Anal. Chem. 2003, 75, 141-144
All-Solid-State Calcium Solvent Polymeric Membrane Electrode for Low-Level Concentration Measurements Agata Michalska,† Anna Konopka, and Magdalena Maj-Zurawska*
University of Warsaw, Department of Chemistry, Pasteura 1, 02-093 Warsaw, Poland
An all-solid-state calcium-selective electrode with a plastic membrane phase containing a calcium ionophore ETH1001 placed on poly(3-methylthiophene) ion-to-electron transducting layer functionalized to bind calcium cations has been constructed. The obtained potentiometric sensors were characterized with a calibration line of slope close to Nernstian within the activity from 10-5 to 0.1 M. For Ca2+ activity lower than 10-5 M, super-Nernstian behavior was observed. The super-Nernstian response that is observed for electrodes with an internal solution strongly binding primary ions was in this case attributed to incorporation of calcium ions in the modified solid-contact phase. With this arrangement, the evaluation of selectivity coefficients much closer to those that depend on the properties of the ion-selective membrane itself was possible. The observed detection limit of the conventional solvent polymeric membranes potentiometric sensors is usually close to 10-6 M, both for the internal solution and so-called all-solid-state or coated wire arrangements. Recently, it was shown that the detection limit is not an intrinsic property of the ion-selective membrane, but rahter, it results from leakage of primary ions from the solvent polymeric membrane into the aqueous surface layer in the sample.1-5 For a traditional arrangement in which the polymericsusually PVC-basedsmembrane separates the sample and internal solution, the problem was solved upon recognition. The small flux of primary ion from the membrane to the sample has been successfully eliminated by choosing an inner solution characterized with buffered, constant, and low activity of the primary ion and a high activity of an interfering ion.1 Originally, inner solutions with ion buffers (e.g., EDTA) were used for this purpose, and the obtained sensors were characterized with the detection limit shifted down to the picomolar level. New possibilities of practical analytical measurements have arisen;6,7 * Corresponding author. Phone: +48 22 8200211. Fax: +48 22 8225996. E-mail:
[email protected]. † Co-corresponding author. Phone: +48 22 8200211. Fax: +48 22 8225996. E-mail:
[email protected]. (1) Sokalski, T.; Ceresa, A.; Zwickl, T.; Pretsch, E. J. Am. Chem. Soc. 1997, 119, 11347-11348. (2) Maj-Zurawska, M.; Erne, D.; Ammann, D.; Simon, W. Helv. Chim. Acta 1982, 65, 55-62. (3) Mathison, S.; Bakker, E. Anal. Chem. 1998, 70, 303-309. (4) Sokalski, T.; Zwickl, T.; Bakker, E.; Pretsch, E. Anal. Chem. 1999, 71, 2041209. (5) Sokalski, T.; Ceresa, A.; Fibbioli, M.; Zwickl, T.; Bakker, E.; Pretsch, E. Anal. Chem. 1999, 71, 1210-1214. 10.1021/ac025916y CCC: $25.00 Published on Web 11/06/2002
© 2003 American Chemical Society
studies of the system are also continued. Steady-state model calculations predicting the effect of a variety of factors were recently published.8 The effects of the concentration of the compexing ligand and resulting activity of the primary ion in the inner solution, the concentration of the ionophore in the membrane, the thickness of the diffusion layer at the outer interface of the electrode,9,10 and the influence of the type of plasticizer used in the ion-selective membrane11 were reported recently. The lowering of the detection limit of the potentiometric sensors has an effect on the selectivity coefficients determination, as well. Despite the earlier reports on the influence of main ion leakage from the membrane phase affecting the selectivity of the conventional ion-selective electrodes,12,13 until recently, the selectivity coefficients were evaluated under conditions of primary ion leakage from the membrane phase, that is, under the conditions that do not enable determination of the unbiased values representing the intrinsic properties of the membranes. The advantages and disadvantages of different methods of determining selectivity coefficients have recently been discussed.14 The conventional methods to determine selectivity coefficients (separate solution method (SSM), fixed interference method (FIM), and matched potential method (MPM)) are still the same as those originally proposed by IUPAC.15 However, several conditions have to be fulfilled to obtain meaningful data. Errors arise when the response to a weakly interfering ion is influenced by the primary ion leaching from the membrane. It should be stressed that lowering of the detection limit of conventional potentiometric sensors has made it possible to evaluate the unbiased selectivity coefficients of solvent polymeric membrane electrodes closer to the values that depend on the properties of the ion-selective membrane itself.4,5,9,11 The best evaluation of the selectivity of a sensor results from comparison of its characteristics obtained for the individual ions.14 The electrodes with lowered detection limits for indi(6) Ceresa, A.; Bakker, E.; Hattendorf, B.; Gu ¨ nther, D.; Pretsch, E. Anal. Chem. 2001, 73, 343-351. (7) Ion, A.; Bakker, E.; Pretsch, E. Anal. Chim. Acta 2001, 440, 71-79. (8) Zwickl, T.; Sokalski, T.; Pretsch, E. Electroanalysis 1999, 11, 673-680. (9) Sokalski, T.; Bedlechowicz, I.; Maj-Zurawska, M.; Hulanicki, A. Fresenius’ J. Anal. Chem. 2001, 370, 367-370. (10) Cereza, A.; Sokalski, T.; Pretsch, E. J. Electronal. Chem. 2001, 501, 7076. (11) Bedlechowicz, I.; Maj-Zurawska, M.; Sokalski, T.; Hulanicki, A. J. Electronal. Chem., in press. (12) Maj-Zurawska, M.; Sokalski, T.; Hulanicki, A. Talanta 1988, 35, 281-286. (13) Sokalski, T.; Maj-Zurawska, M.; Hulanicki, A. Mikrochim. Acta I 1991, 285291. (14) Bakker, E.; Pretsch, E.; Bu ¨ hlmann, P. Anal. Chem. 2000, 72, 1127-1133. (15) Umezawa, Y.; Umezawa, K.; Sato, H. Pure Appl. Chem. 1995, 67, 507-521.
Analytical Chemistry, Vol. 75, No. 1, January 1, 2003 141
vidual ions show better differentiated responses with nearNernstian slope,9 which allows use of SSM for comparison of different electrodes used under given specific conditions. All-solid-state ion-selective solvent polymeric membrane electrodes16 are of especial interest in the field of potentiometric sensors because they are free from some limitations resulting from the presence of the inner solution, for example, the need to work in a vertical position, evaporation of the inner solution, and maintenance requirement. For obvious reasons, the approach leading to prevention of the leakage of main ion from the plastic membrane phase was not readily available for this electrode arrangement. Conducting polymers (CP) poly(pyrrole),17-20 poly(thiophene),21 and poly(aniline)22,23 characterized both with ionic and electronic conductivity prove to be excellent materials for the solid-contact phase. In construction of this type, the solvent polymeric membrane is usually placed directly on the CP surface; thus, obtained sensors were characterized with a stability equivalent of the internal solution electrodes and comparable selectivity coefficients values.18-20 Recent studies have pointed to the possibility of inducing cation sensitivity of conducting polymer films by doping them with complexing agents.24 A similar concept has now been applied for stabilization of metal ions’ activity on a low level at the inner side of the solvent polymeric membranes. The properties of the calcium-selective electrode with a poly(3-methylthiophene) film doped with BF4- ions and modified with EDTA complexing agent solid-contact layer are reported here. EXPERIMENTAL SECTION Reagents. Tetrahydrofuran (THF) (Merck, Darmstadt, Germany) was freshly distilled. Ethylenediaminetetraacetic acid (EDTA), poly(vinyl chloride) (PVC), 2-nitrophenyl octyl ether (oNPOE), calcium ionophore (ETH 1001) [(-)-(R,R)-N,N′-bis-[11(ethoxycarbonyl)undecyl]-N,N′-4,5-tetramethyl-3,6-dioxaoctanediamide], and potassium tetrakis(4-chlorophenyl)borate (KTClPB) were from Fluka (Buchs, Switzerland). Doubly distilled and freshly deionized water (resistance 18.2 MΩcm, Milli-Q Plus, Millipore, Austria) was used for preparing all solutions. The stock solutions (0.1 M) of salt used were obtained by weighing the appropriate salts and dissolving them in water. Salts were of analytical grade (POCh, Gliwice, Poland). Ion-Selective Membranes. The composition of the calciumselective membrane used was as follows (m/m): 32.5% PVC, 0.7% KTClPB, 65.8% o-NPOE, 1.0% calcium ionophore ETH 1001; 200 mg of membrane components was dissolved in 2 mL of THF. Apparatus. In the potentiometric experiments, a multichannel data acquisition setup and software (Lawson Labs, Inc., 3217 Phoenixville Pike, Malvern, PA 19355) was used. In other electrochemical measurements, electrochemical analyzer EA9C (MTM, Krako´w, Poland) was used. Pump systems 700 Dosino and 711 Liquino (Metrohm, Herisau, Switzerland) were used to obtain sequential dilutions of the calibrating solutions. A double-junction Ag/AgCl reference electrode (Mo¨ller, Zu¨rich, Switzerland) with 0.1 M KCl solution in the outer sleeve was used in potentiometric measurements. Stable potential readings were taken (( 0.2 mV). Potentiostatic polarization was performed in the conventional cell. Pt counter electrode, silver/silver chloride gel electrode as 142
Analytical Chemistry, Vol. 75, No. 1, January 1, 2003
reference, and glassy carbon disks working electrodes were used. The single ion activities were calculated according to DebyeHu¨ckel theory.25 Ion-Selective Electrodes. The working electrodes consisted of two parts, the ordinary glassy carbon electrodes (diameter 3 mm) were equipped with a screwed cup with the opening located on the top of the working electrode surface; the diameter of the opening was 4.5 mm. This arrangement was used to prevent the ion-selective membrane from peeling from the solid-contact phase. The cup was added after the polymerization and modification of the conducting polymer film. Polymer films were deposited potentiostatically at +1.8 V, passing the charge 3.3 C/cm2. The monomer solution used was 0.1 M LiBF4 and 0.2 M 3-methylthiophene to give poly(3methylthiophene), PMT, films. Both components of the monomer solution were dissolved in propylene carbonate. Before polymerization, the monomer solution was deaerated with argon flow for 15 min, and deaeration was continued during polymerization. Then one of the two procedures of all-solid-state electrode preparation described below was followed. 1. The THF solution of the PVC-based ion-selective membrane was pipetted to the opening of protective cup screwed on the substrate electrode coated with conducting polymer film while it was placed in an upside-down position. After overnight evaporation of the membrane solvent, thus obtained so-called traditional allsolid-state electrodes were conditioned as described bellow. 2. The PMT films were modified by 2 days immersion in an aqueous solution containing 0.1 M EDTA and 0.05 M CaCl2, pH adjusted to 9 with 1 M KOH. The modified polymer films, after rinsing with deionized water and drying with filter paper, were covered with PVC-based calcium-selective membrane as described above. After overnight evaporation of the membrane solvent, thus obtained so-called all-solid-state electrodes with modified conducting polymer contact were conditioned as described bellow. The usual amount of membrane component solution applied to obtain the all-solid-state electrodes was 15 µL. In parallel, the conventional ion-selective electrodes were tested and the mentioned above membrane components solution in THF was used to prepare the membrane disk; the internal solution used was 0.01 M CaCl2. In this case, the membranes were fixed inside the Philips IS 561 electrode bodies. All obtained sensors tested were conditioned in solution containing both EDTA and Ca2+ ions. The solution contained 0.05 M EDTA solution and 0.025 M CaCl2 and was spiked with 1 M KOH to pH ) 9. RESULTS AND DISCUSSION The exemplary calibration lines obtained for three electrodes tested are presented in Figure 1. Both for conventional internal (16) Nikolskii, B. P.; Materowa, E. A. Ion. Sel. Electrode Rev. 1985, 7, 3-39. (17) Cadogan, A.; Gao, Z.; Lewenstam, A.; Ivaska, A.; Diamond, D. Anal. Chem. 1992, 64, 2496-2501. (18) Hulanicki, A.; Michalska, A. Electroanalysis 1995, 7, 692-693. (19) Michalska, A.; Hulanicki, A.; Lewenstam, A. Analyst 1994, 119, 2417-2420. (20) Michalska, A.; Hulanicki, A.; Lewenstam, A. Microchem. J. 1997, 57, 59-64. (21) Bobacka, J.; Lewenstam, A.; Ivaska A. Talanta 1993, 40, 1437-1441. (22) Lindfors, T.; Ivaska, A. Anal. Chim. Acta 1999, 400, 101-110. (23) Lindfors, T.; Ivaska, A. Anal. Chim. Acta 2000, 404, 111-119. (24) Migdalski, J.; Blaz, T.; Lewenstam, A. Anal. Chim. Acta 1996, 322, 141149. (25) Meier, P. C. Anal. Chim. Acta 1982, 136, 363-368.
Figure 1. The calibration lines obtained for tested electrodes recorded in CaCl2 solutions: (2) traditional all-solid-state arrangement, (9) traditional internal solution type, ([) EDTA-modified conducting polymer solid-state contact.
solution type and for traditional all-solid-state electrodes (i.e., unmodified conducting polymer was used to prepare the sensor), the typical responses were recorded. For the electrolyte activities within the range from 1 to 10-5 M, linear Nernstian responses, characterized with slope 25.0 ( 0.9 mV/dec (R2 ) 0.995) and 25.0 ( 0.9 mV/dec (R2 ) 0.996) mV/dec ISE for internal solution and traditional all-solid-state sensors were recorded, respectively. For a concentration of electrolyte lower than 10-6 M, the potential values recorded for both sensors were practically unaffected by electrolyte activity changes. This effect is well-documented in the literature and is attributed to the continuous release of ions from the membrane phase to the solution.1-5 The potentiometric sensor is insensitive to the electrolyte activity changes since the solution boundary layer is being enriched with primary ions originating from the membrane phase. For ion-selective electrodes, the internal solution phase leaking process has been successfully suppressed by lowering the primary ion activity in the internal solution by complex formation.1 However, it was observed that the composition of the internal solution influences both the achievable lower detection limit and the shape of the electrode response function. If the transport of the primary ion through the membrane phase from the sample toward the internal solution is induced, the super-Nernstian response is observed as a result of depletion of the primary ion in the diffusion layer at the membrane surface.1,4,5,8-11 We report here this effect observed for an all-solid-state solvent polymeric ion-selective electrode, Figure 1. The curve recorded for the electrode with surface modified conducting polymer contact is characterized with a typical super-Nernstian response over the electrolyte activities range from 10-5 to 10-8 M, although for activities higher than 10-5 M, the linear Nerstian dependence, characterized by slope 26.3 ( 0.8 mV/dec (R2 ) 0.997), was obtained. In our opinion, the super-Nernstian slope has to be attributed to the effective binding of calcium ions in the conducting polymer film modified with EDTA. Thus, for the electrolyte activity lower than 10-5 M, calcium cations are transported from the diffusion layer of the sample to the solid-contact phase.
The result presented proves that modified conducting polymer contact can be used to eliminate the analyte ion leakage from the ion-selective-based membrane to the sample phase and to lower the useful detection limit of all-solid-state potentiometric sensors. Although it was reported earlier that polypyrrole films can be effectively doped with EDTA,26 we were not able to repeat this procedure for PMT films deposited from a different solvent. The applied procedure of preparation of the solid-contact phase, that is, conditioning in a complexing-agent-containing solution, was applied here, since this method of incorporation of EDTA into the contact phase is extremely simple and yet effective. The slopes of potentiometric dependence recorded for unmodified PMT films recorded in calcium ion solutions was much lower than Nernstian, in the range of single mV/dec.21 This result was also confirmed in our studies. In contrast, the slope of the potentiometric dependence of EDTA-modified PMT films recorded in CaCl2 solution within the concentration range from 10-4 to 0.1 M was equal to 30.5 ( 0.4 mV/dec (R2 ) 0.99). Regardless of the fact that amount of ligand incorporated in the solid-contact phase by conditioning was not high enough to ensure the long term effect, it was sufficient to result in a super-Nernstian slope of the ionselective electrode characteristic that was obtained. It is well-documented in the literature that ion fluxes in the membrane phase strongly influence the selectivity pattern of the electrode.12,13 Thus, the super-Nernstian behavior of the all-solidstate sensor can be used to determine the selectivity coefficients of tested electrodes much closer to those which depend on the properties of the ion-selective membrane itself. We have obtained response curves for each individual ion, for all the solid-contact EDTA-modified electrodes tested. The inferfering ions exhibited near-Nernstian responses within the concentration range from 0.1 to 10-5 M. The calculated selectivity coefficients resulting from the SSM method for 0.01 M metal chloride solutions were used to compare the electrodes under the same interfering conditions. The results of selectivity coefficient determinations for all sensors (26) Arrigan, D. W. M.; Lowens, M. J. Electroanalysis 1999, 11, 647-652.
Analytical Chemistry, Vol. 75, No. 1, January 1, 2003
143
Figure 2. The selectivity coefficients obtained for the tested electrodes. The coefficients were determined with separate solution methods for the main and the interfering ion concentration equal to 0.01 M.
tested are presented in Figure 2. For the conventional internal solution electrode, the selectivity coefficients obtained in this study were close, within the range of experimental error, to the values reported earlier for similar membrane composition.9 The effect of contact type is easily seen when the above values are compared with the one obtained for traditional all-solid-state potentiometric sensor, Figure 2. Using a conventional internal solution electrode, it was practically impossible to discriminate between divalent (Ba2+ and Mg2+) ions tested. This effect results from a relatively high concentration of the main cation (Ca2+) in the sample solution layer close to the membrane of the sensor, masking the response for the interfering ions. Elimination of the internal solution in favor of a conducting polymer transducing layer resulted in selectivity coefficient values more dependent on the membrane properties. In this arrangement, the sole source of the main ion leaking to the solution is the membrane. It can be expected that a disturbing concentration of Ca2+ in the sample diffusion layer is lower, and log KCa,X values obtained are lower. A similar effect was reported earlier for other all-solid-state sensors with conducting polymer film contact layers.18-20 When the EDTA-modified poly(3-methylthiophene) solid contact was applied, the contamination of the solution with calcium cations leaking from the membrane was prevented, and the selectivity coefficients determined were expected to better represent the properties of the membrane used, Figure 2. The separation between individual monovalent and divalent ions tested is well-pronounced. The selectivity values obtained for Mg2+ and Ba2+ were similar to the ones reported for an electrode with a low activity of calcium ions in the internal solutions.9 The results reported here point out that a modified conducting polymer film solid contact could be used to evaluate selectivity resulting from properties of the polymeric membranes in the internal solution free arrangements. Since the complexing agent used to modified conducting polymer phase is present only at the surface layer of the poly(3methylthiophene) film, the effect of EDTA becomes weaker with time. After a week, the responses of the all-solid-state sensor tested 144
Analytical Chemistry, Vol. 75, No. 1, January 1, 2003
changed to reach those of a traditional (unmodified) conducting polymer-based solid-state electrode. This additionally proves that the EDTA capacity to complex a primary ion is restricted. Further studies in this field are in progress. CONCLUSIONS The results presented show that considerations concerning detection limit, and in particular, the super-Nernstian response function and selectivity pattern of the ion selective electrodes, with an internal solution are also valid for the all-solid-state ion-selective electrodes incorporating complexing ligand-modified conducting polymer film. To our knowledge, these effects have not been reported previously. The responses of the thus obtained sensor in terms of calibration line and selectivity coefficients were similar to the one obtained for an electrode with an internal solution of low activity of the primary ion.9 Because the effect of EDTA incorporated at the surface of the conducting polymer was only temporarysnot exceeding 10 dayssmore investigation with other conducting polymers and other methods of incorporation of the complexing agent are still needed. The results presented above indicate that the surface modification of conducting polymer film can be used as an alternative in those cases in which the direct doping with a complexing ligand is not applicable. Presented results point to a new possibility of constructing allsolid-state solvent polymeric membrane ion-selective electrodes for low-level concentration measurements. ACKNOWLEDGMENT The authors are grateful to Professor Adam Hulanicki for valuable discussion. The financial support from KBN, Grant no. 3T09A 11118 (M.M.Z.) and 7T09A01720 (A.M.) are kindly acknowledged. Received for review July 3, 2002. Accepted October 15, 2002. AC025916Y
