- Email: [email protected]
Flow injection analysis of sulfite ion with a potentiometric titanium phosphate–epoxy based membrane sensor
Talanta 54 (2001) 773– 782 www.elsevier.com/locate/talanta
Flow injection analysis of sulfite ion with a potentiometric titanium phosphate –epoxy based membrane sensor Saad S.M. Hassan a,*, Sayed A. Marei b, I.H. Badr a, H.A. Arida b a
Department of Chemistry, Faculty of Science, Ain Shams Uni6ersity, Cairo, Egypt b Hot Laboratory Center, Atomic Energy Authority, Inshas, Egypt
Received 5 October 2000; received in revised form 4 January 2001; accepted 22 January 2001
Abstract A potentiometric sensor based on the use of titanium phosphate (TP) in epoxy matrix membrane is prepared and characterized. The sensor exhibits near-Nernstian response for many anionic species over the concentration range 10 − 1 –10 − 5 mol l − 1. The origin of response is explained on the basis of the conversion of titanium phosphate cation exchanger into hydrated titanium oxide anion exchanger by the effect of the high pH of the epoxy matrix. The sensitivity and selectivity of the sensor for sulfite ions are optimized by conversion of sulfite into gaseous SOx by acidification, and diffusion of the gas through a membrane-based gas dialyzer followed by potentiometric detection of sulfite ions formed within a flowing recipient stream. No interferences are caused by many common anions and acidic gas releasing species except sulfide and nitrite ions. Determination of sulfite ion at levels as low as 10 − 4 mol l − 1 or less in the presence of nitrite and sulfide ions is performed by using a modified carrier buffer stream (10 − 2 mol l − 1 MES, pH 5.0 containing sulfamic acid) and pretreatment with Pb2 + . Advantages offered by the proposed gas dialyzer/flow injection system with TP–epoxy membrane based sensor over traditional ion exchange based sensors includes long life time ( \ 8 months), excellent stability and reproducibility ( 1 mV), fast response time ( B 30 s), wide pH working range (pH 5–9), high sample throughput ( 60 samples h − 1), low detection limit (8 × 10 − 6 mol l − 1) and high thermal stability (up to 80°C). © 2001 Elsevier Science B.V. All rights reserved. Keywords: Inorganic ion exchangers; Sulfite sensor; Potentiometry; Flow injection analysis; Gas dialyzer; Titanium phosphate ion exchanger; Epoxy membrane
1. Introduction Methods in current use for the determination of sulfite ion include gravimetry [1], spectrophotome-
* Corresponding author. Tel./fax: + 20-2-6822991. E-mail address: [email protected] (S.S.M. Hassan).
try [2–7], chromatography [8–12], chemiluminescence [13 –15] and amperometry [16 –18]. Most of these methods, however, involve several manipulation steps, not used for turbid and colored solutions and require sophisticated instruments. Although potentiometric methods of analysis are simpler, attractive, much faster and less expensive, little is known about the use of potentiometric sensors for selective determination of sulfite ion.
0039-9140/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 0 1 ) 0 0 3 3 0 - 7
774
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
A potentiometric sensing membrane probe for sulfite ion consisting of a flat-bottom glass electrode located behind a porous polymeric membrane whereby a thin film of sulfite electrolyte solution is sandwiched between the electrode and the membrane has been described [19]. Sulfur dioxide released by acidification of sulfite solutions diffuses through the porous membrane, rapidly equilibrates and changes the pH of the electrolyte. Sulfite biosensors based on the use of immobilized yeast, bacteria or sulfite oxidase enzyme and an oxygen probe have been described for determination of sulfite in water, environmental materials, food and wine samples [20–25]. These gas sensors and biosensors suffer from significant interference by many gas releasing anions, low stability and long response time. Guanidinium ionophore based sulfite selective membrane sensor has been suggested [26]. A PVC membrane sensor based on the use of mercury(II) diethyldithiocarbamate complex has been described for potentiometric determination of sulfite in wine [27,28]. Metal wires coated with graphite– epoxy resin and metal mercury(II) methylpiperidine dithiocarbamate have also been suggested as sensors for sulfite ion [29]. A cluster compound film sensor prepared by impregnation of a copper wire into a solution of (NH4)2MoS4 followed by oxidation and coating with PVC membrane displayed linear response for sulfite [30]. Some of these sensors, however, suffer from a narrow response range, short life span, long response time and significant interferences by most common anions. The present work describes the preparation and characterization of a novel potentiometric sensor based on the use of titanium phosphate (TP) ion exchanger as an electroactive material and epoxy resin as a matrix membrane. The high chemical and thermal stability of TP ion exchanger [31] is expected to provide a suitable sensor with long life, high reproducibility and stability. Incorporation of the proposed sensor in a flow injection cell with gas dialyzer provides an efficient flow-injection detector for determination of sulfite with an enhancement selectivity. The sensor also offers the advantages of simplicity, sensitivity, reliability, automation feasibility and low cost.
2. Experimental
2.1. Reagents All chemicals used were of analytical reagent grade unless otherwise stated and the standard solutions and buffers were prepared with distilled, de-ionized water. 4-Morpholino-ethanesulfonic (MES) and tetrahydrofuran (THF) were obtained from Aldrich Chemicals Co. A commercially available metal free epoxy adhesive (Araldite super glue, Hollis, NY 11423) was obtained from Ciba-Geigy Co. The sulfite-sensing flow injection system utilized a 0.5 M H2SO4 as simple diluent, and a 0.01 M MES buffer of pH 5 (adjusted with 50% (w/v) NaOH solution) as a recipient carrier stream. Standard working sulfite solutions (10 − 1 –10 − 5 mol l − 1) were freshly prepared daily and stabilized by dilution of appropriate aliquots of the stock sodium sulfite solution with distilled de-ionized water containing 5% (v/v) glycerol to mask the heavy metals.
2.2. Equipment All potentiometric measurements were made at ambient temperature (259 1°C) using an Orion digital pH/mV meter (Model SA 720) and TP–epoxy membrane based sensor in conjunction with Hana silver-chloride single junction reference electrode filled with 10% (w/v) KCl. A combination Ross glass pH electrode (Orion 81-02) was used for all pH measurements. A schematic diagram of the flow injection analysis (FIA) system manifold used for sulfite quantitation is shown in Fig. 1. The system consists of a laboratory made flow cell containing the sulfite sensor, peristaltic pump (MSREGLO), and an Omnifit 4-port injection valve (Omnifit, Cambridge, UK). A home made modified gas dialyzer similar to that described earlier [32] was used. Each channel contains two support ribs to hold the gas-permeable membrane securely in place, in order to reduce the movement of the membrane. The dialyzer was fitted with a polytetrafluoroethylene membrane (0.2 mm pore size). The recipient buffer (10 − 2 mol l − 1 MES, pH 5.0), sample diluent 0.5 mol l − 1 H2SO4 solution and
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
the sample stream were supplied from separate reservoirs through 1.0, 0.4 and 0.4 mm i.d. Tygon tubings and propelled with the flow rates 4, 4 and 2 ml min − 1, respectively. A mixing coil of 6 cm length and 0.8 mm i.d. was used. The potentiometric signals were monitored with a flow through TP –epoxy based membrane sensor and an Orion SA 720 digital pH/mV meter and displayed on an (X-T) strip chart recorder (Linear 1200).
2.3. Titanium phosphate ion exchanger Amorphous TP was prepared as described earlier [33]. A 20.8 ml of TiCl4 (pure TiCl4 sp. g 1.73 g ml − 1) was diluted to 1 l with 4 mol l − 1 HCl. The resulting solution was slowly added with continuous stirring to 1 l of 12% (v/v) phosphoric acid in 4 mol l − 1 HCl. The precipitated TP gel was left to settle for 48 h before filtration and washing with 4 mol l − 1 HCl. The precipitate was then washed by double distilled deionized water till the acidity of the aqueous washing solution reached pH 3. The formed TP was then dried at 40°C and the final cake was disintegrated by soaking in hot double distilled water at 70°C. The granules produced was filtered and dried at 80°C
775
for 2 h. After sieving, the portion of 30 mesh size was used for subsequent work.
2.4. Titanium phosphate–epoxy membrane sensor The sulfite membrane sensor was prepared by thorough mixing of 0.6 g finely powered TP ion exchanger with 0.4 g of epoxy resin. The resulting paste was spread with a glass rod over a piece of filter paper to give a thin membrane and lift in air to cure. After drying and hardening of the membrane, a circular disc of 10 mm diameter and 1 mm thickness was cut, and fixed to the end of a PVC tube with epoxy resin. The tube was filled with a mixture of equal volume of 10 − 2 mol l − 1 KCl and 10 − 2 mol l − 1 Na2SO3. An Ag/AgCl internal reference electrode was immersed into the solution and the sensor was soaked overnight in a solution of 10 − 2 mol l − 1 Na2SO3. Sensor calibration was carried out by placing TP –epoxy membrane sensor in conjunction with a single junction Ag/AgCl reference electrode in 10 − 1 –10 − 6 mol l − 1 standard sulfite solutions. Calibration plot connecting the relation between logarithm sulfite concentration and sensor response in mV was made and used for subsequent measurements of unknown samples. Sensor selectivity was evaluated by measuring the responses of the sensor toward 10 − 5 –10 − 1 mol l − 1 sulfite solution and different interferent anion solutions (sodium or potassium salts). Calibration of the sensor was made in each tested anion solution under static conditions (batch measurements) and under a flow injection conditions in the presence and in absence of the gas diffusion cell.
2.5. Gas dialyzer/flow through sensor detector
Fig. 1. Schematic diagram of flow injection dialyzer manifold for determination of sulfite: (A) 10 − 2 M MES buffer pH 5; (B) 5× 10 − 1 M H2SO4; (C) sulfite test sample solution; (D) peristaltic pump; (E) pulse suppressor; (F) mixing coil; (G) flow injection valve; (H) dialysis chamber; (I) gas permeable membrane; (J) waste; (K) flow through TP –epoxy membrane sensor; (L) Ag/AgCl reference electrode; (M) electric ground attached to a metal inlet tube; (N) electrolyte solution; (O) pH/mV meter; and (P) chart recorder.
Flow through potentiometric cell equipped with TP –epoxy membrane sensor with an internal filling solution consisting of 2 ml of an equal mixture of 10 − 2 mol l − 1 Na2SO3 and 10 − 2 mol l − 1 NaCl and an Ag/AgCl internal reference electrode was used. The sensor was conditioned after preparation by soaking the membrane sensor overnight in 10 − 2 mol l − 1 Na2SO3. The flow cell is a flow through cap type (Fig. 2) with a gap thin enough between the membrane and sensing ele-
776
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
electrode were connected to an Orion 720 pH/mV meter and attached to a strip-chart recorder to monitor the FI signals. The waste from the beaker was continuously removed by the peristaltic pump.
2.6. FIA measurements of sulfite
Fig. 2. Schematic diagram of the flow-through cell: (A) Ag/ AgCl reference electrode; (B) 10 − 2 mol l − 1 Na2SO3 +10 − 2 mol l − 1 NaCl internal filling reference solution; (C): TP – epoxy membrane; (D): carrier solution; and (E): waste.
ment, not to retain any of the flowing solution after injection cycle. The whole cell acted as a potentiometric flow-through detector with adjusted zero dead volume. The cell was assembled and integrated with the flow injection system as shown in (Fig. 1). A gas dialysis unit similar to that described earlier [32] was used in the flow injection system. The upper channel of the dialyzer acted as an injection loop. The used valve has two positions; one to allow the carrier buffer stream (10 − 2 mol l − 1 MES, pH 5.0) to flow through the upper channel of the dialyzer and then downstream through the flow detector. The second position of the valve allowed separation of the upper channel of the dialyzer, and released the recipient stream to flow directly through the sensor detection unit. The sample diluent (0.5 mol l − 1 H2SO4) was thoroughly mixed with the sulfite sample before passing through the lower channel of the dialyzer using T-shaped glass connector and mixing coil. Both the recipient steam and the sample diluent stream were followed in an opposite direction to each other, to ensure a good trapping of SOx gases. The flow cell detection unit was placed in conjunction with a single junction Ag/AgCl reference electrode in a beaker filled with the carrier electrolyte solution. The sulfite sensor and reference
Sulfite test solutions were injected in the sample diluent stream containing 0.5 mol l − 1 sulfuric acid and then entered a mixing coil where SOx gases were generated. At the moment the sample plug began to enter the dialyzer, the rotary injection valve controlling the flow of the recipient carrier stream was turned to allow this stream bypass the upper channel of the dialyzer unit. The carrier buffer stream provided a good washing of the sensor detection unit, at the same time the acidified sample flow through the lower channel of the dialyzer before discharge to the waste. After exactly 1 min, the rotary valve was positioned so that the recipient stream plug contained in the upper channel of the dialyzer, was injected through the detector. At the same time the sample was changed into the next concentration. After 20 s, the new sample began to enter the dialyzer and the injection valve turned black. A time of 20 s was sufficient to obtain a complete signal on the recorder. The tubing distance between the output of the upper channel of the dialyzer and the input of the flow through TP detector was 3 cm, in order to minimize the dilution of the sample by the recipient stream. At least three signals for each sample were recorded and their average potential change was measured.
3. Results and discussion
3.1. Sensor characteristics and origin of response Epoxy membrane based sensor containing TP, showed strong potentiometric responses for many − − common anions such as SO23 − , NO− 2 , Br , I , − − − 2− 2− 2− − Cl , F , NO3 , HPO4 , SO4 , CrO4 , IO4 and S2 − . Evaluation of the sensor under static mode of operation according to IUPAC recommendations [34] indicated that the sensor displayed a sub
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
Nernstian response (slope 259 0.5 mV decade − 1) over the concentration range 10 − 5 – 10 − 1 mol l − 1 of sulfite and the detection limit was 6× 10 − 6 mol l − 1 Fig. 3. No potentiometric response was noticed by using a pure epoxy resin (blank) membrane based sensor for almost all tested cation and anions. Deviation from Nernstian behavior is known with many inorganic ion exchanger based sensors and attributed to incomplete permselectivity of the membrane [35]. The response of the TP – epoxy sensor in partially non-aqueous media was also tested by repeated calibration in different aqueous-organic media (e.g. ethanol, acetone and n-butanol) with different compositions. The results obtained revealed that the response was strongly affected by the nature and percentage of organic solvent. The calibration plots were generally linear over a narrow sulfite concentration range in the nonaqueous media. Best linear response was only obtained in 20% (v/v) ethanol solution. Table 1 summarizes the response characteristics of TP –epoxy sensor under static and hydrodynamic modes of operation from data collected from four different sensor assemblies for each
Fig. 3. Potentiometric static response of TP –epoxy membrane based sensor for sulfite ion in 10 − 2 mol l − 1 MES buffer of pH 5.
777
Table 1 Response characteristics of TP–epoxy matrix membrane sensor for sulfite ion under static (batch) and hydrodynamic gas dialyzer (FIA) modes of operations Parameter
Batch
FIA
Slope, mV decade−1 Linear range, mol l−1 Lower limit of dection, mol l−1 Response time for [SO2− 3 ] change from 10−4 to 10−1 mol l−1, s Recovery time, s Life time, month Working pH range
25.0 90.5 10−5–10−1 6×10−6
42 9 1.0 10−5–10−1 8×10−6
B20
30 \8 5–9
20
40 \8 5–9
system. It can be seen that the linear range of the sensor was improved under the FIA due to the pre concentration. To study the mechanism and origin of the anionic response of TP–epoxy membrane sensor, the effect of both ion exchanger and epoxy matrix on the potentiometric response of sulfite were examined. Although, TP is a polymeric cationic exchanger with no adsorption properties towards anions, its use in epoxy matrix displayed unexpected anionic response. Agrawal and Abe [36] attributed the anionic response of a sensor incorporating antimonic(V) acid cation exchanger in epoxy resin (Araldite) to the contribution of the polyamide functional group (N+) of Araldite. This explanation, however, failed to explain the cationic response of many epoxy based membrane sensors. Epoxy membranes with 5, 10, 40 and 60 wt.% TP were prepared and tested. A sensor with a membrane incorporating 100 wt.% epoxy resin has neither anionic nor cationic response. As the concentration of TP increased in the membrane, the initial potential, linear range and slope of the calibration plot were significantly changed Fig. 4. A sensor with Araldite membrane consisting of 10–60 wt.% TP–epoxy exhibited the best anionic response towards sulfite ion (calibration slope − 25 mV decade − 1). It was difficult to prepare membranes containing more than 60 wt.% TP with acceptable mechanical and adhesive proper-
778
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
ties. Sensor with pure compressed TP pellet as a membrane gave cationic response for most monovalent cations with a calibration slope of + 44 mV decade − 1. The possible conversion of TP cation exchanger into an anionic exchanger in the epoxy matrix was examined. Measurement of the pH of aqueous leaching solution of uncured epoxy resin was found to be higher than pH 12. This high basicity of the matrix membrane was sufficient to convert TP into hydrated titanium oxide gel (TiO2 · xH2O)n, which is known as an anion exchanger [37]. The gel takes up H+ followed by equivalent amount of exchangeable anions. For this reason hydrated zirconium and titanium oxide gels have been used as sorbent for anions in cooling water reactors [38]. A pure TP pellet based sensor soaked in 1 mol l − 1 NaOH solution rapidly lost its cationic response and displayed an anionic slope of − 12 mV decade − 1 upon soaking for 48 h. Soaking for a longer period of time caused swelling and disintegration of the membrane. Based on these findings it seems that TP cation exchanger is partly converted in the strong basic epoxy matrix into hydrated titanium oxide anion exchanger.
Fig. 4. Effect of membrane composition on the potentiometric response of TP–epoxy membrane for sulfite ion: (A) 100% epoxy resin; (B) 95% epoxy + 5% TP; (C) 90% epoxy +10% TP; (D) 60% epoxy + 40% TP; (E) 40% epoxy + 60% TP; and (F) 100% TP.
3.2. Response time and stability The response time of TP–epoxy membrane sensor was tested for 10 − 5 –10 − 1 mol l − 1 sulfite solutions. The time required for the sensor to reach 95% of the steady potential response, after successive addition of 10–1000 ml aliquots of 10 − 1 mol − 1 standard sulfite to 100 ml of the buffer to obtain 10-fold differences in concentration, was almost less than 20 s. Typical dynamic response of TP–epoxy sensor for sulfite ion is shown in Fig. 5. The sensor reached its equilibrium response in a relatively short time (B 20 s) at 10 − 5 –10 − 1 mol l − 1 sulfite. The sensor exhibited rapid recovery time (30 s) to reach 95% of the base line potential even after being exposed to 1.0 mol l − 1 sulfite. Similar response and recovery times were obtained with the flow through detection cell under flow injection conditions. The sample throughput was about 60 sample h − 1. The potential displayed by the proposed sulfite sensor for consecutive measurements of 10 − 5 – 10 − 1 mol l − 1 standard sulfite solutions did not vary by more than 9 2 mV over a period of about 6 months. The changes in the calibration slopes did not exceed 9 1.0 mV decade − 1 change of sulfite concentration. However, the initial potentials displayed by different sensors prepared from different membranes did not vary by more than 9 10 mV. Nevertheless, the calibration slope, detection limit, linear range of response and selectivity coefficients values were almost reproducible over a study period of \ 8 months when the sensor was preconditioned and stored when not in use in Na2SO3 solution and the internal filling solution was changed every 4 weeks. These are further advantages of the present sensor over many of the earlier suggested sulfite ion sensor. Reproducibility studies under flow injection conditions revealed that variation of the sensor potentials did not exceed 2 mV, the slope fluctuated within 1 mV concentration per decade. The high reproducibility and stability of the sensor is attributed to the high stability of TP ion exchanger, the good adhesive properties of epoxy matrix and the less leachibility of the ion exchanger compared with other earlier reported
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
779
Fig. 5. Potentiometric dynamic response and recovery time of TP – epoxy membrane sensor toward sulfite solutions.
sulfite sensors [26– 30]. In addition, the sensor exhibited high thermal stability (20– 80°C) and resistance towards strong oxidizing agents and mineral acids.
3.3. Effects of pH and foreign anions The effect of pH of 10 − 2 and 10 − 3 mol l − 1 sulfite solutions on the sensor potential response was examined by monitoring the variation of potential with the pH change over the range of 2–11. The sensor potentials did not vary by more than 2 mV over the pH range 5– 9. All subsequent measurements were performed at a pH 5. However, the potential of the sensor was significantly affected below pH 4 and above pH 9. The potentiometric responses of the TP–epoxy sensor towards 12 different anions were examined
under static (batch) and hydrodynamic (flow injection) modes of operation. The selectivity coefficient data obtained (Table 2) show that, TP – epoxy sensor is not really selective for sulfite ion since most of the common anions interfere. The order of selectivity pattern for the tested anions is similar in both batch and FI mode of operations. The selectivity sequence is, S2 − \IO− 4 − − − 2− 2− \I−\NO− 3 \NO2 \Br \Cl \CrO4 \SO3 \ SO24 − \ F− \ H2PO− A comparison was 4 . made with the Hofmeister sequence [39,40], − − − − − − ClO− 4 \IO4 \ SCN \ NO3 \ I \ Br \ NO2 − − − − \ HSO3 \ Cl \ HCO3 = H2PO4 . Most often observed with organic ion exchanger based membrane sensor, which is based on the relative lipophilicity of the anions and their partition coefficients into the organic membrane. TP–epoxy membrane exhibits slightly different response
780
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
characteristics toward sulfite compared with some other anions. Elimination of all interfering anions was achieved by incorporation of the TP–epoxy sensor in a flow injection system with a gas diffusion cell.
3.4. Flow injection analysis of sulfite The potentiometric calibration response of TP– epoxy membrane sensor in FI system without gas diffusion cell was examined. A series of standard sulfite fresh solutions and other interferent anions were injected in separate experiments through the injection valve into the carrier stream (10 − 2 mol l − 1 MES buffer of pH= 5.0) using a single line flow injection system. At least three signals were obtained by three different injections for each determination. The data obtained showed significant interference by most lipophilic anion Fig. 6. It has been suggested that lack of selectivity over certain anions can be overcome by using a gas diffusion cell to prevent certain interfering anions from reaching the surface of the membrane sensor [32]. Upon using a flow injection system with gas diffusion cell and repeating the above experiments, responses were only obtained with sulfite, nitrite and sulfide ions. These ions were easily converted under the same acidic conditions to their volatile gaseous counter parts and induced a strong response over a linear range of 10 − 5 –10 − 1 mol l − 1 Fig. 7. Using the gas diffusion cell, the
Fig. 6. Potentiometric calibration response of TP – epoxy based sensor toward various anions.
calibration slope of sulfite increased to 429 1 mV decade − 1 due to preconcentration step during 1 min sample trapping time before reaching the
Table 2 Potentiometric selectivity coefficients (K pot SO3,B) of TP–epoxy matrix membrane sensor Interferent (B)
K pot SO3,B
HPO2− 4 F− SO2− 4 CrO2− 4 Cl− − Br NO− 3 IO− 4 S2− NO− 2 I−
2.1×10−1 2.5×10−1 3.0×10−1 3.9 8.3 19.0 91.2 7.5×102 7.5×104 25 2.2×103
Fig. 7. Potentiometric response of TP – epoxy membrane sensor for various anions using gas diffusion FIA system.
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
781
Fig. 8. Typical strip-chart recording obtained by sulfite by TP– epoxy membrane sensor-gas dialyzer.
surface of the membrane sensor. Typical stripchart recording of the response obtained with standard sulfite solutions is shown in Fig. 8. However, interferences by nitrite and sulfide ions were completely eliminated by using a carrier buffer stream containing 10 − 1 mol l − 1 sulfamic acid in 10 − 2 mol l − 1 MES buffer of pH 5 after sample pre-treatment with 10 − 2 mol l − 1 lead acetate solution. Sulfamic acid completely eliminated the effect of nitrite ion, and the sulfide ion was precipitated as PbS. Based on this approach a selective determination of sulfite ion in the presence of almost all anions was obtained using the FIA/gas dialyzer manifold system. Determination of sulfite at a concentration level of 40 mg ml − 1 –8 mg ml − 1 using the flow injection system showed an average recovery of 98.5% with a mean standard deviation (S.D.) of 1.2%. No interference was noticed due to the presence of 104-fold excess of most common anions.
4. Conclusion TP –epoxy matrix membrane sensor exhibits a near-Nernstian response for many common anions including sulfite ion with a slope of 25– 30 mV decade − 1 over the linear range of 10 − 5 –10 − 1 mol l − 1 at pH 5 – 9. Under a hydrodynamic mode of operation using a flow injection system manifold with a gas dialyzer, the sensor exhibits a linear response with a good reproducibility and high selectivity for sulfite ion. Interferences caused
by nitrite and sulfide ions are eliminated by using a carrier buffer solution containing sulfamic acid and sample pretreatment reaction with lead acetate.
References [1] K. Irgum, Anal. Chem. 57 (1985) 1496. [2] M.P. Keuken, F.P. Bakker, W.A. Lingerak, J. Slanina, Int. J. Environ. Anal. Chem. 31 (1987) 263. [3] W. Frenzel, B. Hillmann, Chem. Anal. (Warsaw) 40 (1995) 619. [4] J.S. Cosano, M.D. Luque-de-Castro, M. Valcarcel, Anal. Chim. Acta 302 (1995) 269. [5] T. Koh, K. Okabe, Analyst 119 (1994) 2457. [6] J.M. Bosque-Sendra, F. Molina, M. Nechar, J. AOAC Int. 77 (1994) 1224. [7] M.S. Abdel-Latif, Anal. Lett. 27 (1994) 2601. [8] J.F. Lawrence, R.K. Chadha, J. Chromatogr. 398 (1987) 355. [9] H.J. Kim, Y.K. Kim, J. Food Sci. 51 (1986) 1360. [10] R.G. Gerrits, J.A. Adeny, J. Chromatogr. 347 (1985) 419. [11] Y. Miura, T. Maruyama, T. Koh, Anal. Sci. 11 (1995) 617. [12] A.H. Liu, L. Xu, T. Li, S.J. Dong, E.K. Wang, J. Chromatogr. 699 (1995) 39. [13] I.I. Koukli, E.G. Sarantonis, A.C. Calokerinos, Analyst 113 (1988) 603. [14] D.A. Paullis, A. Townshend, Analyst 120 (1995) 467. [15] J. Hlavay, G.G. Guilbault, Anal. Chim. Acta 299 (1994) 91. [16] A.G. Fogg, M.A. Fernandez-Archiniega, R.M. Alonso, Analyst 110 (1985) 851. [17] H. Kaden, G. Schoenfeld, GIT-Fachz, Lab. 39 (1995) 301. [18] I.G. Casella, R. Marchese, Anal. Chim. Acta 311 (1995) 199.
782
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
[19] J.W. Ross, J.H. Riseman, J.A. Krueger, Pure Appl. Chem. 36 (1973) 473. [20] X. Xie, Z. Shakhsher, A.A. Suleiman, G.G. Guilbault, Z.M. Yang, Z. Sun, Talanta 41 (1994) 317. [21] M. Suzuki, S. Lee, K. Fujii, Y. Arikawa, I. Kubo, T. Kanagawa, E. Mikami, I. Karube, Anal. Lett. 25 (1992) 973. [22] Y. Kawamura, N. Kubo, H. Arata, Y. Ito, M. Tamura, K. Yamamoto, J. AOAC Int. 77 (1994) 1052. [23] V.J. Smith, Anal. Chem. 59 (1987) 2256. [24] Y. Yao, M. Wade, T. Wasa, Bunseki Kagaku 42 (1993) 397. [25] L. Campanella, P. Cipriani, T.M. Martini, M.P. Sammartino, M. Tomassetti, Anal. Chim. Acta 305 (1995) 32. [26] R.S. Hutchins, P. Molina, M. Alajarin, A. Vidal, L.G. Bachas, Anal. Chem. 66 (1994) 3188. [27] D.M. Paranitis, M.E. Meyerhoff, Anal. Chim. Acta 217 (1989) 123. [28] I.H.A. Badr, M.E. Meyerhoff, S.S.M. Hassan, Anal. Chim. Acta 310 (1995) 211.
.
[29] I. Ibrahim, Y. Cemal, B. Humeyra, Analyst 121 (1996) 1873. [30] W.Z. Gu, X.R. Wang, Fenxi-Huaxue 25 (1997) 308. [31] G. Alberti, M.G. Bernasconi, M. Casclola, U. Costantino, J. Inorg. Nucl. Chem. 42 (1980) 1637. [32] G.B. Martin, M.E. Meyerhoff, Anal. Chim. Acta 186 (1986) 71. [33] G. Alberti, P. Cardini-Galli, U. Costantino, E. Torracca, J. Inorg. Nucl. Chem. 29 (1967) 571. [34] IUPAC, Analytical Chemistry Division, Commision on Analytical Nomenclature, Pure Appl. Chem. 67 (1995) 507. [35] C.J. Coetzee, Ion-Selective Electrodes Rev. 3 (1981) 105. [36] S. Agrawal, M. Abe, Analyst 108 (1983) 712. [37] Sten, C. Gran, J. Inorg. Nucl. Chem. 33 (1971) 2229. [38] L. Maya, Radiochem. Acta 31 (1982) 947. [39] F. Hofmeister, Arch. Exp. Pathol. Pharmakol. 24 (1888) 247. [40] S.B. Park, W. Matuszewski, M. Meyerhoff, Y.H. Liu, K.M. Kadish, Electroanalysis 3 (1991) 909.
Flow injection analysis of sulfite ion with a potentiometric titanium phosphate –epoxy based membrane sensor Saad S.M. Hassan a,*, Sayed A. Marei b, I.H. Badr a, H.A. Arida b a
Department of Chemistry, Faculty of Science, Ain Shams Uni6ersity, Cairo, Egypt b Hot Laboratory Center, Atomic Energy Authority, Inshas, Egypt
Received 5 October 2000; received in revised form 4 January 2001; accepted 22 January 2001
Abstract A potentiometric sensor based on the use of titanium phosphate (TP) in epoxy matrix membrane is prepared and characterized. The sensor exhibits near-Nernstian response for many anionic species over the concentration range 10 − 1 –10 − 5 mol l − 1. The origin of response is explained on the basis of the conversion of titanium phosphate cation exchanger into hydrated titanium oxide anion exchanger by the effect of the high pH of the epoxy matrix. The sensitivity and selectivity of the sensor for sulfite ions are optimized by conversion of sulfite into gaseous SOx by acidification, and diffusion of the gas through a membrane-based gas dialyzer followed by potentiometric detection of sulfite ions formed within a flowing recipient stream. No interferences are caused by many common anions and acidic gas releasing species except sulfide and nitrite ions. Determination of sulfite ion at levels as low as 10 − 4 mol l − 1 or less in the presence of nitrite and sulfide ions is performed by using a modified carrier buffer stream (10 − 2 mol l − 1 MES, pH 5.0 containing sulfamic acid) and pretreatment with Pb2 + . Advantages offered by the proposed gas dialyzer/flow injection system with TP–epoxy membrane based sensor over traditional ion exchange based sensors includes long life time ( \ 8 months), excellent stability and reproducibility ( 1 mV), fast response time ( B 30 s), wide pH working range (pH 5–9), high sample throughput ( 60 samples h − 1), low detection limit (8 × 10 − 6 mol l − 1) and high thermal stability (up to 80°C). © 2001 Elsevier Science B.V. All rights reserved. Keywords: Inorganic ion exchangers; Sulfite sensor; Potentiometry; Flow injection analysis; Gas dialyzer; Titanium phosphate ion exchanger; Epoxy membrane
1. Introduction Methods in current use for the determination of sulfite ion include gravimetry [1], spectrophotome-
* Corresponding author. Tel./fax: + 20-2-6822991. E-mail address: [email protected] (S.S.M. Hassan).
try [2–7], chromatography [8–12], chemiluminescence [13 –15] and amperometry [16 –18]. Most of these methods, however, involve several manipulation steps, not used for turbid and colored solutions and require sophisticated instruments. Although potentiometric methods of analysis are simpler, attractive, much faster and less expensive, little is known about the use of potentiometric sensors for selective determination of sulfite ion.
0039-9140/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 0 1 ) 0 0 3 3 0 - 7
774
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
A potentiometric sensing membrane probe for sulfite ion consisting of a flat-bottom glass electrode located behind a porous polymeric membrane whereby a thin film of sulfite electrolyte solution is sandwiched between the electrode and the membrane has been described [19]. Sulfur dioxide released by acidification of sulfite solutions diffuses through the porous membrane, rapidly equilibrates and changes the pH of the electrolyte. Sulfite biosensors based on the use of immobilized yeast, bacteria or sulfite oxidase enzyme and an oxygen probe have been described for determination of sulfite in water, environmental materials, food and wine samples [20–25]. These gas sensors and biosensors suffer from significant interference by many gas releasing anions, low stability and long response time. Guanidinium ionophore based sulfite selective membrane sensor has been suggested [26]. A PVC membrane sensor based on the use of mercury(II) diethyldithiocarbamate complex has been described for potentiometric determination of sulfite in wine [27,28]. Metal wires coated with graphite– epoxy resin and metal mercury(II) methylpiperidine dithiocarbamate have also been suggested as sensors for sulfite ion [29]. A cluster compound film sensor prepared by impregnation of a copper wire into a solution of (NH4)2MoS4 followed by oxidation and coating with PVC membrane displayed linear response for sulfite [30]. Some of these sensors, however, suffer from a narrow response range, short life span, long response time and significant interferences by most common anions. The present work describes the preparation and characterization of a novel potentiometric sensor based on the use of titanium phosphate (TP) ion exchanger as an electroactive material and epoxy resin as a matrix membrane. The high chemical and thermal stability of TP ion exchanger [31] is expected to provide a suitable sensor with long life, high reproducibility and stability. Incorporation of the proposed sensor in a flow injection cell with gas dialyzer provides an efficient flow-injection detector for determination of sulfite with an enhancement selectivity. The sensor also offers the advantages of simplicity, sensitivity, reliability, automation feasibility and low cost.
2. Experimental
2.1. Reagents All chemicals used were of analytical reagent grade unless otherwise stated and the standard solutions and buffers were prepared with distilled, de-ionized water. 4-Morpholino-ethanesulfonic (MES) and tetrahydrofuran (THF) were obtained from Aldrich Chemicals Co. A commercially available metal free epoxy adhesive (Araldite super glue, Hollis, NY 11423) was obtained from Ciba-Geigy Co. The sulfite-sensing flow injection system utilized a 0.5 M H2SO4 as simple diluent, and a 0.01 M MES buffer of pH 5 (adjusted with 50% (w/v) NaOH solution) as a recipient carrier stream. Standard working sulfite solutions (10 − 1 –10 − 5 mol l − 1) were freshly prepared daily and stabilized by dilution of appropriate aliquots of the stock sodium sulfite solution with distilled de-ionized water containing 5% (v/v) glycerol to mask the heavy metals.
2.2. Equipment All potentiometric measurements were made at ambient temperature (259 1°C) using an Orion digital pH/mV meter (Model SA 720) and TP–epoxy membrane based sensor in conjunction with Hana silver-chloride single junction reference electrode filled with 10% (w/v) KCl. A combination Ross glass pH electrode (Orion 81-02) was used for all pH measurements. A schematic diagram of the flow injection analysis (FIA) system manifold used for sulfite quantitation is shown in Fig. 1. The system consists of a laboratory made flow cell containing the sulfite sensor, peristaltic pump (MSREGLO), and an Omnifit 4-port injection valve (Omnifit, Cambridge, UK). A home made modified gas dialyzer similar to that described earlier [32] was used. Each channel contains two support ribs to hold the gas-permeable membrane securely in place, in order to reduce the movement of the membrane. The dialyzer was fitted with a polytetrafluoroethylene membrane (0.2 mm pore size). The recipient buffer (10 − 2 mol l − 1 MES, pH 5.0), sample diluent 0.5 mol l − 1 H2SO4 solution and
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
the sample stream were supplied from separate reservoirs through 1.0, 0.4 and 0.4 mm i.d. Tygon tubings and propelled with the flow rates 4, 4 and 2 ml min − 1, respectively. A mixing coil of 6 cm length and 0.8 mm i.d. was used. The potentiometric signals were monitored with a flow through TP –epoxy based membrane sensor and an Orion SA 720 digital pH/mV meter and displayed on an (X-T) strip chart recorder (Linear 1200).
2.3. Titanium phosphate ion exchanger Amorphous TP was prepared as described earlier [33]. A 20.8 ml of TiCl4 (pure TiCl4 sp. g 1.73 g ml − 1) was diluted to 1 l with 4 mol l − 1 HCl. The resulting solution was slowly added with continuous stirring to 1 l of 12% (v/v) phosphoric acid in 4 mol l − 1 HCl. The precipitated TP gel was left to settle for 48 h before filtration and washing with 4 mol l − 1 HCl. The precipitate was then washed by double distilled deionized water till the acidity of the aqueous washing solution reached pH 3. The formed TP was then dried at 40°C and the final cake was disintegrated by soaking in hot double distilled water at 70°C. The granules produced was filtered and dried at 80°C
775
for 2 h. After sieving, the portion of 30 mesh size was used for subsequent work.
2.4. Titanium phosphate–epoxy membrane sensor The sulfite membrane sensor was prepared by thorough mixing of 0.6 g finely powered TP ion exchanger with 0.4 g of epoxy resin. The resulting paste was spread with a glass rod over a piece of filter paper to give a thin membrane and lift in air to cure. After drying and hardening of the membrane, a circular disc of 10 mm diameter and 1 mm thickness was cut, and fixed to the end of a PVC tube with epoxy resin. The tube was filled with a mixture of equal volume of 10 − 2 mol l − 1 KCl and 10 − 2 mol l − 1 Na2SO3. An Ag/AgCl internal reference electrode was immersed into the solution and the sensor was soaked overnight in a solution of 10 − 2 mol l − 1 Na2SO3. Sensor calibration was carried out by placing TP –epoxy membrane sensor in conjunction with a single junction Ag/AgCl reference electrode in 10 − 1 –10 − 6 mol l − 1 standard sulfite solutions. Calibration plot connecting the relation between logarithm sulfite concentration and sensor response in mV was made and used for subsequent measurements of unknown samples. Sensor selectivity was evaluated by measuring the responses of the sensor toward 10 − 5 –10 − 1 mol l − 1 sulfite solution and different interferent anion solutions (sodium or potassium salts). Calibration of the sensor was made in each tested anion solution under static conditions (batch measurements) and under a flow injection conditions in the presence and in absence of the gas diffusion cell.
2.5. Gas dialyzer/flow through sensor detector
Fig. 1. Schematic diagram of flow injection dialyzer manifold for determination of sulfite: (A) 10 − 2 M MES buffer pH 5; (B) 5× 10 − 1 M H2SO4; (C) sulfite test sample solution; (D) peristaltic pump; (E) pulse suppressor; (F) mixing coil; (G) flow injection valve; (H) dialysis chamber; (I) gas permeable membrane; (J) waste; (K) flow through TP –epoxy membrane sensor; (L) Ag/AgCl reference electrode; (M) electric ground attached to a metal inlet tube; (N) electrolyte solution; (O) pH/mV meter; and (P) chart recorder.
Flow through potentiometric cell equipped with TP –epoxy membrane sensor with an internal filling solution consisting of 2 ml of an equal mixture of 10 − 2 mol l − 1 Na2SO3 and 10 − 2 mol l − 1 NaCl and an Ag/AgCl internal reference electrode was used. The sensor was conditioned after preparation by soaking the membrane sensor overnight in 10 − 2 mol l − 1 Na2SO3. The flow cell is a flow through cap type (Fig. 2) with a gap thin enough between the membrane and sensing ele-
776
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
electrode were connected to an Orion 720 pH/mV meter and attached to a strip-chart recorder to monitor the FI signals. The waste from the beaker was continuously removed by the peristaltic pump.
2.6. FIA measurements of sulfite
Fig. 2. Schematic diagram of the flow-through cell: (A) Ag/ AgCl reference electrode; (B) 10 − 2 mol l − 1 Na2SO3 +10 − 2 mol l − 1 NaCl internal filling reference solution; (C): TP – epoxy membrane; (D): carrier solution; and (E): waste.
ment, not to retain any of the flowing solution after injection cycle. The whole cell acted as a potentiometric flow-through detector with adjusted zero dead volume. The cell was assembled and integrated with the flow injection system as shown in (Fig. 1). A gas dialysis unit similar to that described earlier [32] was used in the flow injection system. The upper channel of the dialyzer acted as an injection loop. The used valve has two positions; one to allow the carrier buffer stream (10 − 2 mol l − 1 MES, pH 5.0) to flow through the upper channel of the dialyzer and then downstream through the flow detector. The second position of the valve allowed separation of the upper channel of the dialyzer, and released the recipient stream to flow directly through the sensor detection unit. The sample diluent (0.5 mol l − 1 H2SO4) was thoroughly mixed with the sulfite sample before passing through the lower channel of the dialyzer using T-shaped glass connector and mixing coil. Both the recipient steam and the sample diluent stream were followed in an opposite direction to each other, to ensure a good trapping of SOx gases. The flow cell detection unit was placed in conjunction with a single junction Ag/AgCl reference electrode in a beaker filled with the carrier electrolyte solution. The sulfite sensor and reference
Sulfite test solutions were injected in the sample diluent stream containing 0.5 mol l − 1 sulfuric acid and then entered a mixing coil where SOx gases were generated. At the moment the sample plug began to enter the dialyzer, the rotary injection valve controlling the flow of the recipient carrier stream was turned to allow this stream bypass the upper channel of the dialyzer unit. The carrier buffer stream provided a good washing of the sensor detection unit, at the same time the acidified sample flow through the lower channel of the dialyzer before discharge to the waste. After exactly 1 min, the rotary valve was positioned so that the recipient stream plug contained in the upper channel of the dialyzer, was injected through the detector. At the same time the sample was changed into the next concentration. After 20 s, the new sample began to enter the dialyzer and the injection valve turned black. A time of 20 s was sufficient to obtain a complete signal on the recorder. The tubing distance between the output of the upper channel of the dialyzer and the input of the flow through TP detector was 3 cm, in order to minimize the dilution of the sample by the recipient stream. At least three signals for each sample were recorded and their average potential change was measured.
3. Results and discussion
3.1. Sensor characteristics and origin of response Epoxy membrane based sensor containing TP, showed strong potentiometric responses for many − − common anions such as SO23 − , NO− 2 , Br , I , − − − 2− 2− 2− − Cl , F , NO3 , HPO4 , SO4 , CrO4 , IO4 and S2 − . Evaluation of the sensor under static mode of operation according to IUPAC recommendations [34] indicated that the sensor displayed a sub
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
Nernstian response (slope 259 0.5 mV decade − 1) over the concentration range 10 − 5 – 10 − 1 mol l − 1 of sulfite and the detection limit was 6× 10 − 6 mol l − 1 Fig. 3. No potentiometric response was noticed by using a pure epoxy resin (blank) membrane based sensor for almost all tested cation and anions. Deviation from Nernstian behavior is known with many inorganic ion exchanger based sensors and attributed to incomplete permselectivity of the membrane [35]. The response of the TP – epoxy sensor in partially non-aqueous media was also tested by repeated calibration in different aqueous-organic media (e.g. ethanol, acetone and n-butanol) with different compositions. The results obtained revealed that the response was strongly affected by the nature and percentage of organic solvent. The calibration plots were generally linear over a narrow sulfite concentration range in the nonaqueous media. Best linear response was only obtained in 20% (v/v) ethanol solution. Table 1 summarizes the response characteristics of TP –epoxy sensor under static and hydrodynamic modes of operation from data collected from four different sensor assemblies for each
Fig. 3. Potentiometric static response of TP –epoxy membrane based sensor for sulfite ion in 10 − 2 mol l − 1 MES buffer of pH 5.
777
Table 1 Response characteristics of TP–epoxy matrix membrane sensor for sulfite ion under static (batch) and hydrodynamic gas dialyzer (FIA) modes of operations Parameter
Batch
FIA
Slope, mV decade−1 Linear range, mol l−1 Lower limit of dection, mol l−1 Response time for [SO2− 3 ] change from 10−4 to 10−1 mol l−1, s Recovery time, s Life time, month Working pH range
25.0 90.5 10−5–10−1 6×10−6
42 9 1.0 10−5–10−1 8×10−6
B20
30 \8 5–9
20
40 \8 5–9
system. It can be seen that the linear range of the sensor was improved under the FIA due to the pre concentration. To study the mechanism and origin of the anionic response of TP–epoxy membrane sensor, the effect of both ion exchanger and epoxy matrix on the potentiometric response of sulfite were examined. Although, TP is a polymeric cationic exchanger with no adsorption properties towards anions, its use in epoxy matrix displayed unexpected anionic response. Agrawal and Abe [36] attributed the anionic response of a sensor incorporating antimonic(V) acid cation exchanger in epoxy resin (Araldite) to the contribution of the polyamide functional group (N+) of Araldite. This explanation, however, failed to explain the cationic response of many epoxy based membrane sensors. Epoxy membranes with 5, 10, 40 and 60 wt.% TP were prepared and tested. A sensor with a membrane incorporating 100 wt.% epoxy resin has neither anionic nor cationic response. As the concentration of TP increased in the membrane, the initial potential, linear range and slope of the calibration plot were significantly changed Fig. 4. A sensor with Araldite membrane consisting of 10–60 wt.% TP–epoxy exhibited the best anionic response towards sulfite ion (calibration slope − 25 mV decade − 1). It was difficult to prepare membranes containing more than 60 wt.% TP with acceptable mechanical and adhesive proper-
778
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
ties. Sensor with pure compressed TP pellet as a membrane gave cationic response for most monovalent cations with a calibration slope of + 44 mV decade − 1. The possible conversion of TP cation exchanger into an anionic exchanger in the epoxy matrix was examined. Measurement of the pH of aqueous leaching solution of uncured epoxy resin was found to be higher than pH 12. This high basicity of the matrix membrane was sufficient to convert TP into hydrated titanium oxide gel (TiO2 · xH2O)n, which is known as an anion exchanger [37]. The gel takes up H+ followed by equivalent amount of exchangeable anions. For this reason hydrated zirconium and titanium oxide gels have been used as sorbent for anions in cooling water reactors [38]. A pure TP pellet based sensor soaked in 1 mol l − 1 NaOH solution rapidly lost its cationic response and displayed an anionic slope of − 12 mV decade − 1 upon soaking for 48 h. Soaking for a longer period of time caused swelling and disintegration of the membrane. Based on these findings it seems that TP cation exchanger is partly converted in the strong basic epoxy matrix into hydrated titanium oxide anion exchanger.
Fig. 4. Effect of membrane composition on the potentiometric response of TP–epoxy membrane for sulfite ion: (A) 100% epoxy resin; (B) 95% epoxy + 5% TP; (C) 90% epoxy +10% TP; (D) 60% epoxy + 40% TP; (E) 40% epoxy + 60% TP; and (F) 100% TP.
3.2. Response time and stability The response time of TP–epoxy membrane sensor was tested for 10 − 5 –10 − 1 mol l − 1 sulfite solutions. The time required for the sensor to reach 95% of the steady potential response, after successive addition of 10–1000 ml aliquots of 10 − 1 mol − 1 standard sulfite to 100 ml of the buffer to obtain 10-fold differences in concentration, was almost less than 20 s. Typical dynamic response of TP–epoxy sensor for sulfite ion is shown in Fig. 5. The sensor reached its equilibrium response in a relatively short time (B 20 s) at 10 − 5 –10 − 1 mol l − 1 sulfite. The sensor exhibited rapid recovery time (30 s) to reach 95% of the base line potential even after being exposed to 1.0 mol l − 1 sulfite. Similar response and recovery times were obtained with the flow through detection cell under flow injection conditions. The sample throughput was about 60 sample h − 1. The potential displayed by the proposed sulfite sensor for consecutive measurements of 10 − 5 – 10 − 1 mol l − 1 standard sulfite solutions did not vary by more than 9 2 mV over a period of about 6 months. The changes in the calibration slopes did not exceed 9 1.0 mV decade − 1 change of sulfite concentration. However, the initial potentials displayed by different sensors prepared from different membranes did not vary by more than 9 10 mV. Nevertheless, the calibration slope, detection limit, linear range of response and selectivity coefficients values were almost reproducible over a study period of \ 8 months when the sensor was preconditioned and stored when not in use in Na2SO3 solution and the internal filling solution was changed every 4 weeks. These are further advantages of the present sensor over many of the earlier suggested sulfite ion sensor. Reproducibility studies under flow injection conditions revealed that variation of the sensor potentials did not exceed 2 mV, the slope fluctuated within 1 mV concentration per decade. The high reproducibility and stability of the sensor is attributed to the high stability of TP ion exchanger, the good adhesive properties of epoxy matrix and the less leachibility of the ion exchanger compared with other earlier reported
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
779
Fig. 5. Potentiometric dynamic response and recovery time of TP – epoxy membrane sensor toward sulfite solutions.
sulfite sensors [26– 30]. In addition, the sensor exhibited high thermal stability (20– 80°C) and resistance towards strong oxidizing agents and mineral acids.
3.3. Effects of pH and foreign anions The effect of pH of 10 − 2 and 10 − 3 mol l − 1 sulfite solutions on the sensor potential response was examined by monitoring the variation of potential with the pH change over the range of 2–11. The sensor potentials did not vary by more than 2 mV over the pH range 5– 9. All subsequent measurements were performed at a pH 5. However, the potential of the sensor was significantly affected below pH 4 and above pH 9. The potentiometric responses of the TP–epoxy sensor towards 12 different anions were examined
under static (batch) and hydrodynamic (flow injection) modes of operation. The selectivity coefficient data obtained (Table 2) show that, TP – epoxy sensor is not really selective for sulfite ion since most of the common anions interfere. The order of selectivity pattern for the tested anions is similar in both batch and FI mode of operations. The selectivity sequence is, S2 − \IO− 4 − − − 2− 2− \I−\NO− 3 \NO2 \Br \Cl \CrO4 \SO3 \ SO24 − \ F− \ H2PO− A comparison was 4 . made with the Hofmeister sequence [39,40], − − − − − − ClO− 4 \IO4 \ SCN \ NO3 \ I \ Br \ NO2 − − − − \ HSO3 \ Cl \ HCO3 = H2PO4 . Most often observed with organic ion exchanger based membrane sensor, which is based on the relative lipophilicity of the anions and their partition coefficients into the organic membrane. TP–epoxy membrane exhibits slightly different response
780
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
characteristics toward sulfite compared with some other anions. Elimination of all interfering anions was achieved by incorporation of the TP–epoxy sensor in a flow injection system with a gas diffusion cell.
3.4. Flow injection analysis of sulfite The potentiometric calibration response of TP– epoxy membrane sensor in FI system without gas diffusion cell was examined. A series of standard sulfite fresh solutions and other interferent anions were injected in separate experiments through the injection valve into the carrier stream (10 − 2 mol l − 1 MES buffer of pH= 5.0) using a single line flow injection system. At least three signals were obtained by three different injections for each determination. The data obtained showed significant interference by most lipophilic anion Fig. 6. It has been suggested that lack of selectivity over certain anions can be overcome by using a gas diffusion cell to prevent certain interfering anions from reaching the surface of the membrane sensor [32]. Upon using a flow injection system with gas diffusion cell and repeating the above experiments, responses were only obtained with sulfite, nitrite and sulfide ions. These ions were easily converted under the same acidic conditions to their volatile gaseous counter parts and induced a strong response over a linear range of 10 − 5 –10 − 1 mol l − 1 Fig. 7. Using the gas diffusion cell, the
Fig. 6. Potentiometric calibration response of TP – epoxy based sensor toward various anions.
calibration slope of sulfite increased to 429 1 mV decade − 1 due to preconcentration step during 1 min sample trapping time before reaching the
Table 2 Potentiometric selectivity coefficients (K pot SO3,B) of TP–epoxy matrix membrane sensor Interferent (B)
K pot SO3,B
HPO2− 4 F− SO2− 4 CrO2− 4 Cl− − Br NO− 3 IO− 4 S2− NO− 2 I−
2.1×10−1 2.5×10−1 3.0×10−1 3.9 8.3 19.0 91.2 7.5×102 7.5×104 25 2.2×103
Fig. 7. Potentiometric response of TP – epoxy membrane sensor for various anions using gas diffusion FIA system.
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
781
Fig. 8. Typical strip-chart recording obtained by sulfite by TP– epoxy membrane sensor-gas dialyzer.
surface of the membrane sensor. Typical stripchart recording of the response obtained with standard sulfite solutions is shown in Fig. 8. However, interferences by nitrite and sulfide ions were completely eliminated by using a carrier buffer stream containing 10 − 1 mol l − 1 sulfamic acid in 10 − 2 mol l − 1 MES buffer of pH 5 after sample pre-treatment with 10 − 2 mol l − 1 lead acetate solution. Sulfamic acid completely eliminated the effect of nitrite ion, and the sulfide ion was precipitated as PbS. Based on this approach a selective determination of sulfite ion in the presence of almost all anions was obtained using the FIA/gas dialyzer manifold system. Determination of sulfite at a concentration level of 40 mg ml − 1 –8 mg ml − 1 using the flow injection system showed an average recovery of 98.5% with a mean standard deviation (S.D.) of 1.2%. No interference was noticed due to the presence of 104-fold excess of most common anions.
4. Conclusion TP –epoxy matrix membrane sensor exhibits a near-Nernstian response for many common anions including sulfite ion with a slope of 25– 30 mV decade − 1 over the linear range of 10 − 5 –10 − 1 mol l − 1 at pH 5 – 9. Under a hydrodynamic mode of operation using a flow injection system manifold with a gas dialyzer, the sensor exhibits a linear response with a good reproducibility and high selectivity for sulfite ion. Interferences caused
by nitrite and sulfide ions are eliminated by using a carrier buffer solution containing sulfamic acid and sample pretreatment reaction with lead acetate.
References [1] K. Irgum, Anal. Chem. 57 (1985) 1496. [2] M.P. Keuken, F.P. Bakker, W.A. Lingerak, J. Slanina, Int. J. Environ. Anal. Chem. 31 (1987) 263. [3] W. Frenzel, B. Hillmann, Chem. Anal. (Warsaw) 40 (1995) 619. [4] J.S. Cosano, M.D. Luque-de-Castro, M. Valcarcel, Anal. Chim. Acta 302 (1995) 269. [5] T. Koh, K. Okabe, Analyst 119 (1994) 2457. [6] J.M. Bosque-Sendra, F. Molina, M. Nechar, J. AOAC Int. 77 (1994) 1224. [7] M.S. Abdel-Latif, Anal. Lett. 27 (1994) 2601. [8] J.F. Lawrence, R.K. Chadha, J. Chromatogr. 398 (1987) 355. [9] H.J. Kim, Y.K. Kim, J. Food Sci. 51 (1986) 1360. [10] R.G. Gerrits, J.A. Adeny, J. Chromatogr. 347 (1985) 419. [11] Y. Miura, T. Maruyama, T. Koh, Anal. Sci. 11 (1995) 617. [12] A.H. Liu, L. Xu, T. Li, S.J. Dong, E.K. Wang, J. Chromatogr. 699 (1995) 39. [13] I.I. Koukli, E.G. Sarantonis, A.C. Calokerinos, Analyst 113 (1988) 603. [14] D.A. Paullis, A. Townshend, Analyst 120 (1995) 467. [15] J. Hlavay, G.G. Guilbault, Anal. Chim. Acta 299 (1994) 91. [16] A.G. Fogg, M.A. Fernandez-Archiniega, R.M. Alonso, Analyst 110 (1985) 851. [17] H. Kaden, G. Schoenfeld, GIT-Fachz, Lab. 39 (1995) 301. [18] I.G. Casella, R. Marchese, Anal. Chim. Acta 311 (1995) 199.
782
S.S.M. Hassan et al. / Talanta 54 (2001) 773–782
[19] J.W. Ross, J.H. Riseman, J.A. Krueger, Pure Appl. Chem. 36 (1973) 473. [20] X. Xie, Z. Shakhsher, A.A. Suleiman, G.G. Guilbault, Z.M. Yang, Z. Sun, Talanta 41 (1994) 317. [21] M. Suzuki, S. Lee, K. Fujii, Y. Arikawa, I. Kubo, T. Kanagawa, E. Mikami, I. Karube, Anal. Lett. 25 (1992) 973. [22] Y. Kawamura, N. Kubo, H. Arata, Y. Ito, M. Tamura, K. Yamamoto, J. AOAC Int. 77 (1994) 1052. [23] V.J. Smith, Anal. Chem. 59 (1987) 2256. [24] Y. Yao, M. Wade, T. Wasa, Bunseki Kagaku 42 (1993) 397. [25] L. Campanella, P. Cipriani, T.M. Martini, M.P. Sammartino, M. Tomassetti, Anal. Chim. Acta 305 (1995) 32. [26] R.S. Hutchins, P. Molina, M. Alajarin, A. Vidal, L.G. Bachas, Anal. Chem. 66 (1994) 3188. [27] D.M. Paranitis, M.E. Meyerhoff, Anal. Chim. Acta 217 (1989) 123. [28] I.H.A. Badr, M.E. Meyerhoff, S.S.M. Hassan, Anal. Chim. Acta 310 (1995) 211.
.
[29] I. Ibrahim, Y. Cemal, B. Humeyra, Analyst 121 (1996) 1873. [30] W.Z. Gu, X.R. Wang, Fenxi-Huaxue 25 (1997) 308. [31] G. Alberti, M.G. Bernasconi, M. Casclola, U. Costantino, J. Inorg. Nucl. Chem. 42 (1980) 1637. [32] G.B. Martin, M.E. Meyerhoff, Anal. Chim. Acta 186 (1986) 71. [33] G. Alberti, P. Cardini-Galli, U. Costantino, E. Torracca, J. Inorg. Nucl. Chem. 29 (1967) 571. [34] IUPAC, Analytical Chemistry Division, Commision on Analytical Nomenclature, Pure Appl. Chem. 67 (1995) 507. [35] C.J. Coetzee, Ion-Selective Electrodes Rev. 3 (1981) 105. [36] S. Agrawal, M. Abe, Analyst 108 (1983) 712. [37] Sten, C. Gran, J. Inorg. Nucl. Chem. 33 (1971) 2229. [38] L. Maya, Radiochem. Acta 31 (1982) 947. [39] F. Hofmeister, Arch. Exp. Pathol. Pharmakol. 24 (1888) 247. [40] S.B. Park, W. Matuszewski, M. Meyerhoff, Y.H. Liu, K.M. Kadish, Electroanalysis 3 (1991) 909.