Analytica Chimica Acta, 110 (1979) 123-128

0 Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands

SIMULTANEOUS SPECTROPHOTOMETRIC DETERMINATION OF

NITRITE AND NITRATE BY FLOW INJECTION ANALYSIS

LRIF’ ANDERSON
Department of Analytical Chemistry, Chalmers University of Technology and University
of Goihenburg, S-412 96 Giiteborg (Sweden)
(Received 27th February 1979)

SUMMARY

The flow injection principle is used in the photometric determination of nitrite and
nitrate with sulfanilamide and N-(I-naphthyl)ethylenediamine as reagents. An on-line
copper-coated cadmium reductor reduces nitrate to nitrite. The detection limit is 0.05 DM
for nitrite and 0.1 pM for nitrate at a total sample volume of 200 pl. Up to 30 samples
can be analyzed per hour with a relative precision of ca. 1%.

Two of the most frequently required determinations in environmental

investigations are those of nitrite and nitrate. Numerous analyticat methods,
all based on reduction of nitrate to nitrite with subsequent calorimetric
determination of the nitrite formed, have been presented [l--S]. Many of
these methods have been adapted to automatic air-segmented continuous
fiow systems, e.g. the Technicon AutoAnalyzer system. Several reducing
agents, such as zinc [2], cadmium [ 33, amalgamated 14, 51 or copperized
cadmium [6, 71, have been investigated. Reductor columns are, however,
difficult to operate in an air-segmented stream. This problem can be avoided by
using the continuous flow injection technique developed by R&%ka and
Hansen [9, lo].
The flow injection technique is based on three main principles: sample
injection, reproducible timing and controlled dispersion [lo] _ The dispersion
can be described as limited, medium or large; in a calorimetric system based
on a reaction between the sample and a suitable reagent, a medium dispersion
is preferred. Thus in the flow injection determination of nitrate, the
reductor column should not excessively increase the dispersion. In a
copperized cadmium reductor, more than 90% of the total nitrate is reduced
within l-2 s with minimum risk of further reduction of nitrite 1131. Con-
sequently, the reductor can be made very small which results in a minimal
increase of dispersion. The Shinn method [ll] for the determination of
nitrite was chosen for this work because of its high sensitivity and relative
freedom from interferences.
124

EXPERIMENTAL

Apparatus
A schematic diagram of the flow injection system is shown in Fig. 1. An
Ismatec model MP13 peristaltic pump was used. Different flow rates were
obtained by changing the pump tube diameters as indicated in the legend to
Fig_ 1, The injection port was a rotary valve similar in construction to those
described previously [IO, 121, the main difference being that the sample
volume could be varied between 10 and 1000 ~1 simply by changing the
length of the sample loop (see Fig. 2). A Glenco model 5’7V dual absorbance
monitor, with cuvette volumes of 25 ,ul (light paths, 10 mm) was used as de-
tector, and was connected to a Kipp and Zonen BIB41 dual-channel recorder.

ml min-’

C 023

8 023
S
1ocm
A 1.00

5 0.23

C 0.23

1.00
W

Fig_ 1, Flow diagram for the coiorimetric determination of nitriteand nitrate. The internai
diameter of the tygon tubing was 0.64 mm for the 0.23 ml min-’ flow rates aud 1.30 mm
for the 1.00 ml min-’ flow rates; all other tubing was of 0.7 mm i.d_ The reagents were:
(A) carrier stream; (33) sulfanilamide solution and (C) diaminesolution. S denotes the point
of injectionand W waste.

sample

- reagent +-- reagent

Fig. 2. Rotary-valve injection port. (A) Sample loop; (B) O-rings; (C) by-pass of higher
hydrodynamic flow resistance thti the sample loop (100 cm length and 0.5 mm i-d.),
 125

Reugen ts
All reagents were of analytical-reagent grade. The cadmium granules
(particle diameter <O-3 mm) were coated batch-wise with copper in a 1%
(w/v) copper sulphate solution. The reductor tube (60 mm long with an
inner diameter of 1 mm) was filled with copperized cadmium held in position
with glass wool plugs. The copperized cadmium stock was stored under
deaerated water.
The reaction medium (carrier stream) contained 0.4 M ammonium chlo-
ride and 0.3 M sodium chloride. The sulfanilamide reagent was prepared by
dissolving 2.5 g of sulfanilamide and 13 ml of concentrated hydrochloric
acid in 250 ml of the reaction medium. The diamine solution was prepared
by dissolving 0.25 g of IV-( l-naphthyl)ethylenedi~ine dihydrochloride and
10 g of sodium chloride in 250 ml of the reaction medium. The standard
stock solutions of sodium nitrite and sodium nitrate were 0.001 M in 0.7 M
sodium chloride. Working standard solutions were obtained by diluting with
0.7 M sodium chloride-

Procedure
The standards and samples were sucked into the sample loop of the injec-
tion valve by means of the peristaltic pump (see Fig. 2). After an injection,
the valve was not returned to the reload position until the maximum ab-
sorbance value (540 nm) for the previous sample had been reached, because
sample loading creates pressure variations which slightly affect the baseline.
The concentrations of nitrite and nitrate were evaluated from the peak
heights by using calibration curves prepared from standards (see Fig. 3).

RESULTS

Optimum experimental conditions

Peaks obtained by injecting different sample volumes of a 1 r_tM nitrite
standard are shown in Fig, 4. The optimum sample volume is of the order of
100 ~1. Peak broadening is due to mixing in the manifold.
The coil lengths were optimized to give ample reaction times and minimum
dispersion. Optimal conditions were obtained with coil lengths of 100 and
300 mm, respectively, at the fiow rates chosen for the reagents (Fig. 1). The
short reductor column produced only negligible dispersion with minimum
loss in efficiency and reproducibility. The lifetime of the reductor column
naturally depends on the amount of oxidizing material which passes through
it; in continuous analysis, the lifetime was found to be not less than 8 h. Re-
activation of the column in situ is not satisfactory. The column should there-
fore be repacked and replaced in the system on a daily basis.

Precision, detection limit and rate of analysis

Twenty consecutive analyses were made of a sample having nitrite and
126

IQOlA

QOSA

- (/.A1

Fig. 3. Determination of nitrite and nitrate at a rate of 30 samples per h_ For the outputs
shown on the left, the concentrations of nitrite are 2, 1, 0.2 PM for A, B, C, and those
of nitrate are 20, 10, 2 BM for a, b, c, respectively_ The calibration curves prepared from
these data are shown on the right.

nitrate concentrations of 0.4 and 4.0 PM, respectively. The relative standard
deviations were better than 1% in both cases.
The spectrophotometric signal is due not only to the absorbance but also
to the refraction of the sample. Refraction is of no practical importance at
absorbance values above O-05. The detection limit is determined by the
difference in refraction between the reagent stream and the sample. Thus, in
order to achieve low detection limits - 0.05 PM for nitrite and 0.1 PM for
nitrate - it was necessary to adjust the refractive index of the reagents to a
value as close as possible to those of the samples. For sea-water analysis, this
was achieved by adding sodium chloride to the reagents. This also improved
the reproducibility of dispersion because density differences were found to
cause turbulence.
With a sample volume of 200 ,~1(100 ~1 for nitrite and 100 ~1 for nitrate),
30 samples could be analyzed per hour (Fig. 3).

IXthe analysis of sea water, the only significant interference arises from
turbidity caused by particles in the sample. Prior filtration of the sample is
therefore necessary. For anoxic waters, however, sulfide concentrations over
 127

A
4
0.10 -

005 -

Fig. 4. Response curves obtained by injecting different sample volumes of a 1 PM nitrite

standard.

2 PM were found to decrease the absorbance. This was overcome by adding

an excess of either Cd’+ or Hg*’ to the sample [ 15]_

DISCUSSION

The results reported above show that the dispersion pattern of a fIow injec-
tion system is not significantly affected by a small reductor column (see
Fig. 3). Consequently, the flow injection principle is suitable for the simulta-
neous determination of nitrite and nitrate in samples such as sea, tap or
waste water. The optimum concentration range is l-25 PM for each of
these ions. These concentration ranges lie within the probable ranges for any
sort of pulluted waters, although the lowest values reported fall near the
detection limits of the flow injection method. Higher concentrations can be
determined by decreasing the sample volume or by dilution.

The author is indebted to Professor David Dyrssen, Dr. Daniel Jagner and
Dr. Anders Graneli for general discussions.

REFERENCES

1 J. B. Mullin and J. P. Riley, Anal. Chim. Acta, 12 (1955) 464.

2 T. J. Chow and M. S. Johnstone, Anal. Chim. Acta, 27 (1962) 441.
128

3 M. Barnard and G. Macchi, Automation in Analytical Chemistry, Technicon Sym-

posia, 1965, Mediad, N.Y. 1966,252.
4 A. W. Morris and J. P. Riley, Anal Chim. Acta, 29 (1963) 272.
5 K. Grasshoff, Kieler Meeresforsch-, 20 (1964) 5
6 E. D. Wood, F. A. Armstrong and F. A-Richards, J. Mar. BioLAss. U.K.,47 (1967) 23.
7 A. Henriksen and A. R. Sehner-Olsen, Analyst, 95 (1970) 514.
8 F. A. J. hmstrong, C_ R. Stearus and J. D. H. Strickland, Deep-Sea Res., 14 (1967)
381.
9 J. Rfi%ka and E. H. Hansen, Anal. Chim. Acta, 78 (1975) 145.
10 J. R&&&a and E. H. Hansen, Anal. Chim. Acta, 99 (1978) 37.
11 M. B. Shinn, Ind. Eng. Chem., Anal. Ed., 13 (1941) 33.
12 E. H. Hansen and J. R%&a, Anal. Chim. Acta, 87 (1976) 353.
13 F. Nydahe, Taianta, 23 (1976) 349.
14 J. D. H_ Strickland and T. R. Parsons, A Practical Handbook of Seawater Analysis,
Fisheries Research Board of Canada, Bulletin 167,196s.
15 A. Timmer and T_ Hoor, Mar_ Chem., 2 (1974) 149.