ANALYST, AUGUST 1991, VOL. 116 797 Amperometric Monitoring of Sulphur Dioxide in Liquid and Air Samples of Low Conductivity by Electrodes Supported on Ion-exchange Membranes Gilbert0 Schiavon" and Gianni Zotti National Research Council, Institute of Polarograph y and Preparative Electrochemistry, Corso Stati Uniti 4, 1-35100 Padua, Italy Rosanna Toniolo and Gin0 Bontempelli" Institute of Chemistry, University of Udine, Hale Ungheria 43, 1-33100 Udine, Italy An amperometric sensor is described for the determination of sulphur dioxide in both gaseous atmospheres and solutions of low conductivity. It consists of a porous Pt electrode (facing the sample) supported on one face of an ion-exchange membrane (Nafion 417) which serves as a solid polymer electrolyte. The other side of the membrane faces an internal electrolyte solution (1 mol dm-3 aqueous perchloric acid) containing the counter and reference electrodes.This sensor is inserted into a flow cell in which gaseous or electrolyte-free aqueous samples are fed by a peristaltic pump placed in a closed-loop path and SO2 is oxidized at an applied potential of 0.65 V versus Ag-AgCI. The device is found t o be characterized by a high current sensitivity and a short response time, 24 A cm-2 mol-1 dm3 and 1 s respectively for gaseous samples; 0.4 A cm-2 mol-1 dm-3 and 4s, respectively, for water solutions), and by good stability and low background noise. The dynamic range extends up t o 2 x mol dm-3 (gaseous samples) and 1 x 10-3 mol dm-3 (water samples) with good linearity, and detection limits of 8 x 10-9 mol dm-3 (gaseous samples) and 4 x 10-7 mol dm-3 (water samples) are predicted for a signal-to-noise ratio of 3.The advantages offered by this type of sensor over conventional gas-permeation membrane electrodes are discussed. Keywords: Sulphur dioxide; amperometric sensor; ion exchange membrane-supported electrode; elec- trolyte-free sample; electroanalysis Assessment of ecosystem acidification resulting from air pollution necessitates monitoring of the airborne concentra- tions of the contributing acidifying gaseous species, including the trace gas S02. Similarly, monitoring of the SO2 contam- inating various environmental liquid media is of essential interest. Moreover, such a species is used as a reducing agent or as a reactant in a number of industrial processes where the effective use of SO2 is largely dependent on the availability of in situ and instantaneous SO2 sensors suitable for its control, particularly in gaseous media.Several methods and instruments are available to monitor SO2 levels both in gaseous and liquid media using a great variety of systems.1.2 As every instrument has its own characteristics and limitations, new approaches to the detec- tion and analysis of SO2 appear to be desirable. In principle, analysers based on the direct electrochemical oxidation of this polluting species should be characterized by excellent results and should possess the required attributes for providing in situ measurements of SO2 and for its continuous monitoring. Unfortunately, these electroanalytical procedures cannot be applied directly to air samples or to liquid media with a low conductivity in which a sufficient concentration of supporting electrolyte is not present. The addition of electrolyte would cause contamination or perturbation of the natural solution equilibria in polar media, and such an addition is impractic- able in media that have very low dielectric constants because they are unable to dissolve ionic species.In order to overcome this drawback, membrane electrodes3-7 may be adopted but their performance is strongly affected by the transfer of analyte through the gas-permeable membrane, which proves to be a critical step. A fairly low sensitivity and fairly long response time are peculiar to these electrodes and their response also exhibits a high temperature dependence result- ing from the consequent change in the permeability of the membrane. * To whom correspondence should be addressed.The monitoring of SO2 by using an electroanalytical sensor suitable for the determination of electroactive analytes present in gaseous media or in solvents of high resistivity, described previously by Schiavon and co-workers,8-10 is reported. It is based on the use of a porous working electrode supported on one surface of an ion-exchange membrane whose other surface is in contact with an electrolyte solution containing the counter and reference electrodes. The porous working electrode faces the analyte sample in which no supporting electrolyte is needed because electroneutrality in the neighbourhood of the electrode surface is maintained by ionic migration through the ion-exchange membrane.Thus any membrane-permeation step is avoided and the analyte is monitored directly in its natural medium as the membrane separating the sample from the internal electrolyte does not act as a filter for gaseous molecules but serves to ensure the transfer of charged species from the working to the counter electrode. A similar approach has also been adopted by other workers. 11-13 Experimental Chemicals and Instrumentation All the chemicals used were of analytical-reagent grade (obtained from Carlo Erba, RPE) and were employed without further purification. Nitrogen atmospheres containing known amounts of SO2 and stock solutions of SO2 in water obtained from a Milli-Q system (Millipore) (de-aerated preliminarily with nitrogen) were prepared by diluting suitable known amounts of the pure gas.They were standardized by the iodine method14 and diluted further to the desired concentration with nitrogen or Milli-Q water. Voltammetric and amperometric measurements were per- formed by using a three-electrode unit equipped with an Amel 551 potentiostat driven by an Amel 568 digital logic function generator. The recording device was a Hewlett-Packard798 ANALYST, AUGUST 1991, VOL. 116 7080A measurement plotting system. All the tests were conducted at room temperature. Coating Procedure The ion-exchange material used as the solid polymer elec- trolyte (SPE) was a porous Nafion 417 cationic perfluorinated membrane 0.425 mm thick (Aldrich), reinforced with poly- tetrafluoroethylene (PTFE), which was cleaned by boiling in concentrated HN03 for 1 h and then in Milli-Q water for 1 h.The membrane was cut into discs of 1 cm in diameter which were equilibrated in aqueous 1 mol dm-3 NaC104 for 3 h and then in aqueous 1 X 10-3 rnol dm-3 [Pt(NH3)4]C12 (Johnson Matthey) (5 ml of solution for each disc). Following the Pt" loading, these discs were exposed to air on one side and on the other side to a hot (50-60 "C) aqueous alkaline solution (pH = 10) of NaBH4 (1 x 10-3 rnol dm-3) for 3-4h.15716 The concentration and volume of the PtIr solution was adopted in order that a uniform Pt film, 0.5 pm in thickness, could be formed upon total reduction by NaB&. The concentration of this last reducing solution, together with its temperature and the contact time were chosen so as to lead to the formation of a smooth and strongly adherent film located mainly on the surface of the membrane (thus enabling good contact with a Pt wire permitting electrical connection), with a minor portion of the deposited Pt embedded inside the membrane, in order to ensure good adherence of the electrode even under high mechanical tension.Such a Pt distribution can be accom- plished because once the Pt exchange membrane is soaked in the reducing solution, Pt" cations located near the surface are reduced to the metallic state, thus giving rise to a concentra- tion gradient between the bulk of the membrane and its surface which causes further Pt" cations to diffuse towards the surface of the membrane where they are in turn reduced.Different conditions for the Pt deposition step led to less effective electrode films characterized by lower current densities. A decrease of the Pt loading caused poor inter- particle contact, while thicker films in less intimate contact with the ion-exchange sites of the membrane were formed at higher Pt loadings. Electrode Assembly After completion of the coating procedure, Nafion discs were equilibrated in aqueous 1 rnol dm-3 HC104 and then installed in the electrode assembly shown schematically in Fig. 1. The cells were constructed as follows. A disc of Pt-coated ion-exchange membrane was clamped, with the porous conductive layer directed downward, at the bottom of a Pyrex cylinder, the end of which was threaded to a drilled PTFE holder sealing the assembly by means of an elastic O-ring resistant to acids and bases.Electrical contact with the Pt working electrode was made by placing a platinum ring on the inner side of the drilled holder, onto which the border of the Pt film deposited on the ion-exchange membrane was pressed. A Pt wire welded to this Pt ring and piercing through the holder permitted electrical connection. Such an assembly allowed about 0.4 cm2 of Pt film to be exposed to external media. The uncoated side of the membrane faced an internal compartment containing the 'internal electrolyte' (3 ml of aqueous 1 mol dm-3 HC104 solution) and equipped with an aqueous Ag-AgC1 reference electrode and a Pt counter electrode. This compartment was obtained by sealing the Pyrex cylinder to the bottom of a standardized hollow glass stopper in order to make possible the use of this electro- analytical device as a gas-tight stopper for the standardized glass flow cell (shown schematically in Fig.Z), in which all the experiments were conducted. This device enables electrochemical processes to occur in media of high resistivity, without the need for supporting electrolytes, according to the principle illustrated in Fig. 1 C R . a -b - -d Fig. 1 Schematic view of the Pt covered Nafion electrode. a, Internal compartment filled with aqueous 1 mol dm-3 HC104 and equipped with C, a Pt counter electrode and R, a Ag-AgC1 reference electrode; b, Nafion ion-exchange membrane; c, porous Pt coating; d, Pt-ring collector; and e, sample Counter Reference Sa m p I e -- stream Fig.2 Schematic view of the flow cell. Sample volume = 20 ml (enlarged view). When the platinum electrode is held at an appropriate potential, the substrate (SO2) in the working sample is oxidized to yield sulphate and hydrogen ions (the hydrogen ions being trapped in the membrane), coupled with an ionic migration through the membrane to maintain electroneutrality. Concurrently, hydrogen ions in the counter solution are reduced at the counter electrode, thus restoring the ionic content of the internal electrolyte. Hydrogen produced in this cathodic reaction is periodically removed by purging with nitrogen . The flow cell was fed with N2-S02 atmospheres or SO2 solutions in Milli-Q water whose flow rate was kept constant, unless stated otherwise, at 100 ml min-1 by a peristaltic pump placed in a closed-loop path.A flow meter was inserted in the stream in order to monitor this flow rate.799 ANALYST, AUGUST 1991, VOL. 116 4.0 I N I Eu a 2.0 E a 0 0.5 E N 1 .o Fig. 3 Voltammogram recorded at a Pt-Nafion electrode, with aqueous 1 mol dm-3 HC104 as the internal electrolyte, with a nitrogen stream containing SO2 (0.13 mmol dm-3; 8.3 mg 1-1; 2900 ppm v/v). Scan rate, 0.05 V s-1; and flow rate, 100 ml min-1. Current densities refer to the geometric area of the electrode Results and Discussion Anodic Behaviour of SOz at PtSPE Electrodes Platinum porous films have been selected as working elec- trodes because such an electrode material assures an over- voltage which is as small as possible for the SO2 oxidation process.17 A typical steady state linear sweep voltammogram recorded at a Pt-covered Nafion electrode with 1 mol dm-3 aqueous HC104 as the internal electrolyte and a nitrogen stream containing SO2 flowing into the closed-loop flow cell is shown in Fig.3. A single well-formed and reproducible anodic wave is observed, whose limiting current extends over a fairly wide range of potentials. This wave, whose irreversible character has been confirmed by applying the usual criteria to cyclic voltammograms recorded in stationary N2-SO2 atmos- pheres, resembles closely that found at conventional platinum electrodes for the oxidation of SO2 dissolved in aqueous HC1045.17 (the medium used as the internal electrolyte for wetting the rear side of the Nafion membrane). Fairly similar voltammograms, located at the same poten- tials, were also recorded at the same Pt-SPE electrode in electrolyte-free water in which SO2 was dissolved, regardless of the concentration.This last result is rather surprising as a progressive anodic shift of the SO2 oxidation process is expected with a decrease in the amount of SO2 present, owing to the concomitant increase in pH (from about 2.3 up to 7.0 for concentrations decreasing from 1 x 10-2 to 1 x 10-9 rnol dm-3) which causes increasing fractions of SO2 to be present in the solution as the conjugate base HS03-. This apparent contradiction however can be easily explained by considering the irreversibility affecting the electron transfer involved in this anodic oxidation17 which is able to mask the dependence of potential on pH, thus leading to the coin- cidence of voltammograms recorded on strongly acidic and basic buffered SO2 solutions.5 Alternatively, it can be accounted for by admitting that the pH governing the anodic reaction is that of the internal electrolytes rather than the pH of the sample.For both aqueous and gaseous samples, no progressive decrease in the over-all voltammogram was observed in the second and subsequent scans, contrary to the results obtained at conventional platinum electrodes for SO2 aqueous solutions containing supporting electrolytes.'* The ability of ion- exchange membranes to prevent the passivation of Pt elec- trodes, on which they are present as a film, has been previously observed in the surface-conditioned oxidation of methanol occurring at Nafion-covered Pt electrodes.19 On the basis of these voltammograms, a potential of 0.65 V versus an Ag-AgCI electrode was applied to the Pt-SPE electrode when performing the subsequent amperometric measurements of SO2 in both gaseous imd liquid samples. 0.5 N I Eu a E 0.25 0 2 4 6 Time/min Fig.4 Current-time profile recorded at a Pt-Nafion electrode, with aqueous 1 rnol dm-3 HC104 as the internal electyrolyte, on nitrogen streams containing SO2 in the following concentrations: A, 4.7 X mol dm-3 (105 pm); B, 9.4 x 10-6 rnol dm-3 (210 pm); C, 15.1 x mol dm-3 6 3 8 ppm); D, 19.5 x 10-6mol dm-3 637 ppm); and E 23.7 X 10-6 mol dm-3 (531 pprn). Applied potential, 0.65V; flow rate. 100 ml min-1 Table 1 Performance of the amperometric SPE sensor Gaseous Water Feature Unit* samples samples Sensitivity A cm-2 mol-1 dm3 24 0.4 pA cm-2 ppm-l 1.07 6.25 Response Detection time S 1 4 limit mol dm-3 8 x 10-9 4 x 10-7 I.18 I-' 0.6 26 PPb 180 26 Dynamic range mol dm-3 Up to 2 x 10-4 Up to 1 x 10-3 mg 1-1 Up to 12.8 Up to 64 PPm u p to 4500 Up to 64 * For gaseous samples, ppb and ppm (v/v) are employed.Such a value for the potential was selected in order to maximize the current detection sensitivity of the sensor and, at the same time, to minimize its dependence on the working potential. Performance of the Amperometric Sensor In order to test the performance of the Pt-covered Nafion electrode as an amperometric sensor for SO2 in gaseous atmospheres, nitrogen streams containing known and increas- ing concentrations of SO2 were fed into the flow cell shown in Fig.2, typically at a flow rate of 100 ml min-1, and the SO2 content was monitored by measuring the current flowing when a potential of 0.65 V was applied to the working Pt film. The results obtained are summarized in Fig. 4 which shows a typical current-time response recorded during these measure- ments. Each addition of SO2 causes a rapid rise in the current which attains a satisfactory constant value in a fairly short time (about 15 s). This constant current was found to be reprodu- cible within +3% and to depend linearly on the concentration of SO2 over a fairly wide range which extends up to about 2 X 10-4 rnol dm-3 (12.8 mg 1-1,4500 ppm v/v). From a plot of this steady state current density versus SO2 concentration obtainedANALYST, AUGUST 1991, VOL.116 in the range 5 X 10-8-5 x 10-6 mol dm-3, a sensitivity of 24 A cm-2 mol-l dm3 (1.1 PA cm-2 ppm-l) was obtained with a correlation coefficient of 0.998. As the residual current density at the working potential was about 1 FA cm-2 with a standard deviation of about 0.06 pA cm-2 (background noise), a detection limit of approximately 8 x 10-9 rnol dm-3 (0.6 pg 1-1, 180 ppb v/v) could be achieved for a signal-to-noise ratio of 3. These results are summarized in Table 1 where they are compared with those obtained for electrolyte-free water samples containing SO2. Table 1 also gives the response time which cannot be inferred correctly from the current-time profiles shown in Fig. 4, as the time required to attain a steady-state current is due mainly to the inertia opposing the flowing system required in order to achieve equilibrium conditions after each increase in the concentration of S02.Consequently, the response time was investigated by carrying out suitable experiments. A 95% response was observed in 1 s when the electrode was transferred rapidly from air to the cell, shown in Fig. 2, fed with a gas stream containing 1 X 10-5 mol dm-3 (0.64 mg 1-1, 224 ppm v/v) of SO2. This response time remains practically unaffected when the electrode is transferred from air to gas streams with higher or lower contents of S02. The performance of the proposed Pt-SPE sensor with reference to the determination of SO2 dissolved in electrolyte- free water was evaluated after preliminary tests aimed at verifying the effect of the possible transfer of ionic species from the internal electrolyte to samples of high resistance by either permeation through the ion-exchange membrane or incidental leaks caused by defective sealing of the internal electrolyte compartment.With this purpose, a series of Pt-SPE electrodes, all containing aqueous 1 mol dm-3 HC104 as the internal electrolyte and prepared following the pro- cedure reported above, were soaked in Milli-Q water samples (20 ml) in which different amounts of SO2 were dissolved, so as to attain concentrations ranging from 1 x 10-6 to 1 x 10-3 rnol dm-3. The pH due to the presence of SO2 remained virtually unchanged over time (1 h) when each of the samples was monitored using a glass electrode. Only after longer times (about 5 h) could a slight decrease of pH be detected in the more diluted samples (for 1 X 10-6 mol dm-3 solutions a pH decrease from 6 to 4.8 was found), thus suggesting that any electrolyte leakage can be ruled out and no appreciable ionic transfer due to permeation through ion-exchange membranes can be expected to occur during the time typically required for the analysis of a sample (about 2 min).The results concerning solutions containing SO2 in elec- trolyte-free water, summarized in Table 1, were obtained by the same tests described above for nitrogen atmospheres containing S02. In all instances, current responses were practically unaffec- ted by the temperature (1040 "C) as expected on the basis of the absence of gas-permeation steps. Moreover these responses were independent of moderate variations in both the applied potential (k0.05 V) and the sample flow rate, provided that this last parameter was higher than a minimum value (about 50 ml min-1 for gaseous streams and about 80 ml min-1 for aqueous streams).The attainment of constant currents for flow rates above a threshold value can be accounted for by considering that SO2 molecules from the gaseous or liquid stream must diffuse across the thin porous Pt film before reaching the Pt-Nafion interphase where they undergo oxidation. This fact implies that the SO2 oxidation rate is controlled by a diffusion-permeation Pt layer, the thickness of which is independent of the flow rate. Conse- quently, no increase in the current is expected for flow rates above the threshold required to reduce the thickness of the stagnant layer near the electrode surface to a value that allows a more rapid diffusion within such a layer than within the porous Pt film.The long-term stability of these Pt-covered Nafion elec- trodes in gaseous or liquid streams containing SO2 appeared to be totally satisfactory in that no appreciable change in the current response was observed even after 2 months of continuous use. Different Pt-Nafion electrodes, constructed following an identical procedure, led to very similar responses (k4%) when tested on the same SO2 samples, thus indicating that quite reproducible electrode surfaces can be obtained by the coating procedure adopted here. In conclusion, SPE electrodes appear to be useful alterna- tives to conventional amperometric membrane electrodes from which they differ in their inherent conceptual approach.The polymeric membrane in the latter sensors serves as a gas-permeable interphase between the sample and the indica- tor electrode, while Nafion in SPE electrodes acts as an 'ion-permeable, membrane separating the internal electrolyte from the working electrode, which is in direct contact with the sample. Thus, any gas-permeation step is avoided and it is this feature that makes SPE electrodes preferable to conventional membrane electrodes. Thus, sensitivity and response time for the detection of SO2 in gaseous atmospheres, as reported in Table 1 for Pt-covered Nafion electrodes, are undoubtedly better than those found at conventional membrane electrodes [0.3 pA cm-2 ppm-1 and 30s (reference 4) or 0.04 pA cm-2 ppm-1 (reference 7)].In addition the lower detection limit (ppb rather than ~pm4.7)~ the wider dynamic range (up to 4500 ppm as opposed to 1500 ppm7) and the negligible effect of temperature, must be considered among the advantageous characteristics deriving from the absence of a gas-permeation step. Moreover, the electroanalytical sensor proposed in the present paper is advantageous even when compared with a device proposed recently which permits the amperometric determination of SO2 in air samples by a closed-loop flow injection system containing a regenerable chemical probe.20 These advantages notwithstanding, the proposed SPE electrodes cannot be adopted for direct measurements of SO2 in ambient air as a lower limit of detection (of about one order of magnitude) is required.However, they can be used successfully for the direct determination of SO2 in power plant plumes and for industrial hygiene measurements, without the need for preconcentration. All of the characteristics mentioned above make this type of electrode particularly attractive for the monitoring of any electroactive species present in gaseous phase or highly resistive solutions, provided that an appropriate working potential is chosen. Selectivity, however, may be, in principle, unsatisfactory in some instances, owing almost exclusively to the value of the working potential. Thus, for the monitoring of SO2, it has been found that the oxidation of two possible interfering species, NO and H2S, can occur at the high positive potential applied (0.65 V).It must be remarked however that such an insufficient selectivity is not conditioned by the electroanalytical SPE sensor proposed here, but is peculiar to amperometric measurements. Nevertheless, the interferences mentioned are not a real drawback in this instance as much lower concentrations are usually expected for NO and H2S with respect to SO2 so that their effect on the measurement of SO2 can be neglected. Moreover, the chemical nature of these interfering species is such as to permit their easy removal prior to SO2 detection by passing the sample through suitable reaction columns. We thank S. Sitran of the National Research Council, Institute of Polarography and Preparative Electrochemistry (Padova) for skilful experimental assistance.The financial aid of the Italian National Research Council and of the Ministry of the University and of the Scientific and Technological Research is gratefully acknowledged.ANALYST, AUGUST 1991, VOL. 116 801 1 2 3 4 5 6 7 8 9 10 References Hollowell, G. D., Gee, G. Y., McLaughlin, R. D., Anal. Chem., 1973,45,63A. Sickles, J. E., and Grohse, P. M., Sampling and Analysis Methods for Sulphur Dioxide and Nitrogen Dioxide: A Litera- ture Review, Research Triangle Institute Report no. RT1/2823/ 00-011, Research Triangle Institute, Research Triangle Park, NC, 1984. Shaw, M., and Thatcher, I . , US Patent 40 17 373. Chem. Abstr., 1977,87.055005g. Bruckenstein, S . , Tucker, K. A., and Gifford, P. R., Anal. Chem., 1980,52, 2396. Bergman, I., J. Electroanal. Chem., 1983, 157, 59. Langmaier, J., Opekar, F., and Pacakova, V., Talanta, 1987, 34,453. Langmaier, J., Polak, J., and Opekar. F., Analyst, 1988, 113, 501. Schiavon, G., Zotti, G., and Bontempelli, G., Anal. Chim. Acta, 1989, 221,27. Schiavon, G., Zotti, G.. Bontempelli, G., Farnia, G., and Sandona, G., Anal. Chem., 1990, 62, 293. Schiavon. G., Zotti, G., Toniolo, R., and Bonternpelli, G., Efectroanalysis, 1991, in the press. 11 12 13 14 15 16 17 18 19 20 Kaaret, T. W., and Evans, D. H., Anal. Chem., 1988,60,657. De Wulf, D. W., and Bard, A. J., J. Electrochem. SOC., 1988, 135, 1977. Harth, R., Mor, U., Ozer, D., and Bettelheim, A., J. Electrochem. SOC., 1989,136,3863. Jeffery, G. H., Bassett, J., Mendham, J., and Denney, R. C., Vogef’s Textbook of Quantitative Chemical Analysis, Longman, Harlow, 1989, p. 398. Millet, P., Pineri, M., and Durand, R., J. Appl. Electrochem., 1989, 19, 162. Fedkiw, P. S., and Her, K. W., J. Electrochem. Soc., 1989,136, 899. Zhdanov, S. I., in Encyclopedia of Electrochemistry of the Elements, ed. Bard, A. J., Marcel Dekker, New York, 1975, Spotnitz, R. M., Colucci, J. A., and Langer, S. H., Electrochim. Acta, 1983,28, 1053. Enea, O., J. Electroanal Chem., 1987, 235, 393. Rios, A., Luque de Castro, M. D., Valcarcel, M., and Mottola, H. A., Anal. Chem., 1987, 59, 666. Paper 1 /00349F Received January 23rd, 1991 Accepted April 4th, 1991 vol. IV, pp. 314-335.