Page 1 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Thin Layer Ionophore-Based Membrane for Multianalyte Ion Activity Detection Gastón A. Crespo*, Maria Cuartero and Eric Bakker* Department of Inorganic and Analytical Chemistry, University of Geneva, Quai Ernest-Ansermet 30, CH1211 Geneva, Switzerland. Corresponding Authors: [email protected]; [email protected]
ACS Paragon Plus Environment
1
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 29
Abstract A concept is introduced that allows one to detect the activity of multiple ions simultaneously and selectively with a single ion-selective membrane. This is demonstrated with ca. 300 nm thin plasticized PVC membranes containing up to two ionophores in addition to a lipophilic cation-exchanger, overlaid on an electropolymerized poly-3-octylthiophene (POT) film as the electron to ion transducer. The ionselective membranes are formulated under ionophore depleted conditions (avoiding excess of ionophore over ion-exchanger), which is purposely different from common practice with ion-selective electrodes. Cyclic voltammetry is used to interrogate the films. An anodic scan partially oxidizes the POT underlayer, which results in the expulsion of cations from the membrane at an appropriate potential. During the scan of a membrane containing multiple ionophores, the least bound ion is expelled first, giving distinct Gaussian peak shaped ion transfer voltammetric waves that are analyzed in terms of their peak potential. These potentials are found to change with the logarithm of the ion activity, in complete analogy to ion-selective electrodes, and multiple such waves are observed with multiple ionophores that exhibit no obvious interference from the other ionophores present in the membrane. The concept is established with lithium and calcium ionophores and accompanied by a response model that assumes complete equilibration of the membrane at every applied potential. Based on the model, diffusion coefficients in the membrane or aqueous phase bear no influence on the peak potentials as long as thin layer behavior is observed, further confirming the analogy to a potentiometric experiment. Idealized ion transfer waves are narrower than experimental findings, which is explained by a broader than expected anodic peak for the oxidation of conducting polymer. The correspondence between experiment and theory is otherwise excellent in terms of thin layer behavior and Nernstian shift of the peaks with analyte concentration.
Keywords: thin layer membrane, multianalyte detection, conducting polymer, ion discrimination, ion transfer voltammetry.
ACS Paragon Plus Environment
2
Page 3 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Introduction Solid contact ion-to-electron transducers have been introduced to the field of potentiometric sensors to replace the inner filling liquid solution, which had been a root cause of a range of fundamental and operative drawbacks of polymeric ion-selective electrodes.1-4 Doped conducting polymers (CP) such as poly(3-octylthiophene) (POT) and poly(3,4-ethylenedioxythiophene) (PEDOT) were used to control the potential at the inner interface (metal/CP/membrane).5-7 The excellent stability of the resulting potentiometric signal (better than 1 mV h-1)1 was attributed to the large redox capacitance intrinsic to these polymers (∼1 mF)6, 8, although for solvent cast POT there is evidence that oxidized surface confined groups in combination with reduced bulk polymer provide a well defined redox buffer.9 For the latter case, only a small fraction of the neutral CP is in its oxidized form (CP+) and stabilized by the lipophilic counter ion in the membrane (c.a., tetrakis[3,5-bis(trifluoromethyl)phenyl]borate, TFPB).9-10 At zero current (potentiometry), the concentration of the lipophilic ion exchanger in the membrane phase is practically constant, and a redox buffer composed of a CP/CP+ couple is created in the vicinity of the inner interface and defines the potential.9 In this case, the phase boundary potential originating at the outer interface (membrane/sample) is in addition to the potential provided by the conducting polymer couple. The concept was systematically employed for preparing solid contact reference electrodes and indicator electrodes.11-12 Even more recently, a redox buffer system based on lipophilic Co(III)/Co(II) salts, conceptually similar to CP/CP+ behavior, resulted in highly reproducible potentials between different electrodes.13 The oxidation-reduction properties of CPs were also explored for the modulation of ion-transfer processes across a membrane/sample interface. Amemiya introduced CPs for the voltammetric ion-transfer with thin layer membranes triggered by the oxidation/reduction of either POT or PEDOT. A submicrometer thin membrane doped with a lipophilic electrolyte (ETH 500) and deposited on a thin POT layer14-15 was characterized in a rotating electrode configuration. The membrane was electrochemically enriched with perchlorate during a pre-concentration step, followed by cyclic voltammetric striping of the accumulated anion, giving impressively low detection limits. More recently, calcium was also determined at nanomolar levels by using PEDOT overlaid with a calcium ionophore containing membrane.16 Later, Si and Bakker used cation-exchangers to alter the extraction behavior of thin layer membranes with a POT underlayer.17 With this configuration both cation and anion transfer was observed in the same voltammogram. This suggests, in principle, that the same membrane may be subsequently used for cation and anion detection. Ionophores were not yet used in this early study.
ACS Paragon Plus Environment
3
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
More
recently,
other
electroactive
species
in
Page 4 of 29
polymeric
tetracyanoquinodimethane (TCNQ) and tetrathialfulvalene (TTF))
18-19
membranes
(i.e.,
7,7,8,8-
was also used to control the ion-
transfer process in a similar fashion as the conducting polymers discussed above.14, 17 Here, the authors explored a polarizable ionophore-based membrane composed of an ionophore, lipophilic electrolyte and a mixture of redox active sites.19 A peak shift was reported for Na+, K+ and Ca2+ electrodes in a range of 103.5
to 1M that appeared to obey the Nernst equation. Although it was claimed that a redox confinement is
limiting electrochemical turnover, the shape of the experimental voltammograms did not strongly suggest this.19 This might be explained by a heterogeneous distribution of the redox active sites caused by drop casting of the film instead of other strategies such as electropolymerization. We report here on the electrochemical modulation of ionophore-based thin membranes by controlling the oxidation state of electropolymerized POT. This ensures a more electrochemically accessible and homogenous layer of POT in comparison to other electroactive species reported elsewhere.19 It is confirmed that the incorporation of a limiting amount of either calcium or lithium ionophore shifts the ion transfer peak potential to drastically more positive potentials, as expected for a reversible complexation in the thin layer membrane. Because the potential shifts follow a Nernstian response with increasing concentrations of calcium and lithium, calibration curves analogous to potentiometric sensors are found. The key novelty of this work is the discovery that a membrane containing two ionophores can be used to perform selective multianalyte detection in a single voltammetric scan. A thin layer model developed in this paper explains the multianalyte detection ability and agrees very well with the experimental results.
ACS Paragon Plus Environment
4
Page 5 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Idealized Theory i) Thin layer membrane for the detection of a single monovalent cation The system is composed of a thin layer membrane containing a fixed concentration of cation exchanger (R-). The membrane is backside contacted with a thin film of a redox species (R/O). Two key assumptions are used: i) The membrane layer is sufficiently thin and the current is not limited by mass transport, and ii) the sample is sufficiently concentrated and the detected ion (M+) is therefore not depleted at the membrane surface. The potential for the charge transfer occurring at the boundary between the redox species film and membrane ( ECT ) is, for a reversible one electron transfer reaction, ideally described with the Nernst equation20-22 as follows: 0 ECT = ECT − s log
cR cO
(1)
with s the Nernstian slope (0.059 mV), cR and cO c the concentrations of the reduced and oxidized forms 0 of the redox species, and ECT is the standard reduction potential for this process.
In a similar fashion, the phase boundary potential for the ion transfer at the membrane/solution interface is also ideally written in terms of the Nernst equation (eq. 2) with the uncomplexed primary analyte concentration in the membrane ( cMorg ) and in the aqueous phase ( cMaq ), where EIT0 is the standard iontransfer potential. EIT = EIT0 − s log
cMorg cMaq
(2)
In the absence of ionophore, the mass balance equations for both redox species and cation exchanger RT are:
cR/O = cR + cO
(3)
RT = cO + cMorg
(4)
where cR / O and RT are the total concentrations of the indicated species. The working electrode potential, which corresponds to the applied potential ( Eapp ), is written as the sum of these two phase boundaries (charge and ion transfer potentials):
c c org 0 Eapp = ECT + EIT0 − s log R + log Maq cO cM
ACS Paragon Plus Environment
(5)
5
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 6 of 29
Inserting the mass balances into equation (5) gives:
c −c R −c 0 Eapp = ECT + EIT0 − s log R / O O + log T aq O cO cM
(6)
As the applied potential is time dependent, it can be expressed as a function of scan rate ( υ ) and the initial potential ( Ei ) as follows: Eapp (t ) = Ei + υ t
(7)
Considering that the sensing film is sufficiently thin, the observed current is proportional to its oxidation rate, area of the electrode ( A ) and the thickness of the film ( δ film ) as follows23:
i = −nFA δ film
∂cO (t) ∂t
(8)
Of course, this current is the same for each interface and for the total cell. Note that mass transport is not limiting the current. The film is considered at electrochemical equilibrium at each potential value. The concentration of cO is finally obtained by solving equation 6 (not shown). For the backward scan, the applied potential is written as: Eapp (t ) = Eturn + υ t
(9)
And the procedure is otherwise in complete analogy to the forward scan. ii) Idealized two ionophore based membrane An ideal system based on two ionophores (L and L’) and two monovalent cations (I+ and J+) is also characterized. Mass balances for the cation exchanger and the two ionophores are expressed in equations (10-12): org RT = cO + cILorg + cJL '
(10)
LT = cLorg + cILorg
(11)
org L 'T = cLorg' + cJL '
(12) '
org
where LT and LT are the total concentrations of each ionophore in the membrane, cL org
org
and cL '
the
org
concentrations of the indicated uncomplexed ionophore, cIL and cJL ' the concentrations of complexed
ACS Paragon Plus Environment
6
Page 7 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
aq
aq
cations and, cI and cJ are the concentrations of cations in the aqueous sample. In analogy to above, the applied potential is a combination of the potential at the two interfaces, written for each ion as:
c −c c org 0 Eapp = ECT + EI0 − s log R / O O + log orgIL aq cO cL cI
(13)
c −c c org 0 Eapp = ECT + EJ0 − s log R / O O + log orgJL ' aq cO cL ' c J
(14)
The formation constants
org org ' −1 β IL = cILorg ⋅ (cIorg LT )−1 and β JL ' = cJL for IL and JL’ are included in ' ⋅ (cJ LT )
EI0 and EJ0 (equations 13 and 14). By solving the system of equations 10 to 14 for cR / O , an analytical solution is obtained as a function of Eapp . Since the complexity of the analytical solution increases considerably when allowing for the concurrent existence of divalent and monovalent ions, the treatment was restricted to monovalent ions only.
ACS Paragon Plus Environment
7
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 8 of 29
Experimental Section Reagents, Materials and Equipment. Aqueous solutions were prepared by dissolving the appropriate chloride salts (supporting information) (Sigma Aldrich) in deionized water (>18 MΩ cm). Au-electrode-tip (6.1204.320) and GC-electrode-tip (6.1204.300) with an electrode diameter of 3.00±0.05 mm were sourced from Metrohm (Switzerland). Cyclic voltammograms were recorded with a PGSTAT 302N (Metrohm Autolab B.V., Utrecht, The Netherlands) controlled by Nova 1.8 software (supplied by Autolab) running on a PC. A double-junction Ag/ AgCl/3M KCl/1 M LiOAc reference electrode (6.0726.100 model, Metrohm, Switzerland) and a platinum electrode (6.0331.010 model, Metrohm, Switzerland) were used in the three-electrode cell. A rotating disk electrode (Autolab RDE, Methrom Autolab B.V., Utrecht, The Netherlands) was used for 17, 24
spin coating the films on the electrodes. Electrochemical polymerization of poly(3-octylthiophene)
is
described in the supporting information (Figure S1). Table 1 shows the compositions of membrane cocktails prepared in 1 mL of THF (see supporting information for thin membrane preparation Figure S2S4).
ACS Paragon Plus Environment
8
Page 9 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Results and Discussion The goal of this paper is to demonstrate that a single membrane can be made responsive to multiple analytes, using the peak potential as potentiometric readout. Ion extraction is triggered by the oxidation or re-reduction of a POT layer overlaid with a thin ionophore-based ion-selective membrane. Preliminary experiments with membranes that do not contain ionophore (MI and MII, see Table 1) were performed, allowing us to rationalize the electrochemical processes that occur at both interfaces POT/membrane and membrane/sample when a sweep potential is applied. To produce reproducible and sufficiently thin films (232±20 nm as measured by ellipsometry; see Supporting Information and Figure S5), a careful optimization of the membrane composition and viscosity was performed (see Supporting Information). Cyclic voltammetry at a moderately high scan rate (100 mV s-1) between -0.5 V and 1.4 V was performed in a 10 mM NaCl background electrolyte. The same cation (i.e., Na+) was shared in the membrane and sample for this experiment. An idealized response mechanism for MI is illustrated in Figure 1. As POT is oxidized to POT+, it starts to form ion pairs with the cation exchanger (R-) present in the membrane, resulting in the expulsion of sodium from the membrane (Figure 1a). During the reverse cathodic sweep, sodium is again extracted from the solution to balance the cation exchanger (R-) that is liberated by the reduction of POT+ to electrically neutral POT (Figure 1b). During the forward scan, two distinct peaks are observed for MI (Figure S6) and one for MII (Figure 2a), suggesting that the oxidation of POT is indeed coupled to ion-transfer processes at the membrane/sample interface (note a corresponding pattern for the backward scan). The peak 333 mV is ascribed to the transfer of sodium while the one at 1.2 V to that of chloride. This is confirmed by adding tetrabutylammonium nitrate (TBA+NO3-) to the solution containing 10 mM NaCl. As TBA+ is a more lipophilic ion than sodium and hence is most easily transferred to the membrane, the peak at 333 mV gradually disappears, making way for a new peak at a more positive potential of 780.1 mV (Figure S7). Similarly, the anion wave exhibits a potential shift of 30 mV that is a function of the lipophilicity difference between chloride and nitrate. The Gaussian shape of the voltammetric peak (Figure 2a) and the linear relationship between peak current and scan rate (Figure S8) suggests thin layer behavior, meaning that mass transport is not the rate limiting step. This condition is important for the potentiometric multianalyte methodology introduced here. Indeed, this behavior is only achieved with thin layer membranes, otherwise a larger peak separation and peak currents that change linearly with the square root of the scan rate were observed. The integration of the two peaks, see Figure 2a, are 3.51 µC for the forward and 3.50 µC for the backward scan, exhibiting 0.1% variation. This value indicates the charge transferred across the membrane/sample interface and
ACS Paragon Plus Environment
9
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 10 of 29
suggests an excellent chemical reversibility of the system. This ion-transfer charge corresponds to ca. 27% of the total charge deposited during the electropolymerization of POT (12.70±0.15 µC, Figure S1). This suggests that the charge capacity is limited by the ion-exchanger concentration, not by the ability of POT to oxidize. An incomplete oxidation of POT during the ion transfer experiments is also useful to keep the potential at the POT–membrane interface from changing significantly. Membrane MII was used as initial membrane composition to completely suppress the anion transfer wave. With increasing concentration of NaCl (10 to 100 mM) (Figure S9), the ion-transfer peak was found to shift to higher potentials, in accordance with the Nernst expression described in eq. (2). Calculated thin layer cyclic voltammograms (see theory for single ionophore membranes, supporting information) are shown as a function of the sample concentration (Figure 2b) and scan rate (Figure S8). As mass transport within the thin layer membrane and in the sample solution is assumed not to be rate limiting, the theoretical response curves are independent of diffusion coefficients. The qualitative agreement between theory and experiment suggests that the model assumptions are appropriate for the peak shapes as well as thin layer behavior of the membrane. In addition, the voltammetric peak potentials move as a function of ion activity changes according to the Nernst equation. Under these conditions, the methodology may be adequately described as a potentiometric measurement, albeit not at zero current. We note that the calculated peaks are significantly narrower than the experimental ones (ca. 100 mV for calculated and 220 for experimental peaks, Figures 2a and 2b). In principle, any deviation from idealized behavior must affect the shape of the total peak. To confirm this behavior, both the wave for the idealized oxidation process and for the ion transfer were individually calculated (see Supporting Information). As observed in Figure S10, the potential windows at which both processes occur are completely different. While the calculated oxidation peak spans 300 mV, the width for the ion transfer wave is only 5 mV. This means that the overall peak width is mainly determined by the POT oxidation wave, which can be equally approximated by a Gaussian function (Figure S10a). Ideally, therefore, the peak width is rather independent of the ion transfer processes, and peak widths are expected to be similar to that of a reversible surface confined process, which is 90.6 mV/n. A Gaussian function of varying width for the POT oxidation may be combined with the reversible ion transfer wave (Figure S10b) to obtain different peak widths (Figure S10c). The experimental anodic peak obtained for the detection of sodium ions perfectly agrees when a broader Gaussian wave for POT oxidation is considered (Figure S10c). Without having to introduce a potentially inadequate theoretical model that describes the experimentally observed behavior, this treatment nonetheless suggests that the broader than expected peak width originates from non-ideal POT electrochemistry. On other hand, the Nernstian shift of the ion transfer wave for different ion activities of the primary ion will result in a Nernstian shift of the
ACS Paragon Plus Environment
10
Page 11 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
total peak, because the two individual potential contributions are additives (Figure S11). This behavior is therefore independent of the specific peak shape of the underlying electrochemistry of the conducting polymer. As expected for MII, a Hofmeister selectivity pattern is observed (Figure 3a) for sodium, lithium and calcium. The selectivity coefficients were calculated by the separate solution method25, a common pot pot protocol for potentiometric sensors, as log K Na.Li = −0.27 and log K Na.Ca = −1.9 . The integrated charge for
the individual peaks confirms that a similar and reproducible amount of charge is transferred for each separately measured ion (2.719 µC for Na+, 2.702 µC for Li+ and 2.669 µC for Ca2+). Further experiments were carried out with membranes that contained either calcium or lithium ionophore (M3, M4, M5, M6, see Table 1). The individual voltammograms obtained for 10 mM calcium and 10 mM lithium chloride in comparison with 10 mM NaCl are shown in Figure 3b and 3c respectively. As a consequence of the incorporated ionophore, a remarkable shift in the ion transfer waves is observed (337 mV and 141 mV displacement for Ca2+ and Li+, respectively). The calculated selectivity coefficients 26 pot pot correspond to the reported ones for the same ionophores ( log K Li,Na = −12.8 ). = −2.4 and log KCa.Na
These membranes were further explored in a mixture of 10 mM NaCl and variable levels of CaCl2 (from 10-7 to 10-5 M). Cyclic voltammetry shows two anodic ion-transfer peaks. While the first wave is assigned to the expulsion of sodium from the membrane, the second one corresponds to calcium (Figure S12). The ionophore promotes the complexation of calcium (L3Ca2+) and provokes the mentioned potential shift. In the backward scan, the process is reversed and calcium and sodium are extracted back into the polymeric film. If only calcium is present in the solution or at higher concentration levels than 5 x 10-5 M, just a single wave is obtained in the forward and backward scans as shown in Figure S12. An excellent reproducibility is demonstrated for fifty consecutive scans, resulting in a current variation of just 0.03 µA at 965.9 mV. Incremental sample concentration changes of either lithium or calcium were characterized with three identical electrodes measured at the same time. After each addition, a cyclic voltammogram was recorded at 100 mV s-1, taking about 20-s per scan. The peak potential was found to gradually move to more positive values by obeying the Nernst equation (Figure 4a, c). Note that this was not observed in the absence of ionophore because of the highly interfering background electrolyte. The analytical responses of a range of membranes (MII-MVII) with different plasticizers are listed in Table S1. The use of o-NPOE slightly improves the limit of detection (Figure 4b, d). The limit of detection for these thin membranes is at least one order higher than for traditional ion-selective membranes (~200 µm)27, and no attempt was made here to optimize the detection limit in this early work.
ACS Paragon Plus Environment
11
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 12 of 29
We emphasize, however, that such thin layer membranes can be conveniently operated in stripping ion transfer voltammetry to achieve very low detection limits, as demonstrated by the group of Amemiya.14,15 The integrated charge and position of the forward and backward peaks was found to diminish slightly when the electrodes were washed with pure water (see Figure S13). To avoid leaching of the membrane components, electrodes were therefore rinsed with NaCl solution or wiped off with a paper tissue. Alternatively, a more rigid membrane (higher PVC amount, membrane VII) did not show evidence of leaching, but a deterioration of the limit of detection, slope and thin layer behavior was found (Table S1). The shapes of the forward and backward waves for calcium ionophore containing membrane are significantly different to the NaCl wave (Figure 4c). The depletion of calcium in the membrane shows a wider peak with a transfer potential of 966 mV with respect to the backward wave (830 mV). This effect was not observed with membranes containing lithium ionophore (Figure 4a). This might be explained by the higher 1:3 stoichiometry between calcium and ionophore, perhaps giving rise to a broadening owing to step-wise calcium complexation.28 Motivated by our recent work on thin layer voltammetry for halide discrimination29,30 and the results demonstrated above with solid contact ion-to-electron transducers (POT), we explored whether selective multianalyte detection would be possible with a membrane containing two ionophores. Clearly, this would not be possible by zero current potentiometry since it is essentially a one-dimensional technique.31 Figure 5a shows the behavior of Membrane IX, containing a lithium and calcium ionophore, when interrogated by cyclic voltammetry. The first peak corresponds to lithium transfer and the second one to calcium, as expected from the previous experiments with membranes containing a single ionophore. In principle, this membrane behaves similarly to that of two independent membranes (see Figure 3b and 3c). Incremental increases of lithium and calcium concentration again result in a Nernst shift to more positive potentials. This suggests that both ionophores can work in tandem with the underlying POT electrochemistry to resolve the two ions in solution. Calibration curves (Epeak vs. log activity) show Nernstian response slopes (59.1±0.1 and 29.0±0.1 mV for Li+ and Ca2+ respectively) with a limit of detection of (5.0±0.1) x 10-5 and (2.1±0.1) x 10-5 M for Li+ and Ca2+ respectively, see Figure 5b. A methodical adjustment of the cation exchanger composition was explored, keeping an invariable concentration of the two ionophores (MVIII, MIX and MX). The intuitive mechanism is put forward in Figure 6. A membrane (MVIII) containing half the molar amount of cation exchanger with respect to ionophore (0.5:1:1 molar ratio) only exhibited calcium ion-transfer (Figure S14a). As the formation constant for Ca2+ is roughly 1020 times greater than for Li+,26 the cation exchanger will preferentially stabilize the charge of [CaL3]2+ complex. Note that ion-selective membranes used in potentiometry must be formulated to exhibit a molar excess of uncomplexed ionophore to give a selective and Nernstian
ACS Paragon Plus Environment
12
Page 13 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
response to the ion of interest. A two-ionophore membrane would therefore invariably respond only to calcium, in analogy to the voltammetric results with Membrane MVIII. Increasing the amount of cation exchanger to a 1:1:1 molar ratio (MIX) resulted in an adequate discrimination between lithium and calcium (Figure S14b). In this case, the cation exchanger forces, at complete oxidation of POT, the conversion of the totality of calcium ionophore to the complex [CaL3]2+ and partially the lithium ionophore to [LiL’]+. On the other hand, if the cation exchanger concentration just exceeded the sum of both ionophore concentrations, a third peak associated to the extraction of an uncomplexed ion appeared in the voltammogram, deteriorating peak separation (Figure S14c). While the ideal membrane should exactly match the concentration of cation exchanger and both ionophores, MIX exhibited the sought-after characteristics and was used further. It is noted that a slight improvement of the detection limit was observed for calcium when using a gold substrate instead of GC (Table S4, Figure S15). The idealized theoretical description put forward above suggests that this methodology may be capable of potentiometric multianalyte detection. Figure 7 and Figure 8 show the influence of the standard iontransfer potential and the sample concentration on the voltammetric readout of a membrane containing two ionophores. As shown in Figure 7, a standard potential difference of 0.4 V results in complete resolution of both peaks. Smaller potential differences result in lower resolution (see Figure S16). Thin layer behavior is theoretically confirmed with the linear relationship between peak current and scan rate (Figure S17). Most importantly, theory confirms that each peak potential change obeys the Nernst equation for its respective ion when keeping the concentration of the other ion constant (Figure 8, Figure S18). This result can be rationalized as follows. The membrane is interrogated under ionophore deficient conditions. At the beginning of an anodic voltammetric scan, all ionophores in the membrane are in their complexed form, with the lipophilic ion-exchangers forming their counter ions. As POT is oxidized, the weakest bound complex is dissociated first, expelling lithium ions from the membrane. As the calcium ionophore is, at this stage, exhaustively bound to calcium, there is no interference from calcium for this process and the shift of the lithium extraction wave obeys the Nernst equation. At potentials more anodic than the lithium transfer wave, uncomplexed lithium ionophore is present in the membrane, along with calcium complex. Sufficient lithium-calcium peak separation is needed to ensure that this free ionophore does not influence the calcium peak position. It may be estimated that less than 1% error is observed for a peak separation of twice the Nernstian slope, but future work will evaluate this important aspect in more detail. If uncomplexed lithium ionophore is unable to compete with the calcium transfer wave, as in the
ACS Paragon Plus Environment
13
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 14 of 29
experiments observed here, again a selective and Nernstian potential response for calcium is expected, independent of the lithium concentration in the sample.
Conclusions Performing dynamic electrochemistry with solid contact ion-to-electron transducers such as POT may open new horizons for producing reversible, multianalyte selective, easy to miniaturize and disposable sensors in comparison to traditional ion-selective electrodes or other sensors based on ion-transfer voltammetry between two immiscible liquids.32 While the underlying electrochemistry results in broader peaks than expected for a reversible one electron redox couple, conducting polymers are an attractive materials platform for the achievement of electrochemically addressable thin films that also give opportunities for simultaneous optical detection. The key novelty put forward in this paper is the ability for thin layer voltammetric ion transfer membranes to detect multiple analytes selectively by potentiometric readout (change of peak potentials). This was achieved with membranes deficient of ionophore, which was controlled by the concentration of lipophilic ion-exchanger. Two well-resolved ion-transfer waves for lithium and calcium were observed that each gave Nernstian response behavior, independent of the concentration of the other ion. A simple theory adequately explains the results without the need to include diffusion coefficients of the mobile species.
ACS Paragon Plus Environment
14
Page 15 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Acknowledgments The authors thank the Swiss National Science Foundation and the European Union (FP7-GA 614002SCHeMA project) for supporting this research.
Associated Content Supporting Information. POT and membrane preparation, optimization of the membrane composition, calculated voltammograms. This material is available free of charge via the Internet at http://pubs.acs.org.
Author Information Corresponding Authors [email protected]; [email protected]
ACS Paragon Plus Environment
15
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 16 of 29
References (1)
Bobacka, J., Ivaska, A., Lewenstam, A., Chem. Rev. 2008, 108, 329-351.
(2)
Lindner, E., Gyurcsanyi, R. E., J. Solid State Electrochem. 2009, 13, 51-68.
(3)
Duezguen, A., Zelada-Guillen, G. A., Crespo, G. A., Macho, S., Riu, J., Xavier Rius, F., Anal. Bioanal. Chem. 2011, 399, 171-181.
(4)
Michalska, A., Electroanalysis 2012, 24, 1253-1265.
(5)
Bobacka, J., McCarrick, M., Lewenstam, A., Ivaska, A., Analyst 1994, 119, 1985-1991.
(6)
Bobacka, J., Anal. Chem. 1999, 71, 4932-4937.
(7)
Vazquez, M., Danielsson, P., Bobacka, J., Lewenstam, A., Ivaska, A., Sens. Actuators B Chem.
2004, 97, 182-189.
(8)
Bobacka, J., Ivaska, A., Grzeszczuk, M., Synthetic Metals 1991, 44, 21-34.
(9)
Veder, J.-P., De Marco, R., Patel, K., Si, P., Grygolowicz-Pawlak, E., James, M., Alam, M. T., Sohail, M., Lee, J., Pretsch, E., Bakkert, E., Anal. Chem. 2013, 85, 10495-10502.
(10)
Bobacka, J., Grzeszczuk, M., Ivaska, A., J. Electroanal. Chem. 1997, 427, 63-69.
(11)
Mattinen, U., Bobacka, J., Lewenstam, A., Electroanalysis 2009, 21, 1955-1960.
(12)
Cicmil, D., Anastasova, S., Kavanagh, A., Diamond, D., Mattinen, U., Bobacka, J., Lewenstam, A., Radu, A., Electroanalysis 2011, 23, 1881-1890.
(13)
Zou, X. U., Cheong, J. H., Taitt, B. J., Buehlmann, P., Anal. Chem. 2013, 85, 9350-9355.
(14)
Kim, Y., Amemiya, S., Anal. Chem. 2008, 80, 6056-6065.
(15)
Kim, Y., Rodgers, P. J., Ishimatsu, R., Amemiya, S., Anal. Chem. 2009, 81, 7262-7270.
ACS Paragon Plus Environment
16
Page 17 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
(16)
Kabagambe, B., Garada, M. B., Ishimatsu, R., Amemiya, S., Anal. Chem. 2014, 86, 7939-7946.
(17)
Si, P., Bakker, E., Chem. Comm 2009, 5260-5262.
(18)
Zhang, J., Harris, A. R., Cattrall, R. W., Bond, A. M., Anal. Chem. 2010, 82, 1624-1633.
(19)
Harris, A. R., Zhang, J., Cattrall, R. W., Bond, A. M., Anal. Methods 2013, 5, 3840-3852.
(20)
Feldman, B. J., Burgmayer, P., Murray, R. W., J. Am. Chem. Soc. 1985, 107, 872-878.
(21)
Michalska, A., Maksymiuk, K., Talanta 2004, 63, 109-117.
(22)
Bard, A. J., Faulkner, L. R., Electrochemical Methods. Fundamentals and Applications, John Wiley & Sons, Inc., USA, 2001, page 590-593,
(23)
Osteryou.Ra, Christie, J. H., Anal. Chem. 1974, 46, 351-355.
(24)
Skompska, M., Siwiec, D., Kudelski, A., Zagorska, M., Synthetic Metals 1999, 101, 35-36.
(25)
Umezawa, Y., Buhlmann, P., Umezawa, K., Tohda, K., Amemiya, S., Pure Appl. Chem. 2000, 72, 1851-2082.
(26)
Qin, Y., Mi, Y. M., Bakker, E., Anal. Chim. Acta 2000, 421, 207-220.
(27)
Bakker, E., Buhlmann, P., Pretsch, E., Chem. Rev. 1997, 97, 3083-3132.
(28)
Zhu, J., Qin, Y., Zhang, Y., Anal. Chem. 2010, 82, 436-440.
(29)
Cuartero, M., Crespo, G. A., Afshar, M. G., Bakker, E., Anal. Chem. 2014, 86, 11387-11395.
(30)
Cuartero, M., Crespo, G. A., Bakker, E., Anal. Chem. 2015, 87, 1981-1990.
(31)
Han, S. H., Lee, K. S., Cha, G. S., Liu, D., Trojanowicz, M., J. Chromatogr. 1993, 648, 283-288.
(32)
Zhou, M., Gan, S., Zhong, L., Su, B., Niu, L., Anal. Chem. 2010, 82, 7857-7860.
ACS Paragon Plus Environment
17
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
(33)
Page 18 of 29
Wilke, S., J. Electroanal. Chem. 2001, 504, 184-194.
ACS Paragon Plus Environment
18
Page 19 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Figures Captions Figure 1.
Proposed response mechanism of a cation exchanger (R-) based thin layer membrane backside contacted with POT during the a) forward scan and b) backward scan.
Figure 2.
Cyclic voltammograms for the cation exchanger based membrane MII (a) in contact with 10 mM NaCl at a scan rate 100 mV.s-1. (b) Calculated cyclic voltammograms at three concentration levels of NaCl (scan rate 100 mV.s-1); inset: Epeak vs. logaNa+. Parameters used R /O in the calculation: A=7.0686×10-4 dm2; δ film =2.3×10-6 dm; cR / O = 5×10-1 M; RT = 10×10-3 0 = 0.8V; EIT0 = -0.3 V (calculated from the standard Gibbs M; Einit= 0V; Eturn = 0.8V; ECT energy for sodium partition between water | nitrobenzene microinterface)33; s = 0.059 V;
cMaq =0.01, 0.0316 and 0.1 M. Figure 3.
(a) Observed cyclic voltammograms in contact with 10 mM NaCl, LiCl and CaCl2 for membrane MII. (b) Cyclic voltammograms for either 10 mM NaCl or 10 mM CaCl2 using MIII (calcium ionophore). (c) Cyclic voltammograms for either 10 mM NaCl or LiCl using MIV (lithium ionophore). Scan rate: 100 mV s-1.
Figure 4.
Membranes containing a single ionophore. (a) Observed cyclic voltammograms at increasing concentrations of lithium using membrane MIV (lithium ionophore). (b) Calibration curve for lithium using two different plasticizers. (c) Observed cyclic voltammograms at increasing concentrations of calcium using membrane MIII (calcium ionophore). (d) Calibration curve for calcium using two different plasticizers. Cation concentrations: 1×10-6, 2.5×10-6, 5×10-6, 1×10-5, 2.5×10-5, 5×10-5, 1×10-4, 2.5×10-4, 5×10-4, 1×10-3, 2.5×10-3, 5×10-3 and 1×10-2 M. Conditions: 100 mV s-1, background of 10 mM NaCl. Error bars for b) and d) are shown but are smaller than the plot markers.
Figure 5.
Membrane containing two ionophores. (a) Cyclic voltammograms obtained for different concentrations of lithium and calcium at equimolar concentrations (2.5×10-6, 5×10-6,1×10-5, 2.5×10-5, 5×10-5, 1×10-4, 2.5×10-4, 5×10-4, 1×10-3, 2.5×10-3, 5×10-3 and 1×10-2 M) using membrane MIX (b) Calibration curve obtained for calcium and lithium. Error bars for b) are shown but are smaller than the plot markers.
Figure 6.
Response mechanism for membranes containing two ionophores (L and L’) backside contacted with POT during a) the forward scan and b) the backward scan. I+ and J+ are the two analytes and R- is the cation-exchanger.
Figure 7.
Calculated voltammograms for two ionophore-based membranes. Parameters: ∆Eº = 0.4V; R /O =2.3×10-6 dm; cR / O = 5×10-1 M; RT = 10×10-3 M; υ = A=7.0686×10-4 dm-2; F = 96486; δ film aq
aq
0 100 mVs-1; Einit= -0.3 V; Eturn = 1.4V; ECT = 0.8V; s = 0.059 V; cI =0.01 M; c J = 0.01M.
Figure 8.
Calculated voltammograms for two ionophore-based membrane at different concentrations of the species J, keeping the concentration of I constant. Parameters: A=7.0686×10-4 dm-2; F R /O = 96486; δ film =2.3×10-6 dm; cR / O = 5×10-1 M; RT = 10×10-3 M; υ = 100 mVs-1; Einit= -0.3 aq aq 0 V; Eturn = 1.4V; ECT = 0.8V; ∆E I0, J = 0.4V; s = 0.059 V; cI =0.01 M; cJ = 10-2, 3×10-3,
1×10-3, 3×10-4, 10-4 M.
ACS Paragon Plus Environment
19
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 20 of 29
Table 1. Compositions of membrane cocktails prepared in 1 mL of THF. Membrane
a b
Components PVCa
DOSa
I
31
II
NPOEa
NaTFPBb
TDDAClb
Ca Ionophoreb
Li Ionophoreb
63
40
40
33
63
41
III
29
61
40
IV
30
63
42
V
29
60
40
VI
30
62
41
82
VII
47
48
41
82
VIII
29
59
42
80
80
IX
29
57
80
81
80
X
29
55
139
82
80
80 82 80
weight percentage mmol kg-1
ACS Paragon Plus Environment
20
Page 21 of 29
Analytical Chemistry Figure 1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
(+)
(-)
e-
a)
POT+
POT
RMn+
SOLUTION
(+)
(-)
e-
b) POT
POT+ R-
SOLUTION
M+n
ACS Paragon Plus Environment
Analytical Chemistry Figure 2
4
a) CATION RELEASE
3
i/m A
2 1 0 -1 -2 -3
CATION EXTRACTION
-0.6 -0.4 -0.2
0
0.2
0.4
0.6
0.8
1
1.2
1.4
E/V 2
b)
410 E / mV
1.5
390 370
1 i/m A
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
350
0.5
-2.0
-1.5 -1.0 log aNaCl
0 -0.5 -1 -1.5 INCREASING NaCl -2 -0.6 -0.4 -0.2
0
0.2
0.4
0.6
0.8
1
1.2
1.4
E/V
ACS Paragon Plus Environment
Page 22 of 29
Page 23 of 29
Analytical Chemistry Figure 3
3
a) NO IONOPHORE
i/m A
2
1
0 CaCl2
-1
LiCl NaCl -2
-0.5
-0.2
0.1
0.4 E/V
0.7
1
1.3
b) CALCIUM IONOPHORE CaCl2
1.5
i/m A
NaCl 0.5
-0.5
-1.5
-2.5 -0.5
2
-0.2
0.1
0.4 0.7 E/V
1
1.3
c) LITHIUM IONOPHORE
1.5 NaCl 1 i/m A
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
LiCl
0.5 0 -0.5 -1 -1.5 -0.5
-0.2
0.1
0.4 E/V
0.7
1
ACS Paragon Plus Environment
1.3
1.6
Analytical Chemistry
Page 24 of 29
Figure 4
2
a) LITHIUM IONOPHORE
560 b)
510 Epeak / mV
i/m A
1
0
460
DOS o-NPOE
410 -1 360 INCREASING LITHIUM -2 0.4 -0.5 -0.2 0.1 0.7 E/V
310 1
1.3
1.5
-6
-5
-4 log a Li +
-3
-2
1060 d)
c) CALCIUM IONOPHORE 1030
1
1000
Epeak / mV
i/m A
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
0
970
o-NPOE
-1
DOS
940 -1.5
910 INCREASING CALCIUM
-0.5
-0.2
0.1
0.4
0.7
1
1.3
880
-6
E/V
ACS Paragon Plus Environment
-5
-4 log a Ca2+
-3
-2
Page 25 of 29
Analytical Chemistry Figure 5
1.5
i/m A
a) BOTH IONOPHORES
Ca2+
Li+
0.5
-0.5 Ca2+
Li+ -1.5 -0.5
Epeak / mV
950
-0.2
0.1
0.4 E/V
0.7
1
1.3
b)
900 850 log a Ca2+
530 510 Epeak / mV
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
490 470 450 430 410 -6
-5
-4 log a Li +
-3
ACS Paragon Plus Environment
-2
Analytical Chemistry Figure 6 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
(+)
(-)
e-
a)
ePOT+
POT
I++ L
RLI+
POT+
POT
RL’J+
J++ L’
SOLUTION
(+)
(-)
e-
b)
ePOT+
POT
POT+
POT
R-
RL’
L I+
J+
SOLUTION
ACS Paragon Plus Environment
Page 26 of 29
Page 27 of 29
Analytical Chemistry Figure 7
1.6
D E0 = 0.4 V
1.2 i/m A
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
0.8
0.4
0 0
0.4
0.8
1.2
E/V
ACS Paragon Plus Environment
Analytical Chemistry
Page 28 of 29
Figure 8
1.6 FIXED CI
INCREASING CJ
1.2 i/m A
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
0.8
0.4
0 0
0.4
0.8
1.2
E/V
ACS Paragon Plus Environment
Page 29 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
For TOC Only
ACS Paragon Plus Environment
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Thin Layer Ionophore-Based Membrane for Multianalyte Ion Activity Detection Gastón A. Crespo*, Maria Cuartero and Eric Bakker* Department of Inorganic and Analytical Chemistry, University of Geneva, Quai Ernest-Ansermet 30, CH1211 Geneva, Switzerland. Corresponding Authors: [email protected]; [email protected]
ACS Paragon Plus Environment
1
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 29
Abstract A concept is introduced that allows one to detect the activity of multiple ions simultaneously and selectively with a single ion-selective membrane. This is demonstrated with ca. 300 nm thin plasticized PVC membranes containing up to two ionophores in addition to a lipophilic cation-exchanger, overlaid on an electropolymerized poly-3-octylthiophene (POT) film as the electron to ion transducer. The ionselective membranes are formulated under ionophore depleted conditions (avoiding excess of ionophore over ion-exchanger), which is purposely different from common practice with ion-selective electrodes. Cyclic voltammetry is used to interrogate the films. An anodic scan partially oxidizes the POT underlayer, which results in the expulsion of cations from the membrane at an appropriate potential. During the scan of a membrane containing multiple ionophores, the least bound ion is expelled first, giving distinct Gaussian peak shaped ion transfer voltammetric waves that are analyzed in terms of their peak potential. These potentials are found to change with the logarithm of the ion activity, in complete analogy to ion-selective electrodes, and multiple such waves are observed with multiple ionophores that exhibit no obvious interference from the other ionophores present in the membrane. The concept is established with lithium and calcium ionophores and accompanied by a response model that assumes complete equilibration of the membrane at every applied potential. Based on the model, diffusion coefficients in the membrane or aqueous phase bear no influence on the peak potentials as long as thin layer behavior is observed, further confirming the analogy to a potentiometric experiment. Idealized ion transfer waves are narrower than experimental findings, which is explained by a broader than expected anodic peak for the oxidation of conducting polymer. The correspondence between experiment and theory is otherwise excellent in terms of thin layer behavior and Nernstian shift of the peaks with analyte concentration.
Keywords: thin layer membrane, multianalyte detection, conducting polymer, ion discrimination, ion transfer voltammetry.
ACS Paragon Plus Environment
2
Page 3 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Introduction Solid contact ion-to-electron transducers have been introduced to the field of potentiometric sensors to replace the inner filling liquid solution, which had been a root cause of a range of fundamental and operative drawbacks of polymeric ion-selective electrodes.1-4 Doped conducting polymers (CP) such as poly(3-octylthiophene) (POT) and poly(3,4-ethylenedioxythiophene) (PEDOT) were used to control the potential at the inner interface (metal/CP/membrane).5-7 The excellent stability of the resulting potentiometric signal (better than 1 mV h-1)1 was attributed to the large redox capacitance intrinsic to these polymers (∼1 mF)6, 8, although for solvent cast POT there is evidence that oxidized surface confined groups in combination with reduced bulk polymer provide a well defined redox buffer.9 For the latter case, only a small fraction of the neutral CP is in its oxidized form (CP+) and stabilized by the lipophilic counter ion in the membrane (c.a., tetrakis[3,5-bis(trifluoromethyl)phenyl]borate, TFPB).9-10 At zero current (potentiometry), the concentration of the lipophilic ion exchanger in the membrane phase is practically constant, and a redox buffer composed of a CP/CP+ couple is created in the vicinity of the inner interface and defines the potential.9 In this case, the phase boundary potential originating at the outer interface (membrane/sample) is in addition to the potential provided by the conducting polymer couple. The concept was systematically employed for preparing solid contact reference electrodes and indicator electrodes.11-12 Even more recently, a redox buffer system based on lipophilic Co(III)/Co(II) salts, conceptually similar to CP/CP+ behavior, resulted in highly reproducible potentials between different electrodes.13 The oxidation-reduction properties of CPs were also explored for the modulation of ion-transfer processes across a membrane/sample interface. Amemiya introduced CPs for the voltammetric ion-transfer with thin layer membranes triggered by the oxidation/reduction of either POT or PEDOT. A submicrometer thin membrane doped with a lipophilic electrolyte (ETH 500) and deposited on a thin POT layer14-15 was characterized in a rotating electrode configuration. The membrane was electrochemically enriched with perchlorate during a pre-concentration step, followed by cyclic voltammetric striping of the accumulated anion, giving impressively low detection limits. More recently, calcium was also determined at nanomolar levels by using PEDOT overlaid with a calcium ionophore containing membrane.16 Later, Si and Bakker used cation-exchangers to alter the extraction behavior of thin layer membranes with a POT underlayer.17 With this configuration both cation and anion transfer was observed in the same voltammogram. This suggests, in principle, that the same membrane may be subsequently used for cation and anion detection. Ionophores were not yet used in this early study.
ACS Paragon Plus Environment
3
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
More
recently,
other
electroactive
species
in
Page 4 of 29
polymeric
tetracyanoquinodimethane (TCNQ) and tetrathialfulvalene (TTF))
18-19
membranes
(i.e.,
7,7,8,8-
was also used to control the ion-
transfer process in a similar fashion as the conducting polymers discussed above.14, 17 Here, the authors explored a polarizable ionophore-based membrane composed of an ionophore, lipophilic electrolyte and a mixture of redox active sites.19 A peak shift was reported for Na+, K+ and Ca2+ electrodes in a range of 103.5
to 1M that appeared to obey the Nernst equation. Although it was claimed that a redox confinement is
limiting electrochemical turnover, the shape of the experimental voltammograms did not strongly suggest this.19 This might be explained by a heterogeneous distribution of the redox active sites caused by drop casting of the film instead of other strategies such as electropolymerization. We report here on the electrochemical modulation of ionophore-based thin membranes by controlling the oxidation state of electropolymerized POT. This ensures a more electrochemically accessible and homogenous layer of POT in comparison to other electroactive species reported elsewhere.19 It is confirmed that the incorporation of a limiting amount of either calcium or lithium ionophore shifts the ion transfer peak potential to drastically more positive potentials, as expected for a reversible complexation in the thin layer membrane. Because the potential shifts follow a Nernstian response with increasing concentrations of calcium and lithium, calibration curves analogous to potentiometric sensors are found. The key novelty of this work is the discovery that a membrane containing two ionophores can be used to perform selective multianalyte detection in a single voltammetric scan. A thin layer model developed in this paper explains the multianalyte detection ability and agrees very well with the experimental results.
ACS Paragon Plus Environment
4
Page 5 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Idealized Theory i) Thin layer membrane for the detection of a single monovalent cation The system is composed of a thin layer membrane containing a fixed concentration of cation exchanger (R-). The membrane is backside contacted with a thin film of a redox species (R/O). Two key assumptions are used: i) The membrane layer is sufficiently thin and the current is not limited by mass transport, and ii) the sample is sufficiently concentrated and the detected ion (M+) is therefore not depleted at the membrane surface. The potential for the charge transfer occurring at the boundary between the redox species film and membrane ( ECT ) is, for a reversible one electron transfer reaction, ideally described with the Nernst equation20-22 as follows: 0 ECT = ECT − s log
cR cO
(1)
with s the Nernstian slope (0.059 mV), cR and cO c the concentrations of the reduced and oxidized forms 0 of the redox species, and ECT is the standard reduction potential for this process.
In a similar fashion, the phase boundary potential for the ion transfer at the membrane/solution interface is also ideally written in terms of the Nernst equation (eq. 2) with the uncomplexed primary analyte concentration in the membrane ( cMorg ) and in the aqueous phase ( cMaq ), where EIT0 is the standard iontransfer potential. EIT = EIT0 − s log
cMorg cMaq
(2)
In the absence of ionophore, the mass balance equations for both redox species and cation exchanger RT are:
cR/O = cR + cO
(3)
RT = cO + cMorg
(4)
where cR / O and RT are the total concentrations of the indicated species. The working electrode potential, which corresponds to the applied potential ( Eapp ), is written as the sum of these two phase boundaries (charge and ion transfer potentials):
c c org 0 Eapp = ECT + EIT0 − s log R + log Maq cO cM
ACS Paragon Plus Environment
(5)
5
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 6 of 29
Inserting the mass balances into equation (5) gives:
c −c R −c 0 Eapp = ECT + EIT0 − s log R / O O + log T aq O cO cM
(6)
As the applied potential is time dependent, it can be expressed as a function of scan rate ( υ ) and the initial potential ( Ei ) as follows: Eapp (t ) = Ei + υ t
(7)
Considering that the sensing film is sufficiently thin, the observed current is proportional to its oxidation rate, area of the electrode ( A ) and the thickness of the film ( δ film ) as follows23:
i = −nFA δ film
∂cO (t) ∂t
(8)
Of course, this current is the same for each interface and for the total cell. Note that mass transport is not limiting the current. The film is considered at electrochemical equilibrium at each potential value. The concentration of cO is finally obtained by solving equation 6 (not shown). For the backward scan, the applied potential is written as: Eapp (t ) = Eturn + υ t
(9)
And the procedure is otherwise in complete analogy to the forward scan. ii) Idealized two ionophore based membrane An ideal system based on two ionophores (L and L’) and two monovalent cations (I+ and J+) is also characterized. Mass balances for the cation exchanger and the two ionophores are expressed in equations (10-12): org RT = cO + cILorg + cJL '
(10)
LT = cLorg + cILorg
(11)
org L 'T = cLorg' + cJL '
(12) '
org
where LT and LT are the total concentrations of each ionophore in the membrane, cL org
org
and cL '
the
org
concentrations of the indicated uncomplexed ionophore, cIL and cJL ' the concentrations of complexed
ACS Paragon Plus Environment
6
Page 7 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
aq
aq
cations and, cI and cJ are the concentrations of cations in the aqueous sample. In analogy to above, the applied potential is a combination of the potential at the two interfaces, written for each ion as:
c −c c org 0 Eapp = ECT + EI0 − s log R / O O + log orgIL aq cO cL cI
(13)
c −c c org 0 Eapp = ECT + EJ0 − s log R / O O + log orgJL ' aq cO cL ' c J
(14)
The formation constants
org org ' −1 β IL = cILorg ⋅ (cIorg LT )−1 and β JL ' = cJL for IL and JL’ are included in ' ⋅ (cJ LT )
EI0 and EJ0 (equations 13 and 14). By solving the system of equations 10 to 14 for cR / O , an analytical solution is obtained as a function of Eapp . Since the complexity of the analytical solution increases considerably when allowing for the concurrent existence of divalent and monovalent ions, the treatment was restricted to monovalent ions only.
ACS Paragon Plus Environment
7
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 8 of 29
Experimental Section Reagents, Materials and Equipment. Aqueous solutions were prepared by dissolving the appropriate chloride salts (supporting information) (Sigma Aldrich) in deionized water (>18 MΩ cm). Au-electrode-tip (6.1204.320) and GC-electrode-tip (6.1204.300) with an electrode diameter of 3.00±0.05 mm were sourced from Metrohm (Switzerland). Cyclic voltammograms were recorded with a PGSTAT 302N (Metrohm Autolab B.V., Utrecht, The Netherlands) controlled by Nova 1.8 software (supplied by Autolab) running on a PC. A double-junction Ag/ AgCl/3M KCl/1 M LiOAc reference electrode (6.0726.100 model, Metrohm, Switzerland) and a platinum electrode (6.0331.010 model, Metrohm, Switzerland) were used in the three-electrode cell. A rotating disk electrode (Autolab RDE, Methrom Autolab B.V., Utrecht, The Netherlands) was used for 17, 24
spin coating the films on the electrodes. Electrochemical polymerization of poly(3-octylthiophene)
is
described in the supporting information (Figure S1). Table 1 shows the compositions of membrane cocktails prepared in 1 mL of THF (see supporting information for thin membrane preparation Figure S2S4).
ACS Paragon Plus Environment
8
Page 9 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Results and Discussion The goal of this paper is to demonstrate that a single membrane can be made responsive to multiple analytes, using the peak potential as potentiometric readout. Ion extraction is triggered by the oxidation or re-reduction of a POT layer overlaid with a thin ionophore-based ion-selective membrane. Preliminary experiments with membranes that do not contain ionophore (MI and MII, see Table 1) were performed, allowing us to rationalize the electrochemical processes that occur at both interfaces POT/membrane and membrane/sample when a sweep potential is applied. To produce reproducible and sufficiently thin films (232±20 nm as measured by ellipsometry; see Supporting Information and Figure S5), a careful optimization of the membrane composition and viscosity was performed (see Supporting Information). Cyclic voltammetry at a moderately high scan rate (100 mV s-1) between -0.5 V and 1.4 V was performed in a 10 mM NaCl background electrolyte. The same cation (i.e., Na+) was shared in the membrane and sample for this experiment. An idealized response mechanism for MI is illustrated in Figure 1. As POT is oxidized to POT+, it starts to form ion pairs with the cation exchanger (R-) present in the membrane, resulting in the expulsion of sodium from the membrane (Figure 1a). During the reverse cathodic sweep, sodium is again extracted from the solution to balance the cation exchanger (R-) that is liberated by the reduction of POT+ to electrically neutral POT (Figure 1b). During the forward scan, two distinct peaks are observed for MI (Figure S6) and one for MII (Figure 2a), suggesting that the oxidation of POT is indeed coupled to ion-transfer processes at the membrane/sample interface (note a corresponding pattern for the backward scan). The peak 333 mV is ascribed to the transfer of sodium while the one at 1.2 V to that of chloride. This is confirmed by adding tetrabutylammonium nitrate (TBA+NO3-) to the solution containing 10 mM NaCl. As TBA+ is a more lipophilic ion than sodium and hence is most easily transferred to the membrane, the peak at 333 mV gradually disappears, making way for a new peak at a more positive potential of 780.1 mV (Figure S7). Similarly, the anion wave exhibits a potential shift of 30 mV that is a function of the lipophilicity difference between chloride and nitrate. The Gaussian shape of the voltammetric peak (Figure 2a) and the linear relationship between peak current and scan rate (Figure S8) suggests thin layer behavior, meaning that mass transport is not the rate limiting step. This condition is important for the potentiometric multianalyte methodology introduced here. Indeed, this behavior is only achieved with thin layer membranes, otherwise a larger peak separation and peak currents that change linearly with the square root of the scan rate were observed. The integration of the two peaks, see Figure 2a, are 3.51 µC for the forward and 3.50 µC for the backward scan, exhibiting 0.1% variation. This value indicates the charge transferred across the membrane/sample interface and
ACS Paragon Plus Environment
9
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 10 of 29
suggests an excellent chemical reversibility of the system. This ion-transfer charge corresponds to ca. 27% of the total charge deposited during the electropolymerization of POT (12.70±0.15 µC, Figure S1). This suggests that the charge capacity is limited by the ion-exchanger concentration, not by the ability of POT to oxidize. An incomplete oxidation of POT during the ion transfer experiments is also useful to keep the potential at the POT–membrane interface from changing significantly. Membrane MII was used as initial membrane composition to completely suppress the anion transfer wave. With increasing concentration of NaCl (10 to 100 mM) (Figure S9), the ion-transfer peak was found to shift to higher potentials, in accordance with the Nernst expression described in eq. (2). Calculated thin layer cyclic voltammograms (see theory for single ionophore membranes, supporting information) are shown as a function of the sample concentration (Figure 2b) and scan rate (Figure S8). As mass transport within the thin layer membrane and in the sample solution is assumed not to be rate limiting, the theoretical response curves are independent of diffusion coefficients. The qualitative agreement between theory and experiment suggests that the model assumptions are appropriate for the peak shapes as well as thin layer behavior of the membrane. In addition, the voltammetric peak potentials move as a function of ion activity changes according to the Nernst equation. Under these conditions, the methodology may be adequately described as a potentiometric measurement, albeit not at zero current. We note that the calculated peaks are significantly narrower than the experimental ones (ca. 100 mV for calculated and 220 for experimental peaks, Figures 2a and 2b). In principle, any deviation from idealized behavior must affect the shape of the total peak. To confirm this behavior, both the wave for the idealized oxidation process and for the ion transfer were individually calculated (see Supporting Information). As observed in Figure S10, the potential windows at which both processes occur are completely different. While the calculated oxidation peak spans 300 mV, the width for the ion transfer wave is only 5 mV. This means that the overall peak width is mainly determined by the POT oxidation wave, which can be equally approximated by a Gaussian function (Figure S10a). Ideally, therefore, the peak width is rather independent of the ion transfer processes, and peak widths are expected to be similar to that of a reversible surface confined process, which is 90.6 mV/n. A Gaussian function of varying width for the POT oxidation may be combined with the reversible ion transfer wave (Figure S10b) to obtain different peak widths (Figure S10c). The experimental anodic peak obtained for the detection of sodium ions perfectly agrees when a broader Gaussian wave for POT oxidation is considered (Figure S10c). Without having to introduce a potentially inadequate theoretical model that describes the experimentally observed behavior, this treatment nonetheless suggests that the broader than expected peak width originates from non-ideal POT electrochemistry. On other hand, the Nernstian shift of the ion transfer wave for different ion activities of the primary ion will result in a Nernstian shift of the
ACS Paragon Plus Environment
10
Page 11 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
total peak, because the two individual potential contributions are additives (Figure S11). This behavior is therefore independent of the specific peak shape of the underlying electrochemistry of the conducting polymer. As expected for MII, a Hofmeister selectivity pattern is observed (Figure 3a) for sodium, lithium and calcium. The selectivity coefficients were calculated by the separate solution method25, a common pot pot protocol for potentiometric sensors, as log K Na.Li = −0.27 and log K Na.Ca = −1.9 . The integrated charge for
the individual peaks confirms that a similar and reproducible amount of charge is transferred for each separately measured ion (2.719 µC for Na+, 2.702 µC for Li+ and 2.669 µC for Ca2+). Further experiments were carried out with membranes that contained either calcium or lithium ionophore (M3, M4, M5, M6, see Table 1). The individual voltammograms obtained for 10 mM calcium and 10 mM lithium chloride in comparison with 10 mM NaCl are shown in Figure 3b and 3c respectively. As a consequence of the incorporated ionophore, a remarkable shift in the ion transfer waves is observed (337 mV and 141 mV displacement for Ca2+ and Li+, respectively). The calculated selectivity coefficients 26 pot pot correspond to the reported ones for the same ionophores ( log K Li,Na = −12.8 ). = −2.4 and log KCa.Na
These membranes were further explored in a mixture of 10 mM NaCl and variable levels of CaCl2 (from 10-7 to 10-5 M). Cyclic voltammetry shows two anodic ion-transfer peaks. While the first wave is assigned to the expulsion of sodium from the membrane, the second one corresponds to calcium (Figure S12). The ionophore promotes the complexation of calcium (L3Ca2+) and provokes the mentioned potential shift. In the backward scan, the process is reversed and calcium and sodium are extracted back into the polymeric film. If only calcium is present in the solution or at higher concentration levels than 5 x 10-5 M, just a single wave is obtained in the forward and backward scans as shown in Figure S12. An excellent reproducibility is demonstrated for fifty consecutive scans, resulting in a current variation of just 0.03 µA at 965.9 mV. Incremental sample concentration changes of either lithium or calcium were characterized with three identical electrodes measured at the same time. After each addition, a cyclic voltammogram was recorded at 100 mV s-1, taking about 20-s per scan. The peak potential was found to gradually move to more positive values by obeying the Nernst equation (Figure 4a, c). Note that this was not observed in the absence of ionophore because of the highly interfering background electrolyte. The analytical responses of a range of membranes (MII-MVII) with different plasticizers are listed in Table S1. The use of o-NPOE slightly improves the limit of detection (Figure 4b, d). The limit of detection for these thin membranes is at least one order higher than for traditional ion-selective membranes (~200 µm)27, and no attempt was made here to optimize the detection limit in this early work.
ACS Paragon Plus Environment
11
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 12 of 29
We emphasize, however, that such thin layer membranes can be conveniently operated in stripping ion transfer voltammetry to achieve very low detection limits, as demonstrated by the group of Amemiya.14,15 The integrated charge and position of the forward and backward peaks was found to diminish slightly when the electrodes were washed with pure water (see Figure S13). To avoid leaching of the membrane components, electrodes were therefore rinsed with NaCl solution or wiped off with a paper tissue. Alternatively, a more rigid membrane (higher PVC amount, membrane VII) did not show evidence of leaching, but a deterioration of the limit of detection, slope and thin layer behavior was found (Table S1). The shapes of the forward and backward waves for calcium ionophore containing membrane are significantly different to the NaCl wave (Figure 4c). The depletion of calcium in the membrane shows a wider peak with a transfer potential of 966 mV with respect to the backward wave (830 mV). This effect was not observed with membranes containing lithium ionophore (Figure 4a). This might be explained by the higher 1:3 stoichiometry between calcium and ionophore, perhaps giving rise to a broadening owing to step-wise calcium complexation.28 Motivated by our recent work on thin layer voltammetry for halide discrimination29,30 and the results demonstrated above with solid contact ion-to-electron transducers (POT), we explored whether selective multianalyte detection would be possible with a membrane containing two ionophores. Clearly, this would not be possible by zero current potentiometry since it is essentially a one-dimensional technique.31 Figure 5a shows the behavior of Membrane IX, containing a lithium and calcium ionophore, when interrogated by cyclic voltammetry. The first peak corresponds to lithium transfer and the second one to calcium, as expected from the previous experiments with membranes containing a single ionophore. In principle, this membrane behaves similarly to that of two independent membranes (see Figure 3b and 3c). Incremental increases of lithium and calcium concentration again result in a Nernst shift to more positive potentials. This suggests that both ionophores can work in tandem with the underlying POT electrochemistry to resolve the two ions in solution. Calibration curves (Epeak vs. log activity) show Nernstian response slopes (59.1±0.1 and 29.0±0.1 mV for Li+ and Ca2+ respectively) with a limit of detection of (5.0±0.1) x 10-5 and (2.1±0.1) x 10-5 M for Li+ and Ca2+ respectively, see Figure 5b. A methodical adjustment of the cation exchanger composition was explored, keeping an invariable concentration of the two ionophores (MVIII, MIX and MX). The intuitive mechanism is put forward in Figure 6. A membrane (MVIII) containing half the molar amount of cation exchanger with respect to ionophore (0.5:1:1 molar ratio) only exhibited calcium ion-transfer (Figure S14a). As the formation constant for Ca2+ is roughly 1020 times greater than for Li+,26 the cation exchanger will preferentially stabilize the charge of [CaL3]2+ complex. Note that ion-selective membranes used in potentiometry must be formulated to exhibit a molar excess of uncomplexed ionophore to give a selective and Nernstian
ACS Paragon Plus Environment
12
Page 13 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
response to the ion of interest. A two-ionophore membrane would therefore invariably respond only to calcium, in analogy to the voltammetric results with Membrane MVIII. Increasing the amount of cation exchanger to a 1:1:1 molar ratio (MIX) resulted in an adequate discrimination between lithium and calcium (Figure S14b). In this case, the cation exchanger forces, at complete oxidation of POT, the conversion of the totality of calcium ionophore to the complex [CaL3]2+ and partially the lithium ionophore to [LiL’]+. On the other hand, if the cation exchanger concentration just exceeded the sum of both ionophore concentrations, a third peak associated to the extraction of an uncomplexed ion appeared in the voltammogram, deteriorating peak separation (Figure S14c). While the ideal membrane should exactly match the concentration of cation exchanger and both ionophores, MIX exhibited the sought-after characteristics and was used further. It is noted that a slight improvement of the detection limit was observed for calcium when using a gold substrate instead of GC (Table S4, Figure S15). The idealized theoretical description put forward above suggests that this methodology may be capable of potentiometric multianalyte detection. Figure 7 and Figure 8 show the influence of the standard iontransfer potential and the sample concentration on the voltammetric readout of a membrane containing two ionophores. As shown in Figure 7, a standard potential difference of 0.4 V results in complete resolution of both peaks. Smaller potential differences result in lower resolution (see Figure S16). Thin layer behavior is theoretically confirmed with the linear relationship between peak current and scan rate (Figure S17). Most importantly, theory confirms that each peak potential change obeys the Nernst equation for its respective ion when keeping the concentration of the other ion constant (Figure 8, Figure S18). This result can be rationalized as follows. The membrane is interrogated under ionophore deficient conditions. At the beginning of an anodic voltammetric scan, all ionophores in the membrane are in their complexed form, with the lipophilic ion-exchangers forming their counter ions. As POT is oxidized, the weakest bound complex is dissociated first, expelling lithium ions from the membrane. As the calcium ionophore is, at this stage, exhaustively bound to calcium, there is no interference from calcium for this process and the shift of the lithium extraction wave obeys the Nernst equation. At potentials more anodic than the lithium transfer wave, uncomplexed lithium ionophore is present in the membrane, along with calcium complex. Sufficient lithium-calcium peak separation is needed to ensure that this free ionophore does not influence the calcium peak position. It may be estimated that less than 1% error is observed for a peak separation of twice the Nernstian slope, but future work will evaluate this important aspect in more detail. If uncomplexed lithium ionophore is unable to compete with the calcium transfer wave, as in the
ACS Paragon Plus Environment
13
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 14 of 29
experiments observed here, again a selective and Nernstian potential response for calcium is expected, independent of the lithium concentration in the sample.
Conclusions Performing dynamic electrochemistry with solid contact ion-to-electron transducers such as POT may open new horizons for producing reversible, multianalyte selective, easy to miniaturize and disposable sensors in comparison to traditional ion-selective electrodes or other sensors based on ion-transfer voltammetry between two immiscible liquids.32 While the underlying electrochemistry results in broader peaks than expected for a reversible one electron redox couple, conducting polymers are an attractive materials platform for the achievement of electrochemically addressable thin films that also give opportunities for simultaneous optical detection. The key novelty put forward in this paper is the ability for thin layer voltammetric ion transfer membranes to detect multiple analytes selectively by potentiometric readout (change of peak potentials). This was achieved with membranes deficient of ionophore, which was controlled by the concentration of lipophilic ion-exchanger. Two well-resolved ion-transfer waves for lithium and calcium were observed that each gave Nernstian response behavior, independent of the concentration of the other ion. A simple theory adequately explains the results without the need to include diffusion coefficients of the mobile species.
ACS Paragon Plus Environment
14
Page 15 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Acknowledgments The authors thank the Swiss National Science Foundation and the European Union (FP7-GA 614002SCHeMA project) for supporting this research.
Associated Content Supporting Information. POT and membrane preparation, optimization of the membrane composition, calculated voltammograms. This material is available free of charge via the Internet at http://pubs.acs.org.
Author Information Corresponding Authors [email protected]; [email protected]
ACS Paragon Plus Environment
15
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 16 of 29
References (1)
Bobacka, J., Ivaska, A., Lewenstam, A., Chem. Rev. 2008, 108, 329-351.
(2)
Lindner, E., Gyurcsanyi, R. E., J. Solid State Electrochem. 2009, 13, 51-68.
(3)
Duezguen, A., Zelada-Guillen, G. A., Crespo, G. A., Macho, S., Riu, J., Xavier Rius, F., Anal. Bioanal. Chem. 2011, 399, 171-181.
(4)
Michalska, A., Electroanalysis 2012, 24, 1253-1265.
(5)
Bobacka, J., McCarrick, M., Lewenstam, A., Ivaska, A., Analyst 1994, 119, 1985-1991.
(6)
Bobacka, J., Anal. Chem. 1999, 71, 4932-4937.
(7)
Vazquez, M., Danielsson, P., Bobacka, J., Lewenstam, A., Ivaska, A., Sens. Actuators B Chem.
2004, 97, 182-189.
(8)
Bobacka, J., Ivaska, A., Grzeszczuk, M., Synthetic Metals 1991, 44, 21-34.
(9)
Veder, J.-P., De Marco, R., Patel, K., Si, P., Grygolowicz-Pawlak, E., James, M., Alam, M. T., Sohail, M., Lee, J., Pretsch, E., Bakkert, E., Anal. Chem. 2013, 85, 10495-10502.
(10)
Bobacka, J., Grzeszczuk, M., Ivaska, A., J. Electroanal. Chem. 1997, 427, 63-69.
(11)
Mattinen, U., Bobacka, J., Lewenstam, A., Electroanalysis 2009, 21, 1955-1960.
(12)
Cicmil, D., Anastasova, S., Kavanagh, A., Diamond, D., Mattinen, U., Bobacka, J., Lewenstam, A., Radu, A., Electroanalysis 2011, 23, 1881-1890.
(13)
Zou, X. U., Cheong, J. H., Taitt, B. J., Buehlmann, P., Anal. Chem. 2013, 85, 9350-9355.
(14)
Kim, Y., Amemiya, S., Anal. Chem. 2008, 80, 6056-6065.
(15)
Kim, Y., Rodgers, P. J., Ishimatsu, R., Amemiya, S., Anal. Chem. 2009, 81, 7262-7270.
ACS Paragon Plus Environment
16
Page 17 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
(16)
Kabagambe, B., Garada, M. B., Ishimatsu, R., Amemiya, S., Anal. Chem. 2014, 86, 7939-7946.
(17)
Si, P., Bakker, E., Chem. Comm 2009, 5260-5262.
(18)
Zhang, J., Harris, A. R., Cattrall, R. W., Bond, A. M., Anal. Chem. 2010, 82, 1624-1633.
(19)
Harris, A. R., Zhang, J., Cattrall, R. W., Bond, A. M., Anal. Methods 2013, 5, 3840-3852.
(20)
Feldman, B. J., Burgmayer, P., Murray, R. W., J. Am. Chem. Soc. 1985, 107, 872-878.
(21)
Michalska, A., Maksymiuk, K., Talanta 2004, 63, 109-117.
(22)
Bard, A. J., Faulkner, L. R., Electrochemical Methods. Fundamentals and Applications, John Wiley & Sons, Inc., USA, 2001, page 590-593,
(23)
Osteryou.Ra, Christie, J. H., Anal. Chem. 1974, 46, 351-355.
(24)
Skompska, M., Siwiec, D., Kudelski, A., Zagorska, M., Synthetic Metals 1999, 101, 35-36.
(25)
Umezawa, Y., Buhlmann, P., Umezawa, K., Tohda, K., Amemiya, S., Pure Appl. Chem. 2000, 72, 1851-2082.
(26)
Qin, Y., Mi, Y. M., Bakker, E., Anal. Chim. Acta 2000, 421, 207-220.
(27)
Bakker, E., Buhlmann, P., Pretsch, E., Chem. Rev. 1997, 97, 3083-3132.
(28)
Zhu, J., Qin, Y., Zhang, Y., Anal. Chem. 2010, 82, 436-440.
(29)
Cuartero, M., Crespo, G. A., Afshar, M. G., Bakker, E., Anal. Chem. 2014, 86, 11387-11395.
(30)
Cuartero, M., Crespo, G. A., Bakker, E., Anal. Chem. 2015, 87, 1981-1990.
(31)
Han, S. H., Lee, K. S., Cha, G. S., Liu, D., Trojanowicz, M., J. Chromatogr. 1993, 648, 283-288.
(32)
Zhou, M., Gan, S., Zhong, L., Su, B., Niu, L., Anal. Chem. 2010, 82, 7857-7860.
ACS Paragon Plus Environment
17
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
(33)
Page 18 of 29
Wilke, S., J. Electroanal. Chem. 2001, 504, 184-194.
ACS Paragon Plus Environment
18
Page 19 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Figures Captions Figure 1.
Proposed response mechanism of a cation exchanger (R-) based thin layer membrane backside contacted with POT during the a) forward scan and b) backward scan.
Figure 2.
Cyclic voltammograms for the cation exchanger based membrane MII (a) in contact with 10 mM NaCl at a scan rate 100 mV.s-1. (b) Calculated cyclic voltammograms at three concentration levels of NaCl (scan rate 100 mV.s-1); inset: Epeak vs. logaNa+. Parameters used R /O in the calculation: A=7.0686×10-4 dm2; δ film =2.3×10-6 dm; cR / O = 5×10-1 M; RT = 10×10-3 0 = 0.8V; EIT0 = -0.3 V (calculated from the standard Gibbs M; Einit= 0V; Eturn = 0.8V; ECT energy for sodium partition between water | nitrobenzene microinterface)33; s = 0.059 V;
cMaq =0.01, 0.0316 and 0.1 M. Figure 3.
(a) Observed cyclic voltammograms in contact with 10 mM NaCl, LiCl and CaCl2 for membrane MII. (b) Cyclic voltammograms for either 10 mM NaCl or 10 mM CaCl2 using MIII (calcium ionophore). (c) Cyclic voltammograms for either 10 mM NaCl or LiCl using MIV (lithium ionophore). Scan rate: 100 mV s-1.
Figure 4.
Membranes containing a single ionophore. (a) Observed cyclic voltammograms at increasing concentrations of lithium using membrane MIV (lithium ionophore). (b) Calibration curve for lithium using two different plasticizers. (c) Observed cyclic voltammograms at increasing concentrations of calcium using membrane MIII (calcium ionophore). (d) Calibration curve for calcium using two different plasticizers. Cation concentrations: 1×10-6, 2.5×10-6, 5×10-6, 1×10-5, 2.5×10-5, 5×10-5, 1×10-4, 2.5×10-4, 5×10-4, 1×10-3, 2.5×10-3, 5×10-3 and 1×10-2 M. Conditions: 100 mV s-1, background of 10 mM NaCl. Error bars for b) and d) are shown but are smaller than the plot markers.
Figure 5.
Membrane containing two ionophores. (a) Cyclic voltammograms obtained for different concentrations of lithium and calcium at equimolar concentrations (2.5×10-6, 5×10-6,1×10-5, 2.5×10-5, 5×10-5, 1×10-4, 2.5×10-4, 5×10-4, 1×10-3, 2.5×10-3, 5×10-3 and 1×10-2 M) using membrane MIX (b) Calibration curve obtained for calcium and lithium. Error bars for b) are shown but are smaller than the plot markers.
Figure 6.
Response mechanism for membranes containing two ionophores (L and L’) backside contacted with POT during a) the forward scan and b) the backward scan. I+ and J+ are the two analytes and R- is the cation-exchanger.
Figure 7.
Calculated voltammograms for two ionophore-based membranes. Parameters: ∆Eº = 0.4V; R /O =2.3×10-6 dm; cR / O = 5×10-1 M; RT = 10×10-3 M; υ = A=7.0686×10-4 dm-2; F = 96486; δ film aq
aq
0 100 mVs-1; Einit= -0.3 V; Eturn = 1.4V; ECT = 0.8V; s = 0.059 V; cI =0.01 M; c J = 0.01M.
Figure 8.
Calculated voltammograms for two ionophore-based membrane at different concentrations of the species J, keeping the concentration of I constant. Parameters: A=7.0686×10-4 dm-2; F R /O = 96486; δ film =2.3×10-6 dm; cR / O = 5×10-1 M; RT = 10×10-3 M; υ = 100 mVs-1; Einit= -0.3 aq aq 0 V; Eturn = 1.4V; ECT = 0.8V; ∆E I0, J = 0.4V; s = 0.059 V; cI =0.01 M; cJ = 10-2, 3×10-3,
1×10-3, 3×10-4, 10-4 M.
ACS Paragon Plus Environment
19
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 20 of 29
Table 1. Compositions of membrane cocktails prepared in 1 mL of THF. Membrane
a b
Components PVCa
DOSa
I
31
II
NPOEa
NaTFPBb
TDDAClb
Ca Ionophoreb
Li Ionophoreb
63
40
40
33
63
41
III
29
61
40
IV
30
63
42
V
29
60
40
VI
30
62
41
82
VII
47
48
41
82
VIII
29
59
42
80
80
IX
29
57
80
81
80
X
29
55
139
82
80
80 82 80
weight percentage mmol kg-1
ACS Paragon Plus Environment
20
Page 21 of 29
Analytical Chemistry Figure 1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
(+)
(-)
e-
a)
POT+
POT
RMn+
SOLUTION
(+)
(-)
e-
b) POT
POT+ R-
SOLUTION
M+n
ACS Paragon Plus Environment
Analytical Chemistry Figure 2
4
a) CATION RELEASE
3
i/m A
2 1 0 -1 -2 -3
CATION EXTRACTION
-0.6 -0.4 -0.2
0
0.2
0.4
0.6
0.8
1
1.2
1.4
E/V 2
b)
410 E / mV
1.5
390 370
1 i/m A
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
350
0.5
-2.0
-1.5 -1.0 log aNaCl
0 -0.5 -1 -1.5 INCREASING NaCl -2 -0.6 -0.4 -0.2
0
0.2
0.4
0.6
0.8
1
1.2
1.4
E/V
ACS Paragon Plus Environment
Page 22 of 29
Page 23 of 29
Analytical Chemistry Figure 3
3
a) NO IONOPHORE
i/m A
2
1
0 CaCl2
-1
LiCl NaCl -2
-0.5
-0.2
0.1
0.4 E/V
0.7
1
1.3
b) CALCIUM IONOPHORE CaCl2
1.5
i/m A
NaCl 0.5
-0.5
-1.5
-2.5 -0.5
2
-0.2
0.1
0.4 0.7 E/V
1
1.3
c) LITHIUM IONOPHORE
1.5 NaCl 1 i/m A
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
LiCl
0.5 0 -0.5 -1 -1.5 -0.5
-0.2
0.1
0.4 E/V
0.7
1
ACS Paragon Plus Environment
1.3
1.6
Analytical Chemistry
Page 24 of 29
Figure 4
2
a) LITHIUM IONOPHORE
560 b)
510 Epeak / mV
i/m A
1
0
460
DOS o-NPOE
410 -1 360 INCREASING LITHIUM -2 0.4 -0.5 -0.2 0.1 0.7 E/V
310 1
1.3
1.5
-6
-5
-4 log a Li +
-3
-2
1060 d)
c) CALCIUM IONOPHORE 1030
1
1000
Epeak / mV
i/m A
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
0
970
o-NPOE
-1
DOS
940 -1.5
910 INCREASING CALCIUM
-0.5
-0.2
0.1
0.4
0.7
1
1.3
880
-6
E/V
ACS Paragon Plus Environment
-5
-4 log a Ca2+
-3
-2
Page 25 of 29
Analytical Chemistry Figure 5
1.5
i/m A
a) BOTH IONOPHORES
Ca2+
Li+
0.5
-0.5 Ca2+
Li+ -1.5 -0.5
Epeak / mV
950
-0.2
0.1
0.4 E/V
0.7
1
1.3
b)
900 850 log a Ca2+
530 510 Epeak / mV
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
490 470 450 430 410 -6
-5
-4 log a Li +
-3
ACS Paragon Plus Environment
-2
Analytical Chemistry Figure 6 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
(+)
(-)
e-
a)
ePOT+
POT
I++ L
RLI+
POT+
POT
RL’J+
J++ L’
SOLUTION
(+)
(-)
e-
b)
ePOT+
POT
POT+
POT
R-
RL’
L I+
J+
SOLUTION
ACS Paragon Plus Environment
Page 26 of 29
Page 27 of 29
Analytical Chemistry Figure 7
1.6
D E0 = 0.4 V
1.2 i/m A
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
0.8
0.4
0 0
0.4
0.8
1.2
E/V
ACS Paragon Plus Environment
Analytical Chemistry
Page 28 of 29
Figure 8
1.6 FIXED CI
INCREASING CJ
1.2 i/m A
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
0.8
0.4
0 0
0.4
0.8
1.2
E/V
ACS Paragon Plus Environment
Page 29 of 29
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
For TOC Only
ACS Paragon Plus Environment
