ANALYST, AUGUST 1992, VOL. 117 1259 Flow Injection Amperometric Determination of Nitrite at a Carbon Fibre Electrode Modified With the Polymer [O~(bipy)~(PVP)~&l]Cl* Michael M. Malone, Andrew P. Doherty, Malcolm R. Smytht and Johannes G. Vost School of Chemical Sciences, Dublin City University, Dublin 9, Ireland The development of carbon fibre electrodes modified with the polymer [Os( bipy)2(PVP)20CI]CI for the flow amperometric determination of nitrite is described. This osmium polymer modifier greatly enhances the kinetics of nitrite reduction compared with the reaction at bare carbon electrodes. Various electrode characteristics were optimized using both cyclic voltammetry and flow injection. The Cali bration graph yielded a slope of 0.197 nA cm3 pg-1 over the linear range 0-400 pg cm-3 with a limit of detection of 0.1 pg cm-3.The modified electrode was shown to exhibit good short-term reproducibility yielding a relative standard deviation of 2.1 5% ( n = 20). After a 3 week period of monitoring, involving 240 standard injections and 30 meat extract injections, the electrode continued to function with no significant change in sensitivity. The electrode was used to analyse a processed meat sample for nitrite content and the results compared favourably with those obtained using a standard reference spectrophotometric method. Keywords: Carbon fibre electrode; polymer modification; nitrite determination; flow injection amperometric detection; meat analysis The potential hazard of nitrite to human health has been well documented. 1 Conventional techniques for the determination of nitrite are based on spectrophotometric procedures using azo dye formation reactions.2 These methods however have limited sensitivity and dynamic range, and frequently suffer from interferences such as ascorbic acid.Several polarographic methods for the sensitive determina- tion of the nitrite ion have been reported in the literature.3" However, such polarographic methods have obvious disad- vantages for routine sensing and flow applications, and in recent years more emphasis has been placed on the use of solid electrodes for such applications. A large number of methods have been developed for the voltammetric determination of nitrite by oxidation at solid electrodes. Cox and Kulesza7 modified a platinum electrode by chemisorption of iodine, which was found to improve the reproducibility and decrease the peak width in the oxidation of nitrite by linear scan voltammetry.Nitrite oxidation at a bare glassy carbon electrode was reported by Newbery and Lopez de Haddad,g but this method suffered interferences from both ascorbate and chloride ions. The determination of nitrite following oxidation at both electrochemically pre-treated9 and polymer modified1@-12 glassy carbon electrodes at lower operational potentials has therefore been investigated. With the ruthen- ium polymer modified electrodes reported by Barisci et aZ.11 and Wallace et aZ.,1* the major problem associated with the electrodes was their long-term stability. Reduction of nitrite at Au, Pt and carbon electrodes is known to cause severe surface fouling.13 Reductive techniques are also limited by the negative potentials that are required for detection where interferences from metal cations, hydrogen peroxide and oxygen may be problematic.A modified electrode for the reduction of nitrite at moderate potentials, based on modification of glassy carbon macro-electrodes with the electrocatalytic polymer [ O S ( ~ ~ ~ ~ ) ~ ( P V P ) ~ ~ C I ] C I , has recently been reported. 14 Such modification offered several advantages over existing electrochemical and spectropho- tometric procedures. 14 In recent years there has also been much interest in the application of microelectrodes. These electrodes have several advantageous features when compared with macroelectrodes. * Presented at the meeting on Analytical Applications of Chemi- cally Modified Electrodes, Bristol, UK, January 7-8, 1992.Authors to whom correspondence should be addressed. For example, they enable time-independent currents to be monitored, are virtually non-destructive of the analyte and can be used in solutions of very high resistance.15 This paper describes the flow detection of nitrite in meat samples using a carbon fibre microelectrode that has been modified by chemisorption of the electrocatalytic polymer [0~(bipy)~(PVP)~~Cl]Cl, where bipy = 2,2'-bipyridyl and PVP = poly(4-vinylpyridine). The modified microelectrode was applied to the determination of nitrite in processed meat and the results are compared with those obtained using a standard spectrophotometric method.16 Experimental Reagents and Materials All reagents were of analytical-reagent grade.All aqueous solutions were prepared using de-ionized water obtained by passing distilled water through a Milli-Q water purification system. The electrolyte used throughout was 0.1 mol dm-3 HzSO4. Nitrite standard solutions were prepared daily using sodium nitrite [BDH (now Merck)], by appropriate dilutions in 0.1 mol dm-3 Na2S04 (Riedel-de Haen), as nitrite solutions are known to be unstable at low pH.16 The synthesis of the polymers has been reported elsewhere.17,18 Carbon fibres (14 pm diameter) were obtained from Avco. The surface of these fibres had no external coating. The meat sample analysed was Denny processed ham. Instrumentation Cyclic voltammetry was performed using an EG & G Princeton Applied Research (PAR) polarographic analyser/ stripping voltammeter in conjunction with a JJ Lloyd Instru- ments x-y recorder (Model PL4).A 20 cm3 laboratory-built three-electrode cell was employed for batch studies incorpor- ating the modified electrode, an Ag-AgCI reference electrode and a platinum auxiliary electrode. The working electrode used for the batch studies was prepared by inserting a carbon fibre into the narrow end of a plastic micropipette tip and gently heat-sealing it. The electrical connection was made by back-filling with mercury and dipping in a copper wire. The flow injection apparatus consisted of a Gilson Minipuls- 3 peristaltic pump, and a six-port Rheodyne injector with a 0.020 cm3 fixed-volume sample loop which was connected to the three-electrode system described below.The electrode1260 ANALYST, AUGUST 1992. VOL. 117 terminals were connected to an EG & G PAR Model 400 EC detector which was linked to a Philips PM8251 single-pen recorder to record the amperometric signals. A Shimadzu ultraviolet/visible recording spectrophotometer (Model UV- 240) was used in the comparison method for the determination of nitrite in meat. Construction of the Carbon Fibre Flow Cell The preparation of the carbon fibre working electrode was carried out using a method reported previously.19 The Ag-AgC1 reference electrode was prepared by firstly connect- ing a silver wire (0.1 mm diameter) to the anode and a platinum electrode to the cathode of a 1.5 V battery, after which the assembly was immersed in a solution of 1 rnol dm-3 HCl for 2 min.The wire was then inserted in a polyethylene tube (15 x 1 mm i.d.), one end of which was plugged with a ceramic porous rod (2 x 1 mm id.). The tube was then filled with the 1 rnol dm-3 HC1 internal reference solution and closed by heating. A 2 cm piece of stainless-steel tubing (1 x 0.2 mm id.) served as a counter electrode. Throughout this paper, potentials are quoted, after numerical conversion, versus the standard calomel electrode (SCE). After modification of the working electrode (described below) the working, reference and auxiliary electrodes were mounted in a T-tube arrangement reported previously,20 so that the electrolyte passed first through the working electrode and then via the auxiliary electrode to waste.Modification of the Working Electrode The electrode was connected to the tubing of a peristaltic pump via silicon tubing. The carbon fibre was cleaned by pumping 25 cm3 of 5 mol dm-3 HCI, 25 cm3 of de-ionized water and 25 cm3 of methanol, respectively, through the electrode at a rate of 1 cm3 min-1. The electrode was then air-dried by pumping air through for 1 h. When dry, the electrode was modified with the polymer solution of desired concentration in methanol. A 2 cm plug of 0.1% m/v polymer solution was first drawn into the peristaltic tubing, followed by air. The plug was pumped slowly towards the electrode. The pump was stopped for 40 s when the plug surrounded the electrode surface, allowing the polymer to chemisorb onto the fibre surface, after which time the pump was restarted so that air flowed past the electrode surface.In this way the electrode was dried for 3 h before use. Meat Sample Analysis The determination of nitrite in meat was carried out on a sample of processed ham (Denny). This involved the prior extraction of the nitrite from a 5 g sample of meat using a standard Association of Official Analytical Chemists (AOAC) method.21 The procedure involves extraction with hot water for 2 h. The extract was then filtered and made up to 50 cm3 in a calibrated flask. Prior to injection, the extract was diluted 1 + 1 with 0.2 mol dm-3 Na2S04, so that samples and standards were 0.1 mol dm-3 in Na2S04. Injections (0.020 cm3) of standards and samples were made. Results and Discussion Cyclic Voltammetry Studies were first carried out to investigate the retention of the osmium polymer on the carbon fibre surface and its electro- catalytic activity towards nitrite reduction.Well defined oxidation and reduction responses associated with the surface- bound OS~~--OS"~ couple were observed in 0.1 rnol dm-3 H2S04. This supporting electrolyte had previously been shown to be optimum for a macro glassy carbon electrode I I I I I E N 0.8 0.3 -0.2 Fig. 1 T pica1 cyclic voltammogram obtained for [Os(bipyh- (PVP)&$Cl. Scan rate = 0.10 V s-l; electrolyte = 0.1 rnol dm-3 H2S04 1 I I 1 0.6 0.2 -0.2 EN Fig. 2 Cyclic voltammograms for A, modified electrode in blank 0.1 rnol dm-3 H2SO4 electrolyte; B, same electrode as A in 5 x 10-3 rnol dm-3 NO2-; and C, same as B at higher sensitivity outlining the typical electrocatalytic shaped curve obtained on addition of nitrite.Scan rate = 0.010 V s-1; electrolyte = 0.1 rnol dm-3 H2SO4 using a similar polymer.14 Fig. 1 shows a typical cyclic voltammogram obtained from a modified microelectrode in 0.1 rnol dm-3 H2SO4. The peak-to-peak separation (AEp) of these waves was found to be 0.045 V, indicating the favourable kinetics associated with the OS~~-OS~I~ redox couple. A linear dependence of the cathodic and anodic peak currents on scan rate was observed at lower scan rates (less than 0.010 V s-*), indicating the predominantly surface behaviour of the modi- fier, as noted with some other polymer modified electrodes.22 The modified microelectrode stabilized very rapidly, and repetitive cycling over a 2 h period produced no significant change in response, indicating strong adsorption of the polymer.ANALYST, AUGUST 1992.VOL. 117 1261 The effect of nitrite reduction on this redox couple is shown in Fig. 2. On addition of nitrite to the solution, an electro- catalytic reduction occurred at the same potential as that observed for the reduction of Osl*' to 0s". When the voltammogram obtained after addition of nitrite (Fig. 2, B and C) is compared with that obtained from the same electrode in blank electrolyte solution (Fig. 2, A), the nitrite response can clearly be seen. The cyclic voltammogram is significantly elongated along the current axis, which is indicative of an electrocatalytic reaction. The electron transport kinetics at the bare glassy carbon electrode surface were too slow to yield a useful analytical signal.Cyclic voltammetry was used to determine the optimum length of time needed for optimum coating of the microelec- trode with the polymer. This was examined by dipping the fibres in a polymer solution for different lengths of time varying from 10 to 360 s and performing cyclic voltammetry on the resulting modified electrodes. The cyclic voltammetric peak heights were then measured, and 40 s was found to be the coating time giving rise to the best sensor response. At coating times longer than 40 s , the methanol solvent appeared to redissolve the polymer from the carbon fibre surface. The effect of the concentration of polymer solution in modifying the carbon fibres was also investigated.It was found that thicker polymer films produced smaller responses to nitrite, which is probably a result of hindered analyte and counter ion mass transport within the film.23324 An optimum polymer concentration of 0.1% m/v was found and used in further studies. As the osmium response is always present, the response for nitrite must be recorded on top of a substantial background current. However, by using amperometric detection with flow injection the background current can readily be offset, so that the subsequent responses will be due to addition of nitrite. Flow Injection When utilizing the polymer modified carbon fibre electrode in acidic flowing streams, the potential applied ensures that all the osmium is in the 0s" form, so that when the analyte reaches the surface it is reduced at the osmium centres within the polymer film according to the following proposed cross- exchange reaction mechanism14 N02- + H+ HNO2 (1) HN02 + H+ NO+ + HzO (2) NO+ + 0s" -+ NO(g) + 0s"' (3) The potential was varied from +0.195 to -0.250 V in 0.050 V increments, and injections of 0.020 cm3 of a 50 vg cm-3 NO2- standard solution were made at each poten- tial.Decreasing the potential resulted in increased sensitivity for nitrite reduction, but below -0.150 V the background noise also increased. Therefore, the detection potential was set at -0.150 V. The flow rate of the 0.1 mol dm-3 H2S04 electrolyte was kept low (0.2 cm3 min-I), as the kinetics of nitrite reduction are fairly slow. In fact, a useful signal cannot be obtained using these conditions at a bare carbon fibre electrode.However, under the same conditions using the polymer modified electrode, well defined and reproducible responses were obtained. The linearity of the method was determined by injecting a series of nitrite standards in the concentration range WOO pg cm-3 and constructing a calibration graph. The method was linear over this range with a correlation coefficient, r , = 0.999 and the regression equation was: y(nA) = 0.197~ (pg cm-3) - 1.9845. This linear range is better than that obtained using spectrophotometric procedures.25 The limit of detection was determined to be 0.1 pg cm-3 using a signal-to-noise ratio of 3: 1. The precision of the method was evaluated in terms of the variability between 20 replicate injections of various concen- trations of NO2- solution. This resulted in a relative standard deviation of 2.15% for all concentrations examined.Fig. 3 shows some typical amperometric responses obtained from the modified fibre flow cell at different nitrite concentrations, where the reproducibility of response is clearly evident. The long-term stability of the modified microelectrode was monitored over a 3 week period. After 3 weeks of continuous use, involving over 240 standard injections and 30 meat extract injections, the electrode showed no appreciable change in sensitivity. No surface-fouling effects were observed over this period of time. These results demonstrate the long-term stability of the modified microelectrode. This level of stability is rare for modified electrodes.Similar work has been carried out using the ruthenium polymer modified carbon fibre microelectrode analogue as a continuation of the study of ruthenium polymer modified glassy carbon electrodes.11.12 In this instance the oxidation of nitrite was investigated, which is a kinetically faster reaction than that of nitrite reduction. The use of the ruthenium polymer modifier lowered the overpotential for nitrite oxida- tion compared with that at a bare carbon fibre electrode. This method was approximately seven times more sensitive for the determination of nitrite with a slope of 1.41 nA cm3 pg-1 compared with the method reported here. It had a shorter analysis time and comparable short-term reproducibility. However, the main problem associated with this electrode was its long-term stability.When monitored over a 9 d period, a gradual decrease in response was observed owing to stripping of the polymer from the carbon fibre surface. Severe interference from ascorbic acid was also evident. The long- term stability of the osmium polymer modified electrode reported here was shown to be far superior, and deemed to be of more analytical importance in the design of an operational sensor for nitrite. Determination of Nitrite in Meat For both the electrochemical method and the spectropho- tometric reference method, nitrite was extracted from the meat sample using a standard AOAC method.21 For the analysis using the modified microelectrode the extract was diluted 1 + 1 with 0.2 mol dm-3 Na2S04. Using a range of nitrite standards (0-12 pg cm-3 prepared from NaN02), a calibration graph was constructed, which had a correlation coefficient, r , = 0.999 and a regression equation as follows: y (nA) = 0.098~ (pg (3121-3) - 0.016.Electrochemical analysis of the meat sample yielded a result of 88.0 k 1.0 pg g-1 of NO2- (n = 2), which compares favourably with the value of 84.0 k 1.2 pg g-1 ( n = 2) obtained using the reference spectropho- tometric method.16 Thirty consecutive injections of the meat extract were made, without the need for pre-treatment of the electrode, resulting in no deterioration of the response. Hence, in addition to exhibiting electrocatalytic properties, the polymer acts as a protective membrane by inhibition of adsorption of matrix compounds such as proteins on the carbon fibre surface, and thus preventing surface fouling.The polymer also prevented the severe surface fouling encountered at solid electrodes on nitrite reduction.I3 The method has a moderate analytical throughput and is capable of handling 30 samples h-1. Interferences For electrocatalysis at redox polymers such as those described here, the Gibbs free energy for the cross-exchange reaction must be negative. This requirement imposes certain limita- tions on the types of reaction that can occur at electrodes modified with these materials. By exploiting this thermo-ANALYST, AUGUST 1992, VOL. 117 a) 2 min H t I b) T0.79 nA 2 min - Time Fig. 3 Flow amperometric responses of the osmium polymer modified electrode to consecutive injections of (a) 50 and ( b ) 25 pg cm-3 of N02-.Constant potential operation at -0.150 V versus SCE; electrolyte = 0.1 mol dm-3 H2S04; flow rate = 0.2 cm3 min-1 2 min H acid. This feature is of considerable advantage for the determination of nitrite. Considering the limitations imposed by the thermodynamic requirements for cross-exchange reac- tions, then only reduction reactions are feasible with the electrocatalyst in the 0s" redox state. The thiocyanate anion, another potential interferent, was also found to be electro-inactive at the redox polymer. This is because the formal potential for thiocyanate oxidation is more positive than the ElI2 of the electrocatalytic centre, thus eliminating possible surface-fouling effects caused by the oxidation of this anion. The only major interferent in the electrochemical method was found to be iron(n1).However, standard procedures are available to remove iron(1ir) prior to the determination of nitrite.16 Interference from iron(ii1) can also be eliminated by the addition of ethylenediaminetetraacetic acid (EDTA) to the sample and carrier electrolyte. The [ Fe( EDTA)] - com- plex formed has a formal potential more negative than the of the osmium polymer and consequently is thermodynami- cally removed from the reaction at the modified electrode. Conclusion The osmium polymer modified carbon fibre flow electrode has been demonstrated to be sensitive , accurate and reproducible for the determination of the nitrite ion in a meat sample. The electrode was also shown to be very stable over extended periods of use without changes in sensor sensitivity.The sensor can be operated under conditions where common interferents such as ascorbic acid and thiocyanate are non- electroactive. The lack of interference and high stability of the sensor are major advantages of this device. When compared with the standard spectrophotometric techniques, the modified microelectrode has a comparable analysis time; however; considerably wider linear ranges, increased sensitivity and lower limits of detection are charac- teristic of the modified microelectrode. Also, the proposed method does not require the use of potentially unstable complexing or colour-forming reactions. The microelectrode flow cell is also simple and inexpensive to produce and is easy to operate. -Time Fig. 4 Effect of ascorbic acid on the response of the modified electrode to nitrite.A, 500 pg cm-3 ascorbic acid; B, 50 pg cm-3 NO2-; and C, 500 pg cm-3 ascorbic acid plus 50 pg ~ m - ~ NO2-. Operational conditions as for Fig. 3 dynamic limitation, the sensitivity and selectivity of sensors constructed from these materials can be controlled by the synthetic control of the E1/2 of the electrocatalytic centre and by the manipulation of the formal potential of possible analytes and interferents. As the osmium polymer has an Ellz of 0.250 V versus SCE," it therefore has wide scope for mediating reduction reactions. This provides a material that can be used in a range of analytical applications, with sensitivity and selectivity being controlled by the of the redox centre. The ascorbate anion is a common interferent when determining NO2- in processed meats, as sodium ascorbate is frequently added as an antioxidant.This com- pound does not interfere using the osmium polymer modified electrode, as the ascorbate ion cannot undergo oxidation at the modified electrode when the sensor is operating in the 0s" state. This can be seen in Fig. 4, A, when an injection of 500 pg cm-3 ascorbic acid produces a background response. In Fig. 4, B and C the response for 50 pg cm-3 of nitrite and 50 pg cm-3 of nitrite plus 500 pg (3111-3 of ascorbic acid solutions is shown. It can be seen that interference in the determination of nitrite does not occur even with a 10-fold excess of ascorbic 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 References Lijinsky, W., and Epstein, S.S., Nature (London), 1970, 223, 21. Barnes, H., and Folkard, A. R., Analyst, 1951, 76, 599. Chang, S., Kozeniauskas, R., and Harrington, G. W., Anal. Chem., 1977,49, 14. Zhao, Z., and Cai, X., J. Electroanal. Chem.. 1988, 252,361. Gao, Z . , Qing, G., and Zhao, Z., Anal. Chim. Acta, 1990,230, 105. Markusova, K., and Fedurco, M., Anal. Chim. Acta, 1991,248, 109. Cox, J. A., and Kulesza, P. J., J. Electroanal. Chem.. 1984,175, 105. Newbery, J. E., and Lopez de Haddad, M. P., Analyst, 1985, 110, 81. Chamsi, A. Y., and Fogg, A. G., Analyst, 1988, 113, 1723. Cox, J. A., and Kulkarni, K. R., Analyst, 1986, 111, 1219. Barisci, J. N., Wallace, G. G., Wilke, E. A., Meaney, M., Smyth, M. R., and Vos, J. G., Electroanalysis, 1989, 1,245. Wallace, G. G., Meaney, M., Smyth, M. R., and Vos, J. G., Electroanalysis, 1989, 1, 357. Mengoli, G., and Musiani, M. M., J. Electroanal. Chem., 1989, 269,99. Doherty, A. P., Forster, R. J., Smyth, M. R., and Vos, J. G., Anal. Chim. Acta, 1991, 255,45. Pons, S., and Fleischmann, M., Anal. Chem., l987,59,1391A. Vogels Textbook of Quantitative Inorganic Analysis, Longman, New York, 4th edn., 1978, pp. 97, 158 and 755. Forster, R. J., and Vos, J. G., Macromolecules, 1990,59,4372.ANALYST, AUGUST 1992, VOL. 117 1263 18 Clear, J. M., Kelly, J. K., O’Connell, C. M., and Vos, J. G., 1. Chem. Res. ( M ) , 1981, 3039. 19 Hua, C., Yunping, W., Chunghan, J., and Tonghui, Z., Anal. Chim. Acta, 1990,235,273. 20 Hua, C., Sagar, K. A., McLaughlin, K., Jorge, M., Meaney, M. P., and Smyth, M. R., Analyst, 1991, 116, 1117. 21 Official Metho& of Analysis of the Association of Official Analytical Chemists, AOAC, Washington, DC, 13th edn. , 1980. 22 Murray, R. W., in Electroanalytical Chemistry, ed. Bard, A. J., Marcel Dekker, New York, 1984, vol. 13, p. 240. 23 Denisevich, P., Abruna, H. D., Leidner, C. R., Meyer, T. J., and Murray, R. W., Inorg. Chem., 1982,21,2153. 24 Ikeda, T., Shmehl, R., Denisevich, P., Willman, K., and Murray, R. W., J. Am. Chem. SOC., 1982,104,2683. 25 Silva, M., Gallego, M., and Valchrcel, M., Anal. Chim. Acta, 1986,179, 341. Paper 2f00413E Received January 27, 1992 Accepted April 23, 1992