Aqueous Nitrite Ion Determination by Selective Reduction and Gas

Andrew J. Dunham, Robert M. Barkley, and Robert E. Sievers. Anal. Chem. , 1995, 67 (1), pp 220–224. DOI: 10.1021/ac00097a033. Publication Date: Janu...
0 downloads 0 Views 603KB Size
Anal. Chem. 1995,67,220-224

Aqueous Nitrite Ion Determination by Selective Reduction and Gas Phase Nitric Oxide Chemiluminescence Andrew J. Dunham,t Robert M. Barkley, and Robert E. Severs*

Department of Chemistty and Biochemistry and CIRES. Campus Box 215, University of Colorado, Boulder, Colorado 80309-0215

An improved method of flow injection analysisfor aqueous

nitrite ion exploits the sensitivity and selectivity of the nitric oxide (NO) chemiluminescence detector. Trace analysis of nitrite ion in a small sample (5-160 pL) is accomplished by conversion of nitrite ion to NO by aqueous iodide in acid. The resulting NO is transported to the gas phase through a semipenneable membrane and subsequently detected by monitoring the photoemission of the reaction between NO and ozone (03). Chemiluminescence detection is selective for measurement of NO, and, since the detection occurs in the gas-phase, neither sample coloration nor turbidity interfere. The detection limit for a 100-pL sample is 0.04 ppb of nitrite ion. The precision at the 10 ppb level is 2% relative standard deviation, and 60-180 samples can be analyzed per hour. Samples of human saliva and food extracts were analyzed; the results from a standard colorimetric measurement are compared with those from the new chemiluminescencemethod in order to further validate the latter method. A high degree of selectivity is obtained due to the three discriminating steps in the process: (1) the nitrite ion to NO conversion conditions are virtually specific for nitrite ion, (2) only volatile products of the conversion will be swept to the gas phase (avoiding turbidity or color in spectrophotometric methods), and (3)the NO chemiluminescence detector selectively detects the emission from the NO 0 3 reaction. The method is free of interferences, offers detection limits of low parts per billion of nitrite ion, and allows the analysis of up to 180pLsized samples per hour, with little sample preparation and no chromatographic separation. Much smaller samples can be analyzed by this method than in previously reported batch analysis methods, which typically require 5 mL or more of sample and often need chromatographic separations as well.

+

Analysis of nitrite ion is important due to its role in environmental processes and its toxicity and suspected carcinogenicity in humans, as well as its normal function in humans. Measure ment of nitrite ion has recently been examined as a surrogate for in vivo nitric oxide production.'-3 Sensitive and selective methods are needed to measure nitrite ion at low levels in the complex ' Current address: Nalco Chemical Co., One Nalco Center, Naperville, IL 60563-1198. (1) Termin, A: Hoffmann, M.: Bing, R J. Life Sci. 1992,51, 1621-1629. (2) Archer, S. FASEB]. 1993,7, 349-360.

220 Analytical Chemistry, Vol. 67,No. 1, January 1, 1995

matrices found in water, foods, and biological fluids. The wide variety of analytical techniques has been reviewed recently: Chemiluminescence detection method$-9 are based on the conversion of nitrite ion to NO. For the analysis presented here, iodide ion, aciditied with dilute phosphoric acid, was used as the reducing agent. The NO is produced in the liquid phase and then transferred to the gas phase at reduced pressure for detection by chemiluminescence from reaction with ozone. The conversion of nitrite ion to NO and the transport of the NO produced are reproducible. This transfer has been accomplished by purging the solutions with an inert gas6-9 or by stripping NO from the aqueous phase through a semipermeable membrane.5 EXPERIMENTAL SECTION Apparatus. A schematic diagram of the flow injection analysis (FIA) system is presented in Figure 1. The reagent used for conversion was pumped by a Model 1330 HPLC pump @io-Rad, Hercules, CA). All connecting tubing for the liquid stream was l/lbi.0.d. x 0.034in. i.d. PTFE. Afreshly prepared solution of 0.1 M sodium iodide and 0.1 M HaP04(aq) was pumped at a rate of 1.0 mL/min. The samples or nitrite ion standards were injected into this stream using a Model 7125 fixed volume (20-160-pL loops) injection valve (Rheodyne, Cotati, CA). The reaction coil was 1.6" 0.d. x 0.220-mm i.d. x 500-mm Polysil capillary tubing (Supelco, Bellefonte, PA) at ambient temperature (usually 23 "C). Use of the coil allowed suflicient time for reaction to occur between injection and permeation. During temperature experiments, the reaction coil was thermostated by immersion in a Model D2 water bath (Haake, Paramus, NJ) . "he semipermeable membrane which allowed NO to be stripped from the liquid phase was 1.6" 0.d. x 1-mmi.d. Gore Tex tubing with 2-pm pores (Anspec) enclosed in a glass cell. Connections were made between the reaction coil and the porous tubing with flange fittings in a block of poly(chlorotrinuoroethy1ene) plastic (Kel-F) (Figure 1). Glass tubing side arms (1/4in. 0.d.) on the permeation cell provided an inlet for pure air or He as a sweep gas and an outlet through which the permeant can be (3) Misko, T. P.; Schilling, R J.; Salvemini, D.; Moore, W. M.; Cume, M. G.

Anal. Biochem. 1993,214,11-16. Dunham,A J. Nitric Oxide Chemiluminescence and its Application to Nitrite Analysis in Complex Systems. Ph.D.dissertation, University of Colorado at Boulder, Boulder, CO, 1994. (5) Aoki, T. Biomed. Chromafogr. 1990,4,128-130. (6) Braman, R S.; Hendrix, S. A Anal. Chem. 1989,61, 2715-2718. (7)Thornton, A R; F'fab, J.; Massey, R C. Analyst 1989,114,747-748. (8) Garside, C. Mar. Chem. 1982,11, 159-167. (9) Cox, R D. Anal. Chem. 1980,52,332-335. (4)

0003-2700/95/0367-0220$9.00/0 0 1994 American Chemical Society

permeation Integrator sweep gas Aqueous

solution

De tecto

Capillary react ion coil

Liquid waste

Glass

Flgure 1. 1 : Schematic diagram of aqueous nitrite ion conversion and detection apparatus. Cell containing semipermeable membrane for stripping NO(aq) from the liquid phase to the gas phase.

drawn into the NO chemiluminescence detector. A single bead string reactorlo (SBSR), inserted coaxially into the porous tube, facilitated transfer of NO from the liquid to the gas phase. The SBSR was prepared from 10 pound test, Nylon fishing line; overhand knots were tied approximately 1 mm apart along the length of string. When the SBSR was placed inside of the porous tubing, the outside of the knots approached closely the inside perimeter of the tubing. An early Model 207 redox chemiluminescence detector (Sievers Instruments, Inc., Boulder, CO) was used in most of the experiments for the detection of NO (bypassing and removing the catalyst bed). Recent experiments have shown that use of the late Sievers Instruments Model 270 NOA provides even better detection limits (0.04 ppb vs 1 ppb of nitrite). For optimum sensitivity, the nitric oxide analyzer should be specifically designed for low flow rates. A Model 3390A integrator (Hewlett-Packard, Palo Alto, CA) was used to collect and report the data as peak areas. A Hewlett-Packard Model 845lA diode array spectrophotometer was used for all spectrophotometric measurements. Reagents and Chemicals. Sodium nitrite (Aldrich Chemical Co., Inc., Milwaukee, wr) was used to prepare standard solutions. Sodium chloride (Mallinckrodt, Inc., Chesterfield, MO), i r o n 0 sulfate heptahydrate (Fisher Scientitic, Fair Lawn, NJ), sodium iodide (Mallinckrodt), and concentrated phosphoric acid (Mallinckrodt) were used in the preparation of conversion reagents. Sulfanilamide (Aldrich), hydrochloric acid (Mallinckrodt), and N-(1-naphthyl)ethylenediamine dihydrochloride (Eastman, Rochester, NY) were used to prepare the diazotization reagent solutions for the spectrophotometric method. A standard Fe(II) solution was prepared from ferrous ammonium sulfate (Thorn Smith, Beulah, MI); potassium permanganate (Merck, St. Louis, MO) was used to standardizethe nitrite ion solution. All reagents were used as received. Reagent solutions were prepared immediately (10)Sunden, T.;Cedergren, A; Siemer, D. D. Anal. Chem. 1984, 56, 10851089.

before use. Deionized, distilled water was used for all solution preparations. Samples. Saliva samples were collected immediately prior to analysis by expectoration directly into clean test tubes. The samples were centrifuged and the resulting supernatant was decanted from each sample and diluted with water for analysis. Turkey hot dogs (Longmont Foods) and lettuce (Dole) were purchased at a local grocery store and extracted.11 The extraction was performed immediately prior to the analysis, so preservation was not necessary. Procedures. (i) Standardization. The sodium nitrite stock solution was mixed with an excess of potassium permanganate solution that had been standardized by titration with standard ferrous ammonium sulfate. The remaining permanganate was consumed by an excess of standard ferrous ammonium sulfate. The excess iron was back-titrated with standard potassium permanganate.12 NO Standards. Aqueous solutions of NO were prepared as follows. Helium passed through an 02 trap was used to remove oxygen from the aqueous scrubbers and water sample flask prior to exposure to NO. Pure NO from a cylinder was bubbled through an aqueous 1 M sodium hydroxide impinger to remove any NO2 present. A gas phase NO standard (5.1 ppm in Nz) was used as received (Air Products and Chemicals, Inc., Allentown, PA). The standard was introduced directly to the NO chemiluminescence detector via a fixed volume gas injection valve. The pressure of the chemiluminescence reaction cell in the detector was controlled to less than 15 Torr with a needle valve between the injection valve and the instrument. Either pure air or He was used as the sweep gas through the injection port. Comparisons between either aqueous nitrite ion or gas phase standards and samples were made from a calibration curve based on standard peak areas. The total analysis time for each sample or standard injection was on the order of 1 min. Safe& Considerations. The release of NO into air leads to the rapid generation of nitrogen dioxide RJOZ), which is extremely toxic. Concentrated streams of NO (g) should be used with due caution in leak-free gas systems contained in a well-ventilated hood. Access to a self-contained breathing apparatus should be available where large volumes of NO@) are stored or used. (ii) Comparison with Diazotization/UV Spectrophotometric Measurements. The standard method used for comparision to the FIA method is based on the diazotization of nitrite ion followed by spectrophotometric analysis.12 A 50.@mLaliquot of a sample or standard was placed in a volumetric flask. One milliliter of sulfanilamide solution was added to each standard or sample. After 2 min, 1.0 mL of N-(1-naphthy1)ethylenediamine solution was added. The samples and standards were analyzed after at least 8 min, but not longer than 2 h. The response from the spectrophotometerwas zeroed with a blank solution prepared from 50.0 mL of distilled, deionized water, and 1.0 mL of each derivatization reagent. The solutions were then analyzed by spectrophotometry (A,,-,== 540 nm). Sample absorbances were compared to a calibration curve based on absorbances of standard solutions of nitrite ion. (11) Welcher, F. J. In Stundard Methods of Chemical Analysis; Welcher, F. J., Ed.; Robert I. Krieger Publishing Co.: Malabar, FL, 1966. (12)Franson, M. A. Standard Methods For the Eramination of Water and Wustewatq American Public Health Association: 1015 Fifteenth St.NW, Washington, DC 20005,1975.

Analytical Chemistry, Vol. 67, No. 1, January 1, 1995

221

Table 1. Effect of pH on Nltrite Response by Iodide Reduction

reagent

0.1 M HsPOr(aq) 0.1 M KI(aq) 0.1 M NaI in 0.1 M CzH3OOH(aq) 0.1 M NaI in 0.1 M 0.1 M NaI in 0.1 M NaC2H302

2.0 x 106

r

1

'

O

1

average response (area counts), x 10-5 3.21 f 0.296a 5.86 f 1.88 1.70 f 0.439 1.94 f 0.593

NDb

fl standard deviation, n = 5. * None detected.

RESULTS AMD DISCUSSION The FIA system presented here employs a novel, low-volume, gas-permeable membrane which provides a means for rapid, quantitative transport of NO(ad) produced from the reduced sample. The NO@) can then be swept to a NO chemiluminescence detector, where it reacts with ozone in the gas phase ( ~ 5 0 Torr), generating electronically excited nitrogen dioxide. When the electronically excited nitrogen dioxide (NOz*) relaxes by photoemission, the resulting photons are counted by a photomultiplier tube. The dependence of nitrite ion response on the reaction conditions was determined by varying the reducing agent and type of reaction tubing, acid concentration, reducing agent concentration, chloride ion concentration, reaction temperature, and residence time. The efficiency of NO(ad) transport through the membrane was also studied and optimized.

Reaction Conditions. (i) Reducing Agents/Readion Tubing. TWO previously reported reducing agents9 were evaluated

for use as reagents in this flow injection analysis system: ferrous ion and iodide ion. Two reaction chambers were examined: crocheted PTFE13 and fused silica-lined PTFE (Polysil tubing, Supelco). Separate 0.1 M solutions of reducing agents were prepared in 0.1 M H3P04(ad) and evaluated for conversion to NO of a 2@pLsample of 1mM nitrite ion standard in crocheted PTFE. The Fe(II) solutions yielded sharp peaks, with an average peak width at half-height of approximately 12 s. While the iodide reagent stream effected a substantial conversion of nitrite ion to nitric oxide, the peaks showed unacceptable tailing and were significantly broadened (average peak width at half-height of approximately 38 s). The broadening and tailing of the peak resulted in an increase of the sample analysis time from 1 to 5 min in PTFE. However, in fused silica-lined PTFE, the iodide reducing solution generated sharp peaks with average peak areas greater than those produced by the Fe(II) solution. Since combination of the use of an iodide reducing agent together with selection of the fused silica-lined PTFE reaction tubing yielded the highest response and the sharpest peaks, they were chosen for this method. (ii) Iodide Reagent pH. Acid concentrationstested were kept below 0.1 M due to the potential damage that could occur in the HPLC pump and injection valve by prolonged use of more acidic solutions. A 50 pM standard solution of nitrite ion was prepared in deionized distilled water and analyzed using the following reagent solutions: 0.1 M iodide in water, in 0.1 M acetic acid, in 0.1 M H3P04, and 0.1 M in sodium acetate. Responses from the 50 pM nitrite ion standard are compared in Table 1. The reducing solution containing 0.1 M iodide and 0.1 M HsPOd(aq) yielded (13) Birks, J. W.; Poulsen,J. R;Birks, K S.;Gandelman, M.S. Chromatographia 1986,22, 231-234.

222 Analytical Chemistry, Vol. 67, No. 1, January 1, 1995

0.0

'

I 0.2

0.0

0.6

0.8

1

1.2

[NaI] in 0.1 M phosphoric acid

Figure 2. 2: Effect of iodide concentration on nitrite ion standard response. Responses correspond to repetitive injections of a 100 ppb nitrite standard.

the strongest response. The acidity is clearly important, because the aqueous solution containing 0.1 M iodide and 0.1 M sodium acetate did not produce a detectable signal, when an attempt was made to analyze the 50 pM nitrite ion standard solution. Consequently, phosphoric acid (0.1 M) was used in the remainder of the analyses. (iii) Reducing Agent Concentration. Sodium iodide solutions were prepared in a range of concentrations from 0.001 to 1.0 M in 0.1 M HsPOl(aq). Standards were injected in triplicate into the reagent streams. As the iodide concentration increased, the signal increased (Figure 2). There was no increase in the peak-tepeak baseline noise. However, upon standing, the solution of 1.0 M sodium iodide began to turn yellow over the course of a few hours. This solution was unstable due to oxidation by air, forming iodine. The resulting iodine appeared to wet the perme able membrane. The solution of 0.1 M sodium iodide and 0.1 M H3P04 was stable for at least 8 h and did not wet the membrane, so this reagent was used for the remainder of the study. (iv)Reaction Temperature. The temperature of the reaction coil was controlled with a water bath to 4~0.5"C of the indicated value. Standards containing 10pM nitrite ion were injected with the reaction coil temperature varied from 22 to 73 "C. In order to prevent an increase in detector noise due to increased water vapor transport through the porous membrane with increased temperature, a section of tubing between the reaction tube and the porous membrane was cooled in a water bath at 10 "C. As the reaction coil temperature increased, the response of the standard did not change sigdicantly. Therefore, subsequently no temperature control was used; the conversion reaction was performed at room temperature, which was always 21 "C or higher. (v) Potential Interferences. In order to test the effect of commonly encountered, potential interferences, separate 10%(w/ w) solutions of Nafi04, KN03,NaCl, NH4C1, and 1%(w/w) solutions of caffeine and sodium glutamate were prepared and analyzed. AU of the solutions tested yielded very small responses. The responses were small enough to afford more than lo6 selectivity for detection of nitrite ion over these other compounds. Sulfanilamide has been used to effectively remove nitrite ion from solutions that will be treated with strong acid and iodide ion for conversion to NO.14 In order to determine whether the response for these compounds was due to a nitrite ion contamination, 0.6 M sulfanilamide in hydrochloric acid was added to each

sample. No responses were observed for sulfanilamide-treated samples, suggesting that each contained minute levels of nitrite, rather than producing a from the Other compounds' Sulfanilamide or other species that react with nitrite obviously constitute potential negative interferences in nitrite analysis. The conversion of nitrite ion to NO requires the formation of nitrous acid. If the sample is strongly alkaline or has a significant buffering capacity so that in the reagent solution, e.g., 0.1 M iodide ion and 0.1 M HsPOd(aq), the nitrite ion in the sample is not protonated, then the pH of the sample must be adjusted to approximately 5 to yield accurate results. The presence of high concentrations of sulfite ion in samples presents a special problem in nitrite ion analysis. In acidic solution, sulfite and nitrite ions react to form hydroxylamine sulfonates.15 It was observed that nitrite ion was not detected

+ HSO,- - H,O + HONS0,HONSO,- + HSO,- - HON(SOJ,~HNO,

when added to samples containing sulfite ion at pH 5. This indicated that the hydroxylaminesulfonateis not detectable using the chemiluminescent method for nitrite ion. If the analysis of nitrite ion is required in the presence of sulfite ion, then a method in which a high pH is maintained would be required, such as some ion chromatographic methods. (vi) Residence Time. The effect of residence time was determined by varying the flow rate of the reagent solution and monitoring the change in response for a nitrite ion standard. As the flow rates were increased from 1.0 to 5.0 mL/min, both peak areas and widths decreased. At lower flow rates, no signiiicant increase in peak areas was observed, while peak widths increased. A reagent flow rate of 1.0 mWmin resulted in peak areas as large as at any flow rate and a short enough residence time that peaks were sharp, allowing a large number of samples to be analyzed per hour. (vii) Efficiency of NO Liquid to Gas Transport. A key step in the analysis of nitrite ion by this method involves the transport of NO produced in the liquid phase to the gas phase for detection. The semipermeable membrane used in the transport had an internal diameter of 1 mm. If laminar flow and a diffusion coefficient for water of 10-5 cm/s are assumed, the time required for diffusion of an aqueous NO molecule from the center to the surface of the tube for permeation is estimated to be about 2 min. The combination of semipermeable membranes and a mass transport device (such as a SBSR) have been used successfully for stripping carbon dioxide from solution in suppressed ion chromatography.'O The SBSR forces more NO (as) to contact the tubing walls by reducing dead volume, breaking up laminar flow, and introducing turbulence. An additional advantage of engendering secondary flow characteristics is a reduction in band broadening.16 The most benefit from the SBSR is realized in enhanced permeation of NO through the membrane and not through improved mixing in the reaction coil. Our observations support (14)Cox R D.; Frank,C. W.; Nikolaisen, L D.;caputo, R E.Anal. &em, 1982, 54, 253-256. (15)Oblath, s. B.; Markowitz, S. S.; Novakov, T.;Chang, S. G. J. Phys. Chem. 1981, 85, 1017-1023. (16)Reijn, J. M.;Van Der Linden, W. E.; poppe, H.A d , Chim. Acta 1981, 126,1-13.

Effwt of

Tubing Length on NO

ResmonSe

length of permeable average response in average response in tubing in first fist permeation cell second permeation cell permeation cell (cm) (area counts), x 10-6 (area counts), x 10-6 7.5 10 12

1.01 f 0.11" 1.14& 0.06 1.52 f 0.12

0.40f 0.18" 0.33 f 0.11 0.18 f 0.06

fl standard deviation, n = 5.

previous work which suggested that the efficient transport of a dissolved gas through a semipermeable membrane is controlled by the transport of the dissolved gas in the bulk liquid to the membrane and not by diffusion through the membrane.1° In addition to studying the effect of using a SBSR we investigated the effect of changing the length of the porous tubing on the intensity of the nitric oxide chemiluminescencesignal. With two permeation cells arranged in series, the length of porous tubing in the first cell was varied, while the length of the second cell was kept constant at 12 an. The NO responses from repetitive injections of a nitrite ion standard solution were monitored from both cells. Each semipermeable tube in the permeation cells contained a SBSR As the length of the permeable first tubing was increased from 7.5 to 12 cm, the magnitude of the NO response from the first permeation cell increased from 74%to 90% of the total NO measured from both cells, while the responses from the second cell decreased with longer lengths of porous tubing in the first cell (Table 2). Thereafter, the length of porous tubing used in the single stage permeation cell was fixed at greater than 15 cm. The efficiency of NO transport from the aqueous phase to the gas phase was estimated for the permeation cell and a commercially available liquid sample sparging apparatus (Sievers). The responses observed for the aqueous NO standard using each apparatus were compared to the signals from the injections of a gas phase NO standard that was 5.1 ppm in Nz. A saturated, aqueous NO standard was prepared as above and injected repetitively into the sparging apparatus and into the FIA system using distilled water as the liquid carrier at a flow rate of 1.0 mL/ min. The concentration of NO in the saturated solution was calculated from the Henry's law constant, and this was compared to measured values of NO that passed through the membrane. The injections of gas phase NO standards allowed for mass calibration of the NO chemiluminescence detector; comparison with the responses of the aqueous NO standard then led directly to estimates of permeation efficiencies. Stripping efficiencies of 79% and 98% were calculated for the liquid sample sparging apparatus and for the permeation cell used in FIA, respectively. Analysis of Nitrite Ion in Saliva and Food Extract Samples of human saliva, lettuce, and turkey hot dog were analyzed by the FIA method and by spectrophotometric analysis of the dye formed via the diazotization of nitrite ion in the samples with sulfanilamide and N-(l-naphthyl)ethylenediamine.12 The lettuce sample contained 0.19 & 0.03 ppm (w/w) of nitrite ion, which was below the detection limit Of the spectrophotometric method (Table 3). However, our chemiluminescence method detected and measured the presence of nitrite ion in all of the samples. The E S U l t s of the independent measurements of the replicate samples were in good agreement, especially considering the Analytical Chemistry, Vol. 67, No. 1, January 1, 1995

223

Table 4. Concentratlon and Mass Detection Llmits for NltrHe Ion by Various Methods

Table 3. DIasotlzationRIV Compared with Chemiluminescence Results

sample saliva (ppm nitrite ion) hot dog extract (ug nitrite ion/ g food) lettuce extract (ug nitrite ion/ g food) a

diazotization/

uv"

35.0 f 0.8 17.7 f 0.4

chemiluminescencea 34.5 f 0.9 19.8 f 0.2

NDb

0.19 f 0.03

*

Nitrite ion concentration, il standard deviation. Not detected.

complex nature of the matrices (e.g., hot dog extract and saliva). Only ca.3 drops of aqueous sample are needed in the new method. Calibration, Precision, and Detection Limits. The results of linear regression of the standard curves analyzed by the W and FIA methods are y = 0.001l.t - 0.0333 and y = 24300~(5.15 x 1@),respectively. Responses for the methods were AU and area counts, for the W and FIA method, respectively, and the concentration units were parts per billion of nitrite ion. The precisions, expressed as relative standard deviations, at the 50 ppb level for the FJA and W methods were 2% and 8.5%, respectively. Detection limits for the two methods were based on data generated in this study and are 3.29 times the root-meansquare noise. Two measures were used in comparing the detection limits: the concentration (ppb) and the mass (ng) of nitrite ion. The detection limits for the two methods employed here, as well as previously reported methods, are shown in Table 4. CONCLUSIONS Nitrite ion concentration levels in saliva and in food samples were analyzed using a flow injection method. This method is based on the conversion of nitrite ion to NO using iodide ion in acid, on-line transfer of the NO formed across permeable Teflon, and NO detection by gas phase NO chemiluminescence. The advantages over standard spectrophotometric and previously reported methods are as follows: the FIA chemiluminescence (17)Anderson, L.Anal. Chim. Acfa 1979,110, 123-128.

224 Analytical Chemistry, Vol. 67, No. 1, Januaty 1, 1995

method

mass concentration detection sample size detection limit K t (ng needed (mL) (ppb nitrite ion) rutnte ion)

flow injection analysis with chemiluminescence (this report) derivatizationwith spectrophotometry12 static nitrite ion conversion with NO chemiluminescenceg flow in'ection analysis by derivatization and ~pectrophotometry~~ a

0.10 50.0 20.0 0.2

0.005

0.05a 25"

1300

0.05b

1.0

2

0.5

Observed in this study. Reported in ref 9.

method does not suffer from the color and turbidity interferences of the spectrophotometricmethod. S i c a n t l y smaller samples can be analyzed by the new method than by previously reported chemiluminescent techniques without any loss of sensitivity. The rapidity of this method allows the analysis of 60-180 samples/h with little or no sample preparation, and this lends it well to online automation. This method should be particularly useful in analyzing biological fluids and in studying biogeochemical cycling of nitrogen in the environment. ACKNOWLEWMENT Funding for this research has been provided by the National Science Foundation (Grant ATM-9115295) and by NASA through the NASA Specialized Center for Research and Training/Center for Space Environmental Health. The loan of a liquid sample gas sparging apparatus and a Model 270 nitric oxide analyzer from Sievers Instruments, Inc. is appreciated. Received for review May 27, 1994. Accepted October 3, 1994.@ AC940548M Abstract published in Advance ACS Abstracts, November 1, 1994.