. Other researchers have conducted experiments with lake water comparing the second derivative method for NO analysis to the auto3 mated cadmium reduction method and found no signicant dierence between the two techniques (Crumpton et al., 1992). Eects of sample volume, digestion time, and timing of dilution on total N recovery The three-factor ANOVA experiment showed that timing of dilution, digestion time, and sample volume were all signicant factors in obtaining total N recovery. Timing of dilution was the most important factor. For digested QC check samples containing either NO+NH+ or organic N, the 3 4 highest N recoveries were in samples diluted prior to autoclaving. Autoclave time was the second most important factor, with a 60-min autoclave time recovering more N than 30 min. Sample volume was the least important (but still signicant) factor. Overall, better recoveries were obtained using a 10ml sample volume. The paired sample analysis of recovered values for the QC check sample containing NON and 3 NH+N vs the certied value showed that the 4 samples diluted prior to autoclaving and autoclaved for 60 min did not dier signicantly (P r 0.05) from the certied value of the check sample (Fig. 3). All samples autoclaved for 30 min were signicantly less than the certied value, as were the samples

autoclaved for 60 min but diluted after autoclaving. The N concentrations (295% condence interval) recovered for 5 ml and 10 ml samples diluted prior to digestion and autoclaved for 60 min were 18.36 (20.53) mg l1 and 18.06 (20.78) mg l1, respectively, only 2.2 and 3.8% less than the certied value of the check sample (18.78 mg l1). Paired sample analysis of recovered values for the QC check sample containing organic N showed that 10-ml samples diluted prior to digestion (regardless of digestion time) did not dier signicantly (P r 0.05) from the certied value of the QC check sample (Fig. 4). All other samples were signicantly dierent from the certied value of 14.00 mg l1. The best results were obtained with the 10-ml samples diluted prior to digestion and autoclaved for 60 min. The N concentration (295% condence interval) recovered from these replicates was 14.05 (20.51) mg l1, only 0.4% higher than the certied value of the QC check sample. For digested samples, the most consistent N recovery was obtained for inorganic (NO+NH+) 3 4 and organic N samples containing 10 ml that were diluted prior to digestion and autoclaved for 60 min. By following the original method, not all forms of nitrogen were completely oxidized to NO; there3 fore total N concentrations were signicantly less than the actual sample values. Based on the results of our three-factor and paired sample analyses, we modied the method by increasing the autoclave time from 30 to 60 min. Samples were also diluted prior to digestion if necessary to ensure that they were less than 3.0 mg l1 as N, and the digested sample volumes were xed at 10 ml. With these modications, N concentrations recovered from the QC check samples were not signicantly dierent from their certied values, and all subsequent samples were digested with these modications. The mean recovery of nicotinic acid standards containing 3 mg N l1 was 97.3%, similar to recoveries of

Fig. 2. Comparison of nitrate concentrations in constructed wetland samples as determined by second derivative spectroscopy method vs ion chromatography.

Evaluation of a nitrogen analysis technique

331

Fig. 3. Inorganic N recovery from NIST-traceable standards for eight treatment combinations. Treatments refer to: (1) sample volume (5 ml vs 10 ml); (2) digestion time (30 min vs 60 min); and (3) timing of dilution (prior to autoclaving or after). Error bars are 95% condence intervals (n = 3). () Indicates treatments not signicantly dierent from certied concentrations of NIST-traceable standard (Pr0.05).

other organic N compounds (e.g., glutamic acid, sulfanilamide, and methionine) obtained with the method by Crumpton et al. (1992). Comparison of second derivative method with current methods for dissolved inorganic nitrogen Samples containing a range of total N concentrations were obtained from a constructed wetland site. Total N obtained by second derivative spectroscopy was compared to total inorganic N determined by ion chromatography for NO+NO and 2 3 the automated phenate method for NH+ (APHA, 4 1998). Our goals were to assess dierences between the two methods, and to determine if dissolved organic nitrogen was a signicant component in these

constructed wetland samples. All data were compared by regression analysis. The coecient of determination was very high (r 2=0.99), indicating a strong relationship between the two analytical methods (Fig. 5). The y-intercept (295% condence interval) was 0.31 (20.43), indicating that the two methods compared favorably at low concentrations of total N. However, the slope of the best-t line for the data, 1.04 (20.03), was signicantly greater than 1.0. Based on excellent recoveries obtained with the QC sample containing organic N, we concluded that the slightly inated slope was due to a small but signicant fraction of dissolved organic N in the wetland samples. The organic N was detected by the second derivative

Fig. 4. Organic N recovery from NIST-traceable standards for eight treatment combinations. Treatments refer to: (1) sample volume (5 ml vs 10 ml); (2) digestion time (30 min vs 60 min); and (3) timing of dilution (prior to autoclaving or after). Error bars are 95% condence intervals (n = 3). () Indicates treatments not signicantly dierent from certied concentrations of NIST-traceable standard (Pr0.05).

332

Michelle A. Ferree and Robert D. Shannon

Fig. 5. Comparison of total N and total inorganic N concentrations in constructed wetland samples as determined by second derivative spectroscopy method vs ion chromatography (for NO and NO) and 2 3 automated phenate method (for NH+). 4

method, but not by the ion chromatography method or phenate methods used to quantify dissolved inorganic nitrogen. Bachmann and Caneld (1996) have compared recoveries of total N minus NON from the sec3 ond derivative method (the equivalent of Kjeldahl N) with the EPA-approved Kjeldahl method. They found that, for the analysis of lake water samples, the second derivative method was as good or better in terms of accuracy and precision as the Kjeldahl method. They also found the new method to be more ecient; their laboratory could routinely run 100 samples a day with the second derivative spectroscopy method while it formerly took 2 days to run 40 Kjeldahl N samples. Our limited comparison of wastewater samples analyzed for Kjeldahl N and total N minus NON obtained similar results, 3 with no signicant dierence between analytical techniques. We also found that we could rapidly process and analyze a similar number of samples with the second derivative method.
CONCLUSIONS

simple, rapid technique for both NO and total N 3 analysis. The method eliminates many of the disadvantages associated with other analysis techniques, and accuracy and precision are not compromised. The analysis technique can be used to analyze domestic wastewater and other freshwater samples containing a range of inorganic and organic N concentrations. Future work should extend the utility of the method to other wastewaters.

REFERENCES

The second derivative spectroscopy method was found to be accurate, based on comparison with quality control check samples of known concentrations. The method also compared favorably with currently accepted methods for analysis of NO 3 and total N in wastewater samples. Our results agree with and extend the ndings of previous studies testing the accuracy of the second derivative spectroscopy method. The second derivative spectroscopy method is a

APHA (1998) Standard Methods for the Examination of Water and Wastewater, 20th ed. American Public Health Association, Washington, DC. Bachmann R. W. and Caneld Jr D. E. (1996) Use of an alternative method for monitoring total nitrogen concentrations in Florida lakes. Hydrobiologia 323, 18. Crumpton W. G., Isenhart T. M. and Mitchell P. D. (1992) Nitrate and organic N analyses with second-derivative spectroscopy. Limnol. Oceanogr. 37, 907913. Minitab (1996) Minitab for Windows Release 11, Minitab Inc., State College, PA. Simal J., Lage M. A. and Iglesias I. (1985) Second derivative ultraviolet spectroscopy and sulfamic acid method for determination of nitrates in water. Journal of the Association of Analytical Chemists 68, 962964. Suzuki N. and Kuroda R. (1987) Direct simultaneous determination of nitrate and nitrite by ultraviolet secondderivative spectrophotometry. Analyst 112, 10771079. USEPA (1983) Methods for chemical analysis of waters and wastes. U.S. Environmental Protection Agency, EPA-600/4-79-020, Cincinnati, Ohio. Wollin K. M. (1987) Nitrate determination in surface waters as an example of the application of UV derivative spectrometry to environmental analysis. Acta Hydrochim. Hydrobiol. 15, 459469.