Illinois Fertilizer
Conference Proceedings
January 26-28, 1998
| Home | 1998 Index | Search |
|
|
| Home | 1998 Index | Search |
R.L. Mulvaney and S.A. Khan1
The determination of NH4+, NO3-, or NO2- in biological materials having a high content of organic matter presents special difficulties, owing to the potential for interference by organic forms of N. Such interference is apt to arise through hydrolysis of amino sugars and other nitrogenous compounds that are labile under alkaline conditions.
Colorimetric methods are used routinely for quantitative determination of NH4+, NO3-, or NO2- in natural waters and wastewater (EPA, 1983; APHA, 1992). This approach tends to be more successful for NO3- or NO2- than for NH4+, since color development in the latter case requires a high pH and is subject to organic interference (Rhine et al., 1998). A further limitation arises in the use of colorimetric methods for analysis of colored or turbid samples, and from the fact that these methods do not permit isotope-ratio analysis in 15N-tracer investigations or the measurement of natural variations in N-isotope abundance.
Alternatively, inorganic-N analyses can be done by steam distillation or diffusion, in which case both quantitative and 15N analyses can be accomplished in a single operation. To minimize the risk of organic interference, a mild alkali such as MgO (pH 10.5) is commonly employed in steam distillation and diffusion methods for inorganic-N analysis of soil extracts (Mulvaney, 1996), and a borate buffer at pH 9.5 has been recommended for distillation of NH4+-N in natural water and wastewater (EPA, 1983; APHA, 1992).
The use of diffusion would appear to have a fundamental advantage over steam distillation for inorganic-N analysis of organic-laden samples, in that the analysis is performed at a considerably lower temperature, potentially reducing hydrolysis of labile organic-N. Recent work in our laboratory has led to simple, economical, and reliable diffusion methods for inorganic-N and 15N analyses of soil extracts and water (Khan et al., 1997a,b; Mulvaney et al., 1997a). In these methods, diffusion is carried out in a 1-pint wide-mouth Mason jar, either at room temperature (Mulvaney et al., 1997a) or with gentle heating on a hotplate to markedly reduce the period required for the analysis (Khan et al., 1997b).
The primary purpose of the work reported here was to modify existing Mason-jar diffusion methods, so as to permit quantitative and 15N analyses of inorganic N in the presence of a high concentration of alkali-labile organic N. The modified methods were evaluated for analytical accuracy and precision based on recovery of inorganic 15N added to a wide range of samples, and also by comparison to conventional steam distillation techniques.
The samples studied (Table 1) included a test mixture of known composition, plus five other samples collected to obtain a wide range in soluble organic C.
Test Mixture: A solution containing 10 g of N L-1 was prepared by dissolving a sufficient quantity of the following compounds in deionized (DI) water, such that each supplied 1 g of N L-1: arginine HCl, citrulline, galactosamine HCl, glucosamine HCl, glutamine, glycine, histidine, lysine HCl, and methionine. The compounds used to prepare this mixture were obtained from Sigma Chemical Co., St. Louis, MO, or Fisher Scientific, Pittsburgh, PA. The purity of each reagent exceeded 98%.
Urine: The sample used was of human origin and was obtained immediately prior to use.
Manure Extract: A fresh sample of cow (Bos taurus) manure was collected from the University of Illinois Beef Farm. Approximately 135 g of fresh manure (350 g of H2O kg-1) was shaken with 1 L of deionized water in a 2-L bottle for 45 min. An extract was obtained by filtering the resulting slurry through cheesecloth.
Septic Effluent: A sample of effluent was obtained from a domestic septic system. Before use, the sample was passed through a Whatman GF/C glass-fiber filter under vacuum.
Feedlot Leachate: Tile flow originating from the University of Illinois Dairy Science Research Farm was sampled at an outlet located approximately 0.5 km from the facility.
River Water: The sample used was obtained from the Sangamon River at Allerton Park, Monticello, IL. Before use, the sample was filtered under vacuum through a Whatman GF/C glass-fiber filter.
Diffusions were performed with MgO to recover NH4+-N, or with MgO and Devarda's alloy to recover (NH4+ + NO3- + NO2-)-N, using the Mason-jar methods described by Mulvaney et al. (1997a) and Khan et al. (1997a,b). Studies to evaluate the existing Mason-jar methods for interference by labile organic N indicated a need to minimize the diffusion period. Accordingly, the Mason-jar methods were modified as follows:
| - 0.9 g for analyses at 45-50°C | |
| - 1.5 g for analyses at 25°C |
| -18 h at 25°C | |
| -1.5 h at 45-50°C (hotplate) |
Steam distillations were performed as described by Bremner and Keeney (1965), taking care to ensure that the distillation period did not exceed 4 min. In the methods employed, distillation was done with MgO to recover NH4+-N, or with MgO and Devarda's alloy to recover (NH4+ + NO3- + NO2-)-N.
In cases involving diffusion of 15N-treated samples, processing for isotopic analysis was carried out as previously described (Khan et al., 1997a,b; Mulvaney et al., 1997a). Isotopic analyses were performed using a Nuclide Model 3-60-RMS mass spectrometer (Premier American Technologies Corp., Bellefonte, PA) equipped with an automated Rittenberg system (Mulvaney and Liu, 1991; Mulvaney et al., 1997b).
Inorganic-N analyses are complicated considerably by the presence of labile organic N, particularly if the analyses are carried out with prolonged heating at a high pH, which promotes hydrolysis. This problem can lead to serious interference during analyses of organic-laden samples by steam distillation, owing to the high temperatures required.
The Mason-jar diffusion methods recently developed in our laboratory (Khan et al., 1997a,b; Mulvaney et al., 1997a) are a convenient alternative to steam distillation for inorganic-N analysis of soil extracts and water, and because diffusion is performed at a much lower temperature than is distillation, organic interferences should be minimized. Previous work to evaluate the Mason-jar methods for interference by two alkali-labile organic N compounds (300 µg of N as glucosamine or glutamine) showed very little, if any, decomposition of these compounds during diffusions from 10 or 20 mL. Up to 22% of the organic N was liberated as NH4+ when diffusions were performed from 50 or 100 mL at 25°C, owing to the longer period required for the analysis, and a further increase occurred when diffusion was carried out at 30°C (Mulvaney et al., 1997a). As part of the present project, the existing Mason-jar methods were evaluated more rigorously for organic interference, using a concentrated test mixture (10 mg N mL-1) consisting of eight amino acids and two amino sugars, selected as the most decomposable among 25 labile organic N compounds. The results of this evaluation are summarized by Table 2, which confirms the advantage of a small sample volume in minimizing the diffusion period. Table 2 shows that interference can also be reduced by carrying out diffusions from 4 M KCl instead of DI water, which can be attributed to a salt effect in promoting the liberation of NH4+ and thereby reducing the diffusion period (Mulvaney et al., 1997a).
Besides reducing the volume of sample and adding KCl, diffusion can be accelerated by increasing the volume and/or the concentration of H3BO3-indicator solution employed (Mulvaney et al., 1997a). All three modifications have been utilized in minimizing the diffusion period for analysis of inorganic N in the presence of labile organic N; diffusions are performed from 10 mL of sample after addition of sufficient KCl to produce a 4 M concentration, using 7 mL of 4% H3BO3-indicator solution rather than the 4-5 mL specified previously (Khan et al., 1997a,b; Mulvaney et al., 1997a). The benefit of these modifications is apparent from Table 3, which shows no interference from the presence of 10 mg of labile organic N, regardless of whether diffusions were performed at room temperature or with gentle heating on a hotplate.
To avoid the risk of interference by organic N, care should be taken to ensure that the diffusion periods specified are not exceeded. This is illustrated by Figure 1, which shows the results obtained when diffusions were performed with or without heating on a hotplate to recover labeled NH4+-N added to aliquots of the test mixture or river water. In both cases, prolonging the diffusion period led to an increase in the quantity of NH4+-N recovered, and to a decrease in 15N enrichment. Both trends can be attributed to mineralization of unlabeled organic N.
In diffusion or distillation methods of inorganic-N analysis, Devarda's alloy is used to convert NO3--N or NO2--N to NH4+-N. This material is a mixture of 50 parts Cu, 45 parts Al, and 5 parts Zn. Originally, 0.2 g of alloy was recommended for steam distillation (Bremner and Keeney, 1965), and this same addition has been found to be adequate in previous Mason-jar diffusion methods for inorganic-N analysis of soil extracts. Table 4 shows that a much larger addition of alloy was essential to ensure quantitative recovery of NO3--N added to the test mixture, which can be attributed to the presence of reducible organic functional groups and/or to metal chelation. Based on the data in Table 4, 1.5 g of Devarda's alloy is recommended when diffusions are performed at room temperature. If diffusions are to be done on a hotplate, 0.9 g suffices because heating increases the alloy's reactivity.
Table 5 summarizes the results obtained when inorganic-N analyses were performed on the six samples described previously, by the diffusion methods developed in our work and also by conventional steam distillation techniques (Bremner and Keeney, 1965). The data in Table 5 reveal that, with each of the samples studied, a substantial difference occurred between analyses by diffusion and distillation. For example, the test mixture contained no inorganic N, and whereas none was detected by diffusion, substantial amounts of NH4+-N and (NH4+ + NO3- + NO2-)-N were recovered by distillation, indicating hydrolysis of labile organic compounds. In most of the other comparisons, lower values were obtained by distillation than by diffusion, and with some samples, negative values were obtained by distillation after correction for the control. The latter finding can probably be attributed to liberation of acidic constituents during distillation.
In a further comparison of distillation and diffusion, analytical accuracy and precision were evaluated by checking the recovery of NH4+-N, NO3--N, or NO2--N added to each type of sample under study. The results are summarized in Table 6, which shows that diffusion was far superior to distillation. Recoveries by diffusion ranged from 96 to 107%, and the coefficient of variation (4 replicate analyses) was usually < 2%. By comparison, recoveries by distillation were almost always incomplete, and replicate analyses were highly variable (the coefficient of variation usually exceeded 10%). The latter problems were much more serious for NO3--N or NO2--N than for NH4+-N. This can be attributed, at least in part, to incomplete reduction by Devarda's alloy, as distillations were performed according to the conventional methods described by Bremner and Keeney (1965), which use 0.2 g of a finely-ground alloy. Surprisingly, distillation led to incomplete recovery of inorganic N added to river water having a low C content (Table 1), particularly in cases involving addition of NO3- or NO2-. This finding has significant implications, given the fact that distillation has been widely employed in carrying out N-isotope analyses to ascertain the origin of NO3- in surface and ground waters (e.g., Kohl et al., 1971).
Like previous Mason-jar diffusion methods, the methods developed in our work allow isotope-ratio analysis of diffused NH3-N in 15N-tracer investigations. To evaluate the accuracy and precision of 15N analyses by the diffusion methods described, diffusions were performed from aliquots of each of the samples described previously that had been treated with a known amount of 15N-labeled NH4+, NO3-, or NO2-. The results (Table 7) verify that the diffusion methods described permit accurate and precise isotope-ratio analyses of inorganic N in a complex organic matrix. Moreover, the close agreement between the expected and measured values in Table 7 provides conclusive evidence that, when these methods are performed as described, no interference arises from labile organic N.
A major concern in the analysis of organic-laden samples is that chemical or biological processes during storage or transport may affect their chemical composition. Figure 2 confirms the need to refrigerate such samples prior to inorganic-N analyses by the methods described, and also verifies the importance of carrying out analyses as soon as is feasible. The data in Figure 2 indicate that highly carbonaceous samples should be analyzed within 12 h after collection, whereas natural water or other samples that have a low content of organic C can be stored safely for several days if properly refrigerated. Storage without refrigeration is not recommended, as the inorganic-N content may either increase or decrease (Figure 2).
Serious interference can arise when inorganic-N analyses are performed on samples having a high content of labile organic N. Studies to evaluate existing Mason-jar diffusion methods for interference by a wide variety of amino acids and amino sugars showed the need to minimize the diffusion period. Modifications were made to eliminate interference by the aforementioned compounds. In the modified methods, up to 10 mL of sample is treated with sufficient KCl to produce a 4 M concentration, and diffusion is carried out using 7 mL of 4% H3BO3-indicator solution, either at 25°C for 18 h, or for 1.5 h by heating on a hotplate at 45-50°C. The methods are simple and inexpensive, and permit both quantitative and isotopic analyses of inorganic N in animal wastes and other materials containing a large amount of labile organic N.
Table 1. Chemical properties of samples studied
Table 2. Evaluation of existing Mason-jar methods for interference by 10 mg of labile organic N
Table 3. Evaluation of modified Mason-jar methods for interferance by 10mg of labile organic N
Table 5. Comparison of distillation and diffusion for inorganic-N analysis of complex samples.
1R.L. Mulvaney is a Professor and S.A. Khan is a Research Associate, Dept. of Natural Resources and Environmental Sciences, University of Illinois.
