Flow injection analysis of nitrogen dioxide using a galvanic detector

A flow injection configuration (FIA) based on a galvanic detector for the determination of nitrogen dioxide is described. The gaseous sample is directly injected into a gaseous carrier. The sample is transported toward the detector. The steady state measurements are not required to obtain the reproducible peak signals. The features of FIA are compared with that of continuous flow monitoring application. The flow injection system is simple, rapid and capable of detecting NO2 in the range of 1-500ppm (v/v). The measuring range and sensitivity of the galvanic detector in FIA depend on the sample volume. A relative standard deviation is 2.4% (n = 10) for 200ppm (v/v) of nitrogen dioxide. The sampling frequency is about 24 h-1


Introduction
Flow injection analysis (FIA) is based on the injection of a liquid sample into a moving unsegmented continuous carrier stream of a suitable liquid. The injected sample forms a zone, which is transported toward a detector. The hallmark of FIA is that dispersion is reproducible and controllable. Now FIA has evolved into a general technique for solution handling and data gathering, applicable to many areas of chemical research and technology [1]. FIA was mainly used in the analysis of liquid samples. The analysis of gaseous samples is attended by special difficulties [2]. Absorption by suitable solutions is usually used to collect and/or preconcentrate the analyte prior to the measurement. Only a few flow injection methods have been reported for the determination of gaseous samples. These methods can be assembled into these groups: (1) the gaseous sample is directly inserted into a flow injection system and reacts at a gas/liquid interface [2][3][4][5], or at a gas/solid interface [6]; (2) a dual phase gas diffusion/permeation technique which involved a liquid donor and gaseous acceptor [7][8][9][10] or the reverse [11-15] is employed by means of a separated membrane; and (3) a gas sensor is directly used as the detector in FIA [16,17]. Nitrogen dioxide is a major air pollutant that plays a dominant role in acid deposition chemistry, as well as in the production of ozone and hydroxyl radicals [18]. A number of analytical methods has been developed suitable for emission control and the determination of low levels occurring in the atmosphere. The widely adopted Address for correspondence. This paper was presented at the Seventh International Conference on Flow Analysis, Piracicaba, Brazil, 25-28 August 1997. instrumental chemical method for gaseous pollutants is coulometric analysis [19]. This involves measurement of the electrical current produced when strongly oxidizing or reducing pollutant gases react with potassium iodide or bromide solution in an electrochemical cell [20][21][22].
The gaseous sample is drawn through the detector containing a platinum cathode and a carbon anode with a galvanic potential difference between them. The current in the anode-cathode circuit is proportional to the amount of NO2 entering the cell [21][22][23][24]. This type of galvanic detector was designed for continuous monitoring of nitrogen dioxide for a steady state measurement. The present paper describes a study on the features of the galvanic detector of nitrogen dioxide in flow injection analysis. The results of FIA are compared with the usual continuous flow monitoring. While the method of sample introduction changes from continuous flow monitoring (steady state) to FIA, the reproducible peak signals are obtained. The FIA method is simple, rapid and extending the measuring range of the galvanic detector. The effect of FIA variables, i.e. flow rates of sample and carrier, sample volume, delay time and calibration curve, is examined.

Experimental
Gases A certified gas of nitrogen dioxide was obtained from the National Research Centre for Certified Materials (NRCCM, Beijing, China), and the certified value was 0.103% NO/N, mol/mol (No. 406850). The mixtures of nitrogen dioxide with air were also prepared in a 10litre glass bottle connected to an air-pump and sampling coil of the injected valve in the sample enclosure, and were calibrated by means of the certified nitrogen dioxide. Air filtered by the scrubbers of silica gel and activated charcoal was used as the gaseous carrier. CO (500ppm in N) was obtained from a cylinder (NRCCM). The gases SO, H2S and CO were produced by chemical reactions of their respective salts with a solution of sulphuric acid, and were separately prepared in the 10-1itre glass bottle containing air.

Apparatus
The samples of nitrogen dioxide were detected by a NO galvanic detector (The Third Analytical Instrument Factory of Tianjin, China) and a schematic diagram of the component is given in figure l(a). The principle of the detector was the coulometric internal electrolysis [21][22][23][24]. The detector utilized a cyclic oxidation-reduction process and contained a platinum gauze cathode and carbon anode. The composition of the electrolyte con- sampling coilings used to introduce the gaseous samples. The carrier and sample streams were aspirated by the pumps. Flow rates of sample and carrier were indicated by the two flowmeters (701 HB, range of 0.05-0.501/min, Beijing and LZB-3T, range of 0.1-1.01/min, Changzhou, China). The operational amplifier circuit was laboratorymade. The signal peak was displayed by a chart recorder (Dahua, Shanghai). All measurements were performed at room temperature, about 15 C.

Manifold
The scheme of the flow injection system is shown in figure l(b). The clean air obtained by flowing the air through activated charcoal ($1) and silica gel ($2) filters was continuously aspirated as carrier through the galvanic detector (D). The gaseous mixture of nitrogen dioxide was circulated within the sample enclosure and the loop of the injection valve was filled with the sample by means of an air-pump (P2). The flow rates of carrier and sample were adjusted by the adjusting valves (V1,V) to 0.25 and 1.01/min, respectively. Various volume samples of 0.5-3.0ml were injected into the air carrier. The peak signals of the galvanic cell current were monitored as the potential drop across a precision resistor.

Results and discussion
The galvanic detector of nitrogen dioxide was usually employed in the continuous flow monitoring, and the steady state signal was necessary. When the detector is used in a flow injection manifold, a typical peak graph of FIA is observed [25]. The printout of multiple injections of NO2 sample is reproducible for a given concentration of nitrogen dioxide. The hallmark of a conventional FIA experiment is that dispersion is reproducible and controllable. By observing the peak shapes and heights it was clear that the sample zone reproducibly expanded or diluted in the flowing stream. Since identical physical and chemical conditions are consistently obtained in FIA, the steady state is not necessary. A six-port rotary sample valve and syringe were tested for the introduction of gaseous samples. For 200ppm (v/v) NO, the relative standard deviations are 2.4% (n 10) for the quantitative sampling valve and 14.3% (n 8) for the syringe, respectively. The result shows that for the gaseous analysis the method of sample introduction has greatly influenced the reproducibility and that the six-port quantitative sample valve is better than the syringe.
where I is the galvanic current in gA:f is the sample flow rate in ml/min at 20C and 1013mbar; and c is the nitrogen dioxide concentration in ppm (v/v). In a certain range of flow rate, for a given concentration of nitrogen dioxide the output current is dependent on flow rate and linearly increases with flow rates [22].
When the detector is used in the flow injection mode, a graph of the signal as a function of flow rates, for sample and carrier, shows the different behaviours in comparison with that in common continuous flow monitoring mode (steady state, figure 2). At a constant sample flow rate, 1.01/min, the increase of the flow rate of the carrier results in a small initial signal increase at 0.201/min followed by a continuous linear decreasing (curve B). This behaviour shows that, above certain sampling rates, it is not necessary to provide precise flow control of the sample. When the sample is maintained in the stationary mode or at lower flow rates (i.e. less than 0.101/min), the signal is smaller. This is perhaps caused by the absorption of nitrogen dioxide by the sample coil.

Analytical frequency
The effect of the interval At of delay time on the signal reproducibility was examined (table 1). The first sample was injected into the carrier and the first signal peak (i0) obtained. After the time interval At, the second sample was injected and obtained the second signal peak (i).
When At is longer than about 100 s, the relative difference of the two signals is smaller than 5%. Therefore, the minimum time interval between sample introductions must be about 2.5 min to obtain the relative difference between and i0 less than 1.5%. Because the physical mixing and chemical conditions can be carried out reproducibly in FIA, the steady state signal is not necessary. In comparison with that of continuous flow monitoring, the analytical rate of FIA is increased. The sampling frequency is 24 h -1 Sample volume The change of the injected-sample volume is a powerful way to change dispersion. Increasing the volume of the   [1]. The effect of sample volume on the signal at various NO concentrations is shown in figure 3. Below 1.0ml of sample volume, the peak height for various concentrations of NO2 is linearly proportional to the injected sample volumes; above 1.0ml the signal increases gradually with the sample volume. The increase of sensitivity is limited by the rate of adsorption and reaction of nitrogen dioxide at the electrode. Therefore, further increasing sampling volume increases little sensitivity and causes a decreasing of the analytical frequency. The optimum sample volume is considered as 1.0 ml.

Conclusions
The galvanic detector of nitrogen dioxide is suitable for the flow injection analysis of nitrogen dioxide, extending the applications of the detector. The FI method speeds up the analytical process because identical physical and chemical conditions are consistently obtained, and the steady state is not required. The optimum sampling volume for the nitrogen dioxide galvanic cell in flow injection analysis (FIA) is 1.0ml. In FIA the syringe is not suitable for the introduction of gaseous samples, while a quantitative rotary sampling valve gives good reproducibility. The measuring range and sensitivity of flow injection systems are easily adjusted by changing the sample volume.