Determination of nitrite in waters by microplate fluorescence spectroscopy and HPLC with fluorescence detection

Article abstract of DOI:10.1021/ac981330t

A selective and versatile fluorescence spectroscopic method for the determination of nitrite in waters has been developed. Nitrite reacts in the presence of mineral acids with the nonfluorescent N-methyl-4-hydrazino-7- nitrobenzofurazan forming N-methyl-4-amino-7-nitrobenzofurazan, which can be detected by fluorescence spectroscopy with an excitation maximum at λ = 468 nm and an emission maximum at λ = 537 nm in acetonitrile. Three new methods based on this reaction have been developed: Direct fluorescence spectroscopy, HPLC/fluorescence, or HPLC with UV/vis detector may be selected as detection techniques. On microplates, high-throughput fluorescence spectroscopy is achieved, while HPLC/fluorescence provides lower limits of detection, and HPLC with UV/vis detection enables evaluation of the reaction with standard instrumentation. Different water samples were investigated using all detection modes, and a photometric standard procedure was successfully employed to validate the new methods with an independent technique.

Full text of DOI:10.1021/ac981330t

Anal. Chem. 1999, 71, 3003-3007
Determination of Nitrite in Waters by Microplate
Fluorescence Spectroscopy and HPLC with
Fluorescence Detection
Andrea Bu1ldt and Uwe Karst*
Anorganisch-Chemisches Institut, Abteilung Analytische Chemie, Westfa¨lische WilhelmssUniversita¨t Mu¨nster,
Wilhelm-Klemm-Strasse 8, D-48149 Mu¨nster, Germany
established as a European standard.6 Its major drawbacks are
interferences from colored samples and insufficient limits of
detection for ultratrace determination of nitrite. Several fluores-
cence spectroscopic methods have also been published in recent
years.7-10 They are based on the reaction of nitrite with organic
fluorogens to form highly fluorescent compounds. Some of these
methods are based on a direct fluorescence reading;7-9 others
require HPLC separation of the product and the reagents.10 Lee
and Field11 published a method for postcolumn derivatization with
fluorescence detection. A major advantage of the fluorescence
methods is the low limit of detection, but their robustness cannot
compete with photometry for concentrations in the lower to mid-
ppb range which is typical for surface waters in agriculturally used
areas.
Kieber and Seaton12 developed a HPLC method with UV/ vis
detection based on the reaction of nitrite in acidic media with 2,4-
dinitrophenylhydrazine (DNPH), a well-known reagent for the
determination of aldehydes and ketones.13-15 2,4-Dinitrophenyl
azide (DNPA) is formed as a reaction product which is easily
separated from the hydrazone of formaldehyde. The NO+ cation
is the active species in this reaction, as demonstrated before
during investigations on chemical interferences by nitrogen
dioxide in the determination of carbonyls using the DNPH
method.13 Gro¨ mping et al. previously published a method, based
on this reaction, for the determination of nitrogen dioxide in air
samples.16 The method of Kieber and Seaton is characterized by
a low limit of detection at 0.1 µmol/ L (4.6 ppb) nitrite and a rapid
derivatization. The same group has recently reported on the
determination of 15N nitrate and nitrite using the same reaction
with subsequent FT-IR determination.17
A selective and versatile fluorescence spectroscopic method
for the determination of nitrite in waters has been devel-
oped. Nitrite reacts in the presence of mineral acids with
the nonfluorescent N-methyl-4 -hydrazino-7 -nitrobenzo-
furazan forming N-methyl-4 -amino-7 -nitrobenzofurazan,
which can be detected by fluorescence spectroscopy with
an excitation maximum at λ ) 4 6 8 nm and an emission
maximum at λ ) 5 3 7 nm in acetonitrile. Three new
methods based on this reaction have been developed:
Direct fluorescence spectroscopy, HP LC/ fluorescence, or
HP LC with UV/ vis detector may be selected as detection
techniques. On microplates, high-throughput fluores-
cence spectroscopy is achieved, while HP LC/ fluorescence
provides lower limits of detection, and HP LC with UV/
vis detection enables evaluation of the reaction with
standard instrumentation. Different water samples were
investigated using all detection modes, and a photometric
standard procedure was successfully employed to validate
the new methods with an independent technique.
Due to the importance of nitrite in the environmental sciences
and in food chemistry, a large number of analytical methods has
been developed in recent years. For the simultaneous determi-
nation of nitrite and other anions in waters, ion chromatography
with conductometric detection1 is a suitable technique, but
problems arise due to the elution of nitrite in the tailing of the
chloride peak and the high limit of detection at approximately 2
µmol/ L. Enrichment techniques in combination with ion chro-
matography have also been described with the goal to achieve
lower limits of detection.2,3 Another method described for the
determination of nitrite is based on two-dimensional capillary
isotachophoresis4 or on on-line coupled capillary isotachophoresis
with capillary zone electrophoresis,5 both with conductivity detec-
tion. The acid-catalyzed diazotation of p-aminobenzenesulfonamide
with subsequent coupling to N-(1-naphthyl)-1,2-diaminoethane
yields an azo dye which is quantified by means of spectropho-
tometry at a wavelength of 540 nm. This method has been
(6) European Standard, Water qualitysDetermination of nitritesMolecular
absorption spectrometric method. 1993, EN 26777.
(7) Ohta, T.; Arai, Y.; Takitani, S. Anal. Chem. 1 9 8 6 , 58, 3132-3135.
(8) Damiani, P.; Burini, G. Talanta 1 9 8 6 , 33 (8), 649-652.
(9) Axelrod, H. D.; Engel, N. A. Anal. Chem. 1 9 7 5 , 47, 922-924.
(10) Zhou, J. Y.; Prognon, P.; Dauphin, C.; Hamon, M. Chromatographia 1 9 9 3 ,
36, 57-60.
(11) Lee, S. H.; Field, L. R. Anal. Chem. 1 9 8 4 , 56, 2647-2653.
(12) Kieber, R. J.; Seaton, P. J. Anal. Chem. 1 9 9 5 , 67, 3261-3264.
(13) Karst, U.; Binding, N.; Cammann, K.; Witting; U. Fresenius J. Anal. Chem.
1 9 9 3 , 345, 48-52.
(1) German standard methods for the examination of water, wastewater and
sludge; anions (group D). 1988, DIN 38405, part 19.
(2) Haddad, P. R.; Jackson, P. E. J. Chromatogr. 1 9 8 7 , 407, 121-132.
(3) Cassidy, R. M.; Elchuk, S. J. Chromatogr. 1 9 8 3 , 262, 311-315.
(4) Meissner, T.; Eisenbeiss, F.; Jastorff, B. Fresenius J. Anal. Chem. 1 9 9 8 ,
361, 459-464.
(14) Po¨ tter, W.; Karst, U. Anal. Chem. 1 9 9 6 , 68, 3354-3358.
(15) Beasley, R. H.; Hoffmann, C. E.; Rueppel, M. L.; Worley, J. W. Anal. Chem.
1 9 8 0 , 52, 1110-1114.
(5) Kaniansky, D.; Zelensky´, I.; Hybenova´, A.; Onuska, F. Anal. Chem. 1 9 9 4 ,
(16) Gro¨ mping, A. H. J.; Karst, U.; Cammann, K. J. Chromatogr. 1 9 9 3 , 653, 341-
347.
66, 4258-4264.
10.1021/ac981330t CCC: $18.00 © 1999 American Chemical Society
Published on Web 06/25/1999
Analytical Chemistry, Vol. 71, No. 15, August 1, 1999 3003

We have described a new group of hydrazine reagents for
aldehyde and ketone determination which are characterized by
an R-methyl hydrazine function.18 These reagents are less sus-
ceptible to interferences by oxidants, as only one defined byprod-
uct with both nitrogen dioxide and ozone is formed. N-Methyl-
2,4-dinitrophenylhydrazine (MDNPH) is a reagent that reacts with
these oxidants with formation of N-methyl-2,4-dinitroaniline.18 The
product is easily separated from the hydrazones and detected by
UV/ vis spectroscopy. N-Methyl-4-hydrazino-7-nitrobenzofurazan
(MNBDH), a nonfluorescent reagent, however, yields N-methyl-
4-amino-7-nitrobenzofurazan (MNBDA) as a highly fluorescent
product in the reaction with NO+. We have used this new reagent
to develop an array of methods for the determination of nitrite
using microplate fluorescence spectroscopy, HPLC with fluores-
cence detection, and HPLC with UV/ vis detection.
Microplate Fluorophotometer. The microplate fluoropho-
tometer Fluostar (BMG LabTechnologies, Offenburg, Germany)
with software version 2.10-0 was used. Excitation filter, 470 nm;
emission filter, 538 nm.
HP LC Instrumentation. A high-performance liquid chro-
matograph consisting of the following components was used: two
LC-10AS pumps (Shimadzu), an SIL-10A autosampler (Shimadzu),
a SPD-M10Avp diode array detector (Shimadzu), a RF-10Axl
fluorescence detector (Shimadzu), Class LC-10 Version 1.6 soft-
ware (Shimadzu), and a CBM-10A controller unit (Shimadzu). The
injection volume was 10 µL. The column material was Merck
LiChroSpher RP-18 (Merck) in ChromCart cartridges (Macherey-
Nagel, Du¨ ren, Germany): particle size, 5 µm; pore size, 100 Å;
column dimensions, 250 mm × 3 mm; guard column, 8 mm × 3
mm.
HP LC Analysis. For separation, a binary gradient consisting
of acetonitrile and a mixture of water/ triethylamine/ acetic acid
(preparation: 500 mL water with 2415 µL of triethylamine and
975 µL of acetic acid, pH ∼ 7.5) was chosen with the following
profile and a flow rate of 0.62 mL/ min:
EXPERIMENTAL SECTION
Chemicals. All chemicals were purchased from Aldrich
Chemie (Steinheim, Germany) in the highest quality available,
except the following: Sodium nitrite and triethylamine were from
Fluka (Neu-Ulm, Germany). Acids were Merck (Darmstadt,
Germany) analytical grade. Acetonitrile for HPLC was Merck
gradient grade.
time (min)
0
1.5
45
8.5
10.5
100
11.5
45
12.5 (stop)
45
c (CH3CN) (%)
45
100
Synthesis and Characterization of the Reagent. The syn-
thesis and characterization of MNBDH are described in refs 19
and 20.
Nitrite Solutions for the Calibration. A total of 495 mg of
sodium nitrite was dissolved in 1 L of deionized water (7.17 ×
10-3 mol/ L). This solution was diluted 1:100 (7.17 × 10-5 mol/ L)
in water. The following concentrations were used for the calibra-
tions: 2.15 × 10-7, 3.59 × 10-7, 7.17 × 10-7, 1.44 × 10-6, 2.15 ×
10-6, 2.87 × 10-6, 3.59 × 10-6, 7.17 × 10-6, 1.08 × 10-5, and 1.44
× 10-5 mol/ L.
Synthesis and Identification of the Reaction P roduct. The
synthesis, based on a literature procedure by Clusius and
Schwarzenbach,21 was carried out for the synthesis of 2,4-
dinitrophenyl azide and modified as follows: 200 mg of N-methyl-
4-hydrazino-7-nitrobenzofurazan (9.4 × 10-4 mol) was dissolved
in 30 mL of ethanol and 5 mL of concentrated hydrochlorid acid.
While cooling with ice, a solution of 70 mg of sodium nitrite (1 ×
10-3 mol) in 5 mL of distilled water was added. After leaving the
reaction solution for 0.5 h at room temperature, 50 mL of distilled
water was added. Afterward the solution was left at room
temperature again for 1 h. The product precipitated as a yellow
material. This was filtered off and washed with ice/ water. The
yield was 67%. Product characterization (for peak assignment
compare to the structure below): 1H NMR δ 3.24 (d, 3H, N-CH3,
J ) 5.4 Hz), 6.18 (d, 1H, Hb, J ) 8.5 Hz), 8.54 (d, 1H, Ha, J ) 8.6
Hz); MS m/ z 194 (M+, 100), 164 (194 - NO, 27), 118 (29), 103
(26), 91 (31), 76 (29); IR 3040, 2718, 2677, 1650, 1621, 1540, 1312,
1269, 1161 cm-1. Anal. Calcd for C7H6N4O3: C, 43.30; H, 3.12; N,
28.87. Found: C, 43.18; H, 3.43; N, 28.55.
UV/ vis Spectrometer. The HP 8453 diode array spectropho-
tometer (Hewlett-Packard, Waldbronn, Germany) with software
HP Chem Station 845x biochemical UV/ vis system was used.
Spectrofluorophotometer. The RF-5301 PC spectrofluoro-
photometer (Shimadzu, Duisburg, Germany) with software version
1.10 was employed for fluorescence spectroscopy in cuvettes.
Microplate Spectrophotometer. The microplate spectropho-
tometer Spectra Max 250 with software Soft Max Pro Version 1.1
(Molecular Devices, Sunnyvale, CA) was used.
Microplate Fluorophotometer Method. A total of 2.5 mg of
MNBDH was dissolved in 25 mL of acetonitrile (4.8 × 10-4 mol/
L). A 200-µL aliquot of the corresponding nitrite solutions or
deionized water, respectively (blank solution), was pipetted into
each well of the microplate. A 7-µL sample of the MNBDH solution
was added to each solution, and afterward, 15 µL of concentrated
phosphoric acid was pipetted into each well. After a reaction time
of 30 min, the fluorescence was read with an excitation wavelength
of 470 nm and an emission wavelength of 538 nm.
HP LC Method. A 0.9-mL sample of the corresponding nitrite
solutions or deionized water, respectively (blank solution), was
pipetted into a vial. A 33-µL sample of the MNBDH solution
described in the last paragraph and 67 µL of concentrated
phosphoric acid were added. After 30 min, the solutions were
injected into the HPLC system. The substances were detected
with a diode array detector and a fluorescence detector (excitation
wavelength, 468 nm; emission wavelength, 537 nm).
Comparative Measurements on Microplates Using a
Reference Method with Microplate Spectrophotometry. This
method6 is a European Standard method for the spectrophoto-
metric determination of nitrite. The original procedure in cuvettes
has been modified in this work for microplate spectrophotometry
as stated below. Preparation of the Reagent Solution. Sulfanilamide
(1 g) was dissolved in 10 mL of concentrated phosphoric acid
and 50 mL of deionized water. After 100 mg of N-(1-naphthyl)-
ethylenediamine dihydrochloride was added, the resultant solution
(17) Kieber, R. J.; Bullard, L.; Seaton, P. J. Anal. Chem. 1 9 9 8 , 70, 3969-3973.
(18) Bu¨ ldt, A.; Karst, U. Anal. Chem. 1 9 9 7 , 69, 3617-3622.
(19) Bu¨ ldt, A.; Karst, U. German Patent, DE 198 00 537.7, 1998.
(20) Bu¨ ldt, A.; Karst, U. Anal. Chem. 1 9 9 9 , 71, 1893-1898.
(21) Clusius, K.; Schwarzenbach, K. Helv. Chim. Acta 1 9 5 8 , 41, 1413-1416.
3004 Analytical Chemistry, Vol. 71, No. 15, August 1, 1999

was transferred into a 100-mL volumetric flask and filled to volume
with water.
Determination of Nitrite. A 5-µL aliquot of the reagent solution
was pipetted into the wells of the microplate. A 200-µL sample of
the corresponding nitrite solution or deionized water, respectively
(blank solution), was added. Afterward, 45 µL water was added
to adjust the pH to 1.9. After a reaction time of at least 20 min,
the absorbance was read at a wavelength of 540 nm.
RESULTS AND DISCUSSION
Structure Elucidation of the Reaction P roduct. MNBDH
has previously been described as reagent for the determination
of aldehydes and ketones. The reaction is shown below. In acidic
media, MNBDH reacts with nitrite under formation of MNBDA
as follows:
Figure 1. Excitation spectrum (solid) and emission spectrum
(dotted) of MNBDA. The MNBDA concentration was 1.29 × 10-5
mol/L in water and phosphoric acid corresponding to the conditions
chosen for the microplate fluorescence and HPLC methods. Excitation
wavelength, λ ) 468 nm; emission wavelength, λ ) 548 nm.
1
The structure of the reaction product was confirmed by H
NMR, IR, UV/ vis, and fluorescence spectroscopy, mass spec-
trometry, and elementary analysis. The respective analysis data
are listed in the Experimental Section. For a positive identification,
the substance was synthesized as described above. The synthe-
sized product exhibited identical elution properties on different
columns in HPLC and identical UV/ vis spectra compared to the
reaction product. The fluorescence spectrum of MNBDA in a
solution of phosphoric acid and water is depicted in Figure 1.
Under these conditions, the excitation maximum is located at λ
) 468 nm and the emission maximum at λ ) 548 nm. DNPA, the
reaction product of DNPH with nitrite, exhibits long-term stability
only in the freezer, but not at room temperature.14 In contrast to
this, MNBDA is a very stable substance and can be stored for
several weeks at room temperature without decomposition. This
is valid for the solid derivative as well as for solutions in water
and organic solvents, e.g., acetonitrile.
Derivatization. The derivatization reaction requires the pres-
ence of strong acids to achieve complete derivatization within a
reasonable time scale. The development of the fluorescence for
two nitrite concentrations (1.29 × 10-5 and 1.29 × 10-6 mol/ L,
respectively) when reacting with 1.6 × 10-5 mol/ L MNBDH in
the presence of 1 mol/ L phosphoric acid was recorded on a
cuvette spectrofluorometer and is presented in Figure 2. The
relative fluorescence (compared to the fluorescence at a reaction
time at 30 min defined as 100%) with the excitation wavelength of
468 nm and the emission wavelength of 548 nm is plotted versus
time. It is obvious that the reaction is completed after ap-
Figure 2. Fluorescence/time curves for the formation MNBDA at
the emission wavelength of 548 nm. Excitation wavelength, 468 nm.
The concentrations of the nitrite solutions were 1.29 × 10-5 (square)
and 1.29 × 10-6 mol/L (circle) when reacting with 1.6 × 10-5 mol/L
MNBDH in the presence of 1 mol/L phosphoric acid.
proximately 25 min for both analyte concentrations, which were
selected as representative for the real samples to be analyzed. As
no significant decay of fluorescence was observed within 1 h after
completed derivatization, 30 min was selected as reaction time
for all methods described below.
Microplate Fluorescence Method. To achieve low limits of
detection, a large volume of the sample was reacted with small
volumes of acid and reagent as described in the Experimental
Section. The limit of detection of the microplate as well as the
HPLC fluorescence methods is limited by the triple RSD of the
blank, not by the signal-to-noise ratio to 1.5 × 10-8 mol/ L. The
respective limit of quantification is 5 × 10-8 mol/ L. In contrast,
the instrumental limit of detection for a calibration experiment
with pure MNBDA in acetonitrile is 5 × 10-9 mol/ L. Further
purification of the reagent results in lower blanks, but highly
purified reagent is easily oxidized by nitrogen dioxide from
ambient air, thus leading to large standard deviations of the blank.
However, the degree of purity used in this study is sufficient for
most real samples of nitrite taken from waters in agriculturally
used areas. To cover the concentration range expected for the
Analytical Chemistry, Vol. 71, No. 15, August 1, 1999 3005

Table 1. Cross Selectivities of Different Substances
Determined by Means of the Microplate
Fluorophotometer
without nitrite
c (nitrite) ) 10-6 mol/ L
interference
(10-4 mol/ L)
found
RSDb
(%)
recovery of
nitrite (%)
RSDb
(%)
valuea (%)
NaCl
102.0
98.4
97.1
91.5
109.1
95.9
100.2
100.5
126.9
85.7
106.7
89.6
145.8
110.3
97.7
7.1
10.7
5.7
5.0
4.5
5.1
6.0
4.2
7.8
8.8
6.7
9.4
7.0
12.7
7.8
8.6
7.4
103.9
103.9
106.1
97.9
111.9
94.0
116.1
115.5
128.8
103.5
82.9
2.1
1.3
1.3
4.3
7.2
2.0
2.3
2.3
3.4
2.8
4.1
0.5
3.3
3.1
2.7
1.9
3.7
NaNO3
NaHCO3
K2CO3
MgSO4
Na2SO4
KBr
CaCl2
NaI
FeCl3
H2O2
glucose
formaldehyde
acetaldehyde
acetone
acrolein
benzaldehyde
93.2
107.7
100.0
99.0
90.2
101.2
98.8
91.8
a
Blank ) 100%. b n ) 4.
real samples, linear calibration curves have been recorded for
nitrite in the concentration range from 3.59 × 10-7 to 1.44 × 10-5
mol/ L. The relative standard deviation for selected concentrations
was 3.5% for 1.44 × 10-6 mol/ L and 2.5% for 1.44 × 10-5 mol/ L.
The microplate method is well suited to analyze up to 96 samples
on one standard microplate, but a more favorable arrangement
includes 8 nitrite standard concentrations and up to 16 samples,
with each standard and sample in 4 wells of the microplate. This
way, high-throughput screening is combined with some statistical
evaluation. Compared to the cuvette format, the microplate
analysis saves more than 90% of reagent, as the total volume in a
well is below 250 µL, in contrast to 3 mL in a standard cuvette
with a 1-cm optical path length. Due to the convenience of
microplate fluorescence spectroscopy, all interference studies were
carried out with this method. As shown in Table 1, a series of
inorganic salts, oxidants, and organic substances including the
most important carbonyl compounds were added as 0.1 mmol/ L
solutions to a reagent solution with phosphoric acid, but without
nitrite to investigate the influences in the absence of nitrite (left
columns). The same experiment was carried out in the presence
of 1µmol/ L nitrite, thus simulating a 100-fold excess of the
interfering compound (right columns). For the experiments with
and without nitrite, the blank (water added instead of the
interferent solution) was defined as 100%. Without nitrite, the RSDs
are relatively large due to the low absolute values. Only iodide
and formaldehyde exhibit significantly increased values. In the
case of formaldehyde, this may be traced back to the formation
of the respective hydrazone, which is slightly fluorescent. As the
absolute fluorescence signal in the presence of nitrite is far larger,
the formation of the hydrazone can no longer be observed.
Therefore, formaldehyde is only slightly interfering in the micro-
plate method, while the separation in the HPLC methods excludes
an interference from the aldehydes, except due to consumption
of the reagent. The effect of sodium iodide may be due to the
catalytic properties of the iodide ion, which may increase the rate
of formation of MNBDA with nitrite and nitrogen dioxide from
Figure 3. Chromatogram of a nitrite solution (7.17 × 10-6 mol/L)
after reaction with MNBDH and HPLC with UV/vis detection (474 nm).
ambient air. In the presence of nitrite, hydrogen peroxide causes
a reduction of the fluorescence signal which may be due to the
saturation of the solution with oxygen from the peroxide decom-
position. All substances that have been identified as potential
interferences do not occur in these high concentrations in any
nitrite samples known to the authors. Therefore, interferences
from the investigated substances may be neglected.
HP LC Methods with Fluorescence and UV/ vis Detection.
As stated above, the limit of detection is limited by the blank.
However, due to the lower RSDs and the separation from the other
components of the reaction solution, limits of detection are lower
for the HPLC fluorescence method compared to the microplate
method, while the HPLC method with UV/ vis detection exhibits
a similar LOD compared to the microplate method. The extremely
strong fluorescence of MNBDA is obvious from two chromato-
grams of a derivatized sample (1.6 × 10-5 mol/ L reagent, 7.17 ×
10-6 mol/ L nitrite, 1 mol/ L phosphoric acid). Figure 3 shows the
chromatogram with UV/ vis detection at λ ) 474 nm, Figure 4
shows the same chromatogram with fluorescence detection at an
excitation wavelength of λ ) 468 nm and an emission wavelength
of λ ) 537 nm. The delay of the MNBDA peak (8 s) between the
detectors is due to the void volume of the detector cells and lines,
as both detectors are connected in series. The peak at 1.6 min in
the UV/ vis chromatogram is caused by the phosphoric acid. The
chromatograms show that there is almost no fluorescence from
the MNBDH reagent, thus enabling a direct fluorescence spec-
troscopic reading as performed in the microplate method. For the
HPLC method with UV/ vis detection, the blank is almost not
detectable. For HPLC with fluorescence detection, the limit of
detection is limited by the triple RSD of the blank, not by the
signal-to-noise ratio. Linear calibration curves were recorded from
3006 Analytical Chemistry, Vol. 71, No. 15, August 1, 1999

Table 2. Comparison of Water Samples Analyzed with Four Different Methods
reference
method
microplate fluoro-
photometer method
HPLC with
fluorescence detection
HPLC with
UV/ vis detection
concn
(mol/ L)
RSDa
(%)
concn
(mol/ L)
RSDa
(%)
concn
(mol/ L)
RSDa
(%)
concn
(mol/ L)
RSDa
(%)
source of water sample
cattle watering place, Papenburg, Germany
ditch in agriculturally used area
close to Papenburg, Germany
lake Aasee, Mu¨ nster, Germany
tap water (Mu¨ nster, Germany)
2.6 × 10-5
1.6
2.0
2.510-5
5.8
6.6
2.610-5
1.0
0.9
2.6 × 10-5
1.0
0.5
5.9 × 10-6
6.0 × 10-6
6.4 × 10-6
6.4 × 10-6
5.7 × 10-6
4.2
5.1 × 10-6
3.8
11.3
5.7 × 10-6
1.5
1.5
5.7 × 10-6
3.9 × 10-7
1.2
3.5
nqb
4.3 × 10-7
4.1 × 10-7
a
n ) 3. b Below the limit of quantification.
2.15 × 10-7 to 7.17 × 10-6 mol/ L for HPLC with fluorescence
and 2.15 × 10-7 to 1.44 × 10-5 mol/ L for HPLC with UV/ vis
detection, respectively. RSDs for triple determination of nitrite
standards ranged from 1 to 3% for both fluorescence and UV/ vis
detection. The limit of detection is 10-8 mol/ L for HPLC with
fluorescence detection and 3 × 10-8 mol/ L for HPLC with UV/
vis detection. The respective limits of quantification are 3 × 10-8
and 10-7 mol/ L, respectively.
Reference Method. The established reference method based
on the following reaction
•
was converted to a microplate method for more convenient
measurements of large numbers of samples. The procedure for
the analysis is stated in the Experimental Section. Linear calibra-
tion curves were recorded from 7.17 × 10-7 to 1.44 × 10-5 mol/
L. The limit of detection for this method on microplates is 2 ×
10-7 mol/ L and the respective limit of quantification is 6 × 10-7
mol/ L.
Comparative Measurements of Real Samples. Four water
samples of different origin have been investigated by all four
methods. The respective data are listed in Table 2. It is obvious
that all methods are well-suited for the analysis of the four samples,
except the photometric reference method, which does not exhibit
a sufficiently low limit of detection for the tap water sample. All
analyses were carried out in triplicate. The HPLC data are not
for triple injection of the same reaction mixture, but for triple
reaction and chromatographic analysis. The respective RSDs of
all methods are listed in Table 2. The RSDs for the HPLC methods
are excellent, with RSDs not exceeding 3.5% even for the lowest
concentrations. The RSDs for the microplate methods are higher,
as could be expected from pipetting small volumes and from the
Figure 4. Chromatogram of a nitrite solution (7.17 × 10-6 mol/L)
after reaction with MNBDH and HPLC with fluorescence detection
(excitation, 468 nm; emission, 537 nm).
fact that the reaction products are not separated from the reaction
mixture. All four methods correlate well for the water samples.
Discussion. MNBDH is well suited as a new reagent for the
determination of nitrite in waters. A major advantage of the
reaction is the versatility, thus allowing one to quantify nitrite by
fluorescence spectroscopy without or after HPLC separation as
well as by HPLC with UV/ vis detection. The first microplate
fluorometric method for nitrite determination has been developed
and allows for high-throughput screening of a very large number
of nitrite samples within a short time scale.
Received for review December 1, 1998. Accepted April 27,
1999.
AC981330T
Analytical Chemistry, Vol. 71, No. 15, August 1, 1999 3007