Calibration-Free Ionophore-Based Ion-Selective Electrodes With a Co

Article pubs.acs.org/ac

Calibration-Free Ionophore-Based Ion-Selective Electrodes With a Co(II)/Co(III) Redox Couple-Based Solid Contact Xu U. Zou, Xue V. Zhen, Jia H. Cheong, and Philippe Bühlmann* Department of Chemistry, University of Minnesota, 207 Pleasant Street SE, Minneapolis, Minnesota 55455, United States S Supporting Information *

ABSTRACT: A high electrode-to-electrode reproducibility of the emf response of solid contact ion-selective electrodes (SCISEs) requires a precise control of the phase boundary potential between the ion-selective membrane (ISM) and the underlying electron conductor. To achieve this, we introduced previously ionophore-free ion exchanger membranes doped with a well controlled ratio of oxidized and reduced species of a redox couple as redox buffer and used them to make SC-ISEs that exhibited highly reproducible electrode-to-electrode potentials. Unfortunately, ionophores were found to promote the loss of insufficiently lipophilic species from the ionophore-doped ISMs into aqueous samples. Here we report on an improved redox buffer platform based on equimolar amounts of the much less hydrophilic Co(III) and Co(II) complexes of 4,4′-dinonyl2,2′-bipyridyl, which makes it possible to extend the redox buffer approach to ionophore-based ISEs. For example, K+-selective electrodes based on the ionophore valinomycin exhibit electrode-to-electrode standard deviations as low as 0.7 mV after exposure of freshly prepared electrodes for 1 h to aqueous solutions. Exposure of freshly prepared ISE membranes to humidity prior to their first contact to electrolyte solution minimizes the initial (reproducible) emf drift. This redox buffer has also been successfully applied to sodium, potassium, calcium, hydrogen, and carbonate ion-selective electrodes, which all exhibit the high selectivity over interfering ions as expected for ionophore-doped ISE membranes.

I

standard deviation of the cell constant,33,34 but the extension of the Ag+/0 system to ISE membranes doped with different types of ionophores has not been reported and appears to be impeded by Ag+ loss into the aqueous phase. Much work has been invested in developing conducting polymers as intermediate layer materials.12,15,23 Interestingly, synchrotron radiation-X-ray photoelectron spectroscopy (SR-XPS) and near edge X-ray absorption fine structure (NEXAFS) were applied to poly(3-octylthiophene) based solid contact ISEs, confirming that a mixed oxidation state exists, which is consistent with the (moderately) stable potential readings for this type of electrode.35 Problems of continued interest for work with solid contacts based on conducting polymers include the improvement of electrode-to-electrode reproducibilities and stabilities in long-term measurements. Applying a current pulse in the nanoamp range to solid contact ISEs with an intermediate layer consisting of poly(3,4-ethylene dioxythiophene) doped with poly(sodium 4-styrenesulfonate) provides an interesting route to electrochemically control the phase boundary potentials at the intermediate layer,36 which makes it possible to prepare sensors with identical calibration curves, but it requires tuning of individual electrodes and has not been applied to routine applications yet. The less than ideal long-

on-selective electrodes (ISEs) are widely used in clinical analysis, process control, and environmental monitoring.1−6 In view of mass production, affordability, and portability, traditional ISEs with internal electrolyte are less desirable than solid contact ISEs, which comprise an ion-selective membrane (ISM) that is in direct contact to a solid electron-conducting substrate.7−10 Solid contact ISEs usually have an intermediate layer between the ISM and the underlying electron conductor to prevent the formation of a water layer and stabilize the phase boundary potential across the interface of the ISM and the electron conductor. This layer is often a conducting polymer11−25 or a carbon material with a high surface area.26−32 Since the electromotive force (emf) that is measured with an ISE is the sum of the sample-dependent potential at the sample/ISM interface and all sample-independent interfacial potentials in the electrochemical cell,1,3 poorly defined phase boundary potentials at the interface of the ISM, this intermediate layer, and the electron conductor of solid contact ISEs make it necessary that each sensor is individually calibrated once before its first use and in frequent intervals thereafter.3 The goal of calibration-free ISEs has so far been elusive. Unfortunately, calibrations are cumbersome and timeconsuming not only for routine analysis but also for single use applications by less trained personnel. Efforts to develop solid contact ISEs without poorly defined phase boundary potentials have been made in several research groups. Ion-selective membranes doped with Ag+ ionophore complexes on a silver epoxy substrate were shown to reduce the © 2014 American Chemical Society

Received: May 3, 2014 Accepted: August 12, 2014 Published: August 12, 2014 8687

dx.doi.org/10.1021/ac501625z | Anal. Chem. 2014, 86, 8687−8692

Analytical Chemistry

Article

of the synthesis of these compounds are included in the Supporting Information. Elemental analyses were performed by MHW Laboratories (Phoenix, AZ). Glassy carbon and gold electrodes were purchased from CH Instruments (Austin, TX). 1H NMR spectroscopy was performed with a VAC-300 Varian Mercury or Varian Inova 500 spectrometer (Varian, Palo Alto, CA). UV−vis spectroscopy was performed with a JASCO V-630 spectrophotometer (Easton, MD). Electrode Fabrication. The 3 mm diameter glassy carbon disk electrodes were polished with aqueous dispersions of alumina (0.3 μm, Buehler, Lake Bluff, IL) over polishing cloths. The 2 mm diameter gold disk electrodes were polished over polishing cloths with aqueous dispersions of alumina (0.3 and 0.05 μm) and then cleaned in piranha solution (concentrated sulfuric acid and 30% hydrogen peroxide solution in a 3:1 ratio). Caution: piranha solution is a strong oxidizing reagent, is highly corrosive, and should be handled with care. All electrodes were then cleaned by ultrasonication in water and ethanol and dried with a flow of nitrogen. For the formation of a self-assembled monolayer (SAM) of 1-hexanethiol, gold electrodes were immersed for 24 h into a 1 mM solution of 1hexanethiol in ethanol. Solvent polymeric membranes doped with the redox couple were prepared by dissolving the membrane components in 500 μL of freshly distilled THF. The membrane components included 33 mg of PVC as polymer matrix, 66 mg of DOS as plasticizer, 0.8 mmol/kg [Co(C9,C9-bipy)3](TPFPB)2, 0.8 mmol/kg [Co(C9,C9-bipy)3](TPFPB)3, ionophore, and ionic sites (reported in the following in mol % relative to the ionophore). Specifically, the Na+-selective membranes contained 1.0% (w/w) sodium ionophore and 51 mol % LiTPFPB, the K+-selective membranes 1.2% (w/w) valinomycin and 59 mol % LiTPFPB, the Ca2+-selective membranes 1.3% (w/w) calcium ionophore and 43 mol % LiTPFPB, and the H+selective membranes 1.0% (w/w) hydrogen ionophore and 71 mol % LiTPFPB. The CO32−-selective membranes were prepared from 41 mg of PVC, 55 mg of DOS, 2.6% (w/w) carbonate ionophore, 43 mol % TDDMACl, 8 mmol/kg [Co(C9,C9-bipy)3](TPFPB)2, and 8 mmol/kg [Co(C9,C9bipy)3](TPFPB)3). In each case, 40 μL of the solution with the ISM component was drop-casted onto a SAM-modified gold or glassy carbon electrode, which was then left to dry overnight. The electrodes prepared this way were conditioned in 1.0 mM KCl solution for 1 h prior to measurements. The short conditioning time ensured that none of the experiments focusing on the observation of the redox capacity was affected by the formation of a water layer between the ISM and the underlying solid contact, which was a concern for the glassy carbon electrodes. Potentiometric Measurements. Potentials were measured with an EMF 16 potentiometer (input impedance 10 TΩ) controlled with EMF Suite 1.03 software (Lawson Labs, Malvern, PA). A double-junction type external reference electrode (DX200, Mettler Toledo, Switzerland; 3.0 M KCl saturated with AgCl as inner filling solution and 1.0 M LiOAc as bridge electrolyte) was used. The solutions’ pH was measured with a pH glass electrode. CO32− concentrations were calculated as suggested by Herman and Rechnitz.42 Activity coefficients were calculated according to a twoparameter Debye−Hückel approximation,43 and all emf values were corrected for liquid-junction potentials with the Henderson equation.44

term emf stability characteristics of solid contact electrodes based on conducting polymers are likely related to the occurrence of numerous local conformation of these polymers, resulting in a redox activity over a wide range of potentials as opposed to the redox buffer properties of a well-defined, individual redox couple.37 Redox-active self-assembled monolayers (SAMs) with a redox potential controlled by an applied current38 gave sensors with small drifts but had only a small redox buffer capacity. Also, an approach based on a redox buffer provided by gold nanoclusters present in two different charge states39 has suffered from the complexity of preparing these nanoclusters and has not been tested extensively. To address the need for a better redox buffer for solid contact ISEs, we previously reported that an ion exchanger membrane with the tris(1,10-phenanthroline) complexes of Co(II) and Co(III) in a 1:1 ratio significantly reduced the standard deviation of the y-axis intercept, E°, of the emf calibration curve.40,41 However, preliminary results in our laboratory showed that that this redox couple was not suitable for use with ionophore-based cation-selective electrodes because binding of the target cations to the ionophore was energetically so favorable that it caused primary cation transfer from aqueous samples into the ISE membranes, coupled with loss of the redox-active cations from the ISE membrane into the aqueous solution. Here, we describe a new, more lipophilic redox couple buffer system consisting of the tetrakis(pentafluorophenyl)borate (TPFPB−) salts of cobalt(II) and cobalt(III) tris(4,4′-dinonyl-2,2′-bipyridyl) (abbreviated in the following as [Co(C9,C9-bipy)3]2+ and [Co(C9,C9-bipy)3]3+, respectively) and their use to dope ionophore-based plasticized poly(vinyl chloride) (PVC) membranes. This new redox couple does not interfere with the Nernstian response of ionophorebased ISEs to their target ions and gives high electrode-toelectrode reproducibilities.

■

EXPERIMENTAL SECTION Materials. All reagents were of the highest commercially available purity and were used as received. Deionized water (0.18 MΩ m specific resistance) was purified with a Milli-Q PLUS reagent grade water system (Millipore, Bedford, MA). 1Hexanethiol was purchased from Acros Organics (Geel, Belgium). High-molecular weight poly(vinyl chloride) (PVC) and 2-nitrophenyl octyl ether (o-NPOE) were purchased from Fluka (Buchs, Switzerland), dioctyl sebacate (DOS) from Wako Pure Chemical Industries (Osaka, Japan), bis[(12-crown4)methyl] dodecylmethylmalonate (sodium ionophore), and N,N-dicyclohexyl-N′,N′-dioctadecyl-3-oxapentanediamide (calcium ionophore) from Fluka (St. Louis, MO), lithium tetrakis(pentafluorophenyl)borate ethyl etherate from Boulder Scientific (Boulder, CO), and valinomycin (a potassium ionophore), heptyl 4-trifluoroacetylbenzoate (a carbonate ionophore), octadecyl isonicotinate (a hydrogen ionophore), and 4,4′-dinonyl-2,2′-bipyridyl from Sigma-Aldrich (St. Louis, MO). Co(II) and Co(III) complexes with a total of six nonyl groups were synthesized as follows: Briefly, [Co(II) (C9,C9bipy) 3 ]Cl 2 was prepared from CoC1 2 ·6H 2 O and the commercially available 4,4′-dinonyl-2,2′-bipyridyl as reported.45 Oxidation to the Co(III) was performed with Br2 as oxidant using a modified literature procedure,45,46 and the TPFPB− salts of [Co(C9,C9-bipy)3]3+ and [Co(C9,C9-bipy)3]2+ were prepared by metathesis using LiTPFPB ethyl etherate. Details 8688

dx.doi.org/10.1021/ac501625z | Anal. Chem. 2014, 86, 8687−8692

Analytical Chemistry

■

Article

RESULTS AND DISCUSSION Selection of the Redox Couple. It was previously reported that ion exchanger membranes doped with the tetrakis(pentafluorophenyl)borate salts of the redox couple cobalt(II/III) tris(1,10-phenanthroline) exhibited a high redox buffer capacity and emf stability.40 On the basis of the same principle as the well-known traditional pH or redox buffers for aqueous solutions,37 the hydrophobic redox couple in the ISM determined the phase boundary potential at the interface to the underlying electron conductor, as predicted by the Nernst equation. C 2.303RT emf = E°′ − log red nF Cox (1)

small changes in the scan rate, which suggests that the peak separation of 73 mV (rather than the theoretically expected 59 mV) is probably affected by a double layer effect and is not an indication of slow electron transfer. We conclude that the electron transfer at both bare gold electrodes and at SAM modified gold electrodes is fast enough to be able to determine the phase boundary potential at the interface of ion-selective membranes and the underlying electron conductor, as this was similarly the case also for the more hydrophilic redox couple cobalt(II/III) tris(1,10-phenanthroline) described earlier.40,41 Selectivities and Reproducibility of Electrode Potentials. The [Co(C9,C9-bipy)3]3+/2+ redox buffer was successfully applied to solid contact ISEs with glassy carbon as underlying electron conductor and ion-selective membranes doped with one of five ionophores, as shown by the electrode performances summarized in Table 1. For the cation-selective electrodes, the slopes of the emf response were all very close to Nernstian. All electrodes exhibited a high selectivity over interfering ions (e.g., pot log Kpot Na+,K+ = −1.65 ± 0.03, log KK+,Na+ = −4.22 ± 0.03, log pot KH+,Na+ = −5.60 ± 0.05 for the Na+, K+, and H+ ISEs, pot respectively, and log Kpot Ca2+,K+ = −5.57 ± 0.21, log KCa2+,Na+ = 2+ + + −4.10 ± 0.09 for the Ca ISE). For the Na , K , and H+ ISEs, these values are all very close to the values reported in the literature31,47,48 (see Table 1). For the Ca2+ ISEs, a smaller selectivity was observed in this work, which is likely caused by the use of DOS as a plasticizer here while o-NOPE had been used as a plasticizer in the work reported in the literature.49 The selectivity coefficients may have also been affected somewhat by minor differences in the method used for the determination of the selectivity coefficients.4 As an example for an anion-selective electrode, a CO32−selective electrode was prepared. It showed a Nernstian slope of −29.9 ± 1.5 mV/decade, and a selectivity over Cl− of log Kpot CO32−,Cl− = −5.34 ± 0.12, which is a slight improvement over the literature value of −4.8 for ISE membranes without a redox couple.50 Evidence for the effective role of the redox couple [Co(C9,C9-bipy)3]3+/2+ as redox buffer was obtained from the electrode-to-electrode reproducibility of the calibration curves. The standard deviations of E° were 2.6, 1.4, 1.4, 2.8, and 1.7 mV for the Na+, K+, H+, Ca2+, and CO32− ISEs, respectively. These values are all very similar to the 1.7 mV reported for the ion-exchanger electrodes previously. Polymeric Membranes Doped with a Redox Couple on Top of Gold Contacts Modified with a SelfAssembled Alkanethiol Monolayer. As shown previously, modified gold electrodes coated with ionophore-free ion exchanger membranes immersed in a 1.0 mM KCl solution showed a good long-term stability over 2 weeks. The formation of a water layer between the membrane and the underlying substrate could be avoided more efficiently than for electrodes with glassy carbon contacts.40 Therefore, Na+ and K+ selective ionophore-doped membranes on modified gold electrodes were fabricated here in the same way. They exhibited not only the same high selectivities (log Kpot Na+,K+ = −1.65 ± 0.01 and log + Kpot and K+ ISEs, respectively) + + = −4.08 ± 0.01, for the Na K ,Na as the corresponding solid contact electrodes with glassy carbon contacts but indeed performed better in terms of the E° electrode-to-electrode reproducibility (see Table 2). After being conditioned for 1 h in water, the Na+ and K+ ISEs with gold contacts exhibited an excellent standard deviation of E° of 0.9 and 0.7 mV, respectively.

As a result, these ion exchanger electrodes exhibited emf calibration curves with a standard deviation of the y-axis intercept, E°, as low as 1.7 mV after conditioning of the freshly prepared electrodes for 1 h.40 However, when solid contact ISEs were prepared with plasticized PVC membranes doped with the same redox couple and valinomycin as potassium ionophore, the standard deviation of E° was in the range of tens of millivolts, i.e., comparable to the standard deviation of E° of 63.5 mV for solid contact electrodes based on ion exchanger membranes without a redox couple.40 Besides this significant worsening of the electrode-to-electrode reproducibility of E°, the selectivity of these ISEs over interfering ions was also lost immediately, apparently because valinomycin was so good at facilitating the transfer of K+ from the aqueous samples into the ISE membranes that this process was driving the loss of the redox-active cations from the sensing membrane into the aqueous phase through ion exchange. To overcome this problem, a more hydrophobic redox couple was synthesized in this work using a bipyridyl ligand substituted with a nonyl group on each pyridine ring, resulting in redox-active Co(II) and Co(III) complexes with a total of six nonyl groups (Figure 1). [Co(C9,C9-bipy)3](TPFPB)2 and

Figure 1. Structural formulas of cobalt(II/III) (4,4′-dinonyl-2,2′bipyridyl) and tetrakis(pentafluorophenyl)borate (n = 2 or 3).

[Co(C9,C9-bipy)3](TPFPB)3) are soluble in plasticized PVC membranes and are electrochemically stable. Cyclic voltammograms (CVs) of [Co(C9,C9-bipy)3]3+ in acetonitrile taken with a scan rate of 100 mV/s show the same peak separation of 73 mV for both bare and 1-hexanethiol-modified gold macroelectrodes (for the CV, see Figure S2 of the Supporting Information). This is consistent with a fast one-electron transfer from [Co(C9,C9-bipy)3]3+ to [Co(C9,C9-bipy)3]2+ and planar diffusion of the redox active species. This indicates that the self-assembled layer has no substantial effect on the heterogeneous electron transfer at the SAM-modified gold electrode. The peak separation did not change significantly with 8689

dx.doi.org/10.1021/ac501625z | Anal. Chem. 2014, 86, 8687−8692

Analytical Chemistry

Article

Table 1. E° Values of Electrodes with Glassy Carbon As Solid Contact Based on Different Ionophores after Conditioning for 1 h in 1.0 mM Primary Ion Solutions (n = 6)a electrode type

K+

Na+

H+

Ca2+

CO32−

slope (mV/decade) E° (mV) log Kpot i,j (this work)

60.0 ± 0.8 731.3 ± 2.6 j = Na+ −4.22 ± 0.03

62.1 ± 0.5 571.7 ± 1.4 j = K+ −1.65 ± 0.03

56.4 ± 0.4 328.2 ± 1.4 j = Na+ −5.60 ± 0.05

29.9 ± 1.5 273.3 ± 1.7 j = Cl− −5.34 ± 0.12

(literature31,47−50)

j = Na+ −4.50

j = K+ −1.52

j = Na+ −5.6

32.1 ± 0.7 586.0 ± 2.8 j = Na+ −5.57 ± 0.21 j = K+ −4.10 ± 0.09 j = Na+ −6.4 j = K+ −5.6

j = Cl− −4.8

a All membranes contained PVC and DOS in a 1:2 weight ratio as well as ionophore and ionic sites. E° values refer to the potential of the ISE versus a reference electrode, as obtained by extrapolation of the linear section of the emf response to the primary ion activity of 1.0 M.

Table 2. Response of ISEs Composed of a Solvent Polymeric Membrane Doped with a Redox Buffer on a SAM Modified Gold Contact after Storage in 1.0 mM KCl or 1.0 mM NaCl K+ (n = 5)

type

Na+ (n = 6)

time (h)

slope (mV/decade)

E° (mV)

logKpot K+,Na+

1 24 48 72

60.1 ± 0.3 60.9 ± 0.5 N/A N/A

731.4 ± 0.7 716.8 ± 16.3 N/A N/A

−4.08 ± 0.01 −4.29 ± 0.12 N/A N/A

slope (mV/decade) 62.5 63.4 62.5 62.2

± ± ± ±

0.6 0.3 0.2 0.3

E° (mV) 561.1 557.0 549.6 537.9

± ± ± ±

0.9 2.0 4.5 8.2

log Kpot Na+,K+ −1.65 −1.65 −1.61 −1.59

± ± ± ±

0.01 0.03 0.03 0.03

mM KCl solution and observation of the emf over 1 h. As Figure 3 shows, the prior exposure to humidity reduced the

While oxygen was shown to form a half cell on Pt and Cu substrates,52,53 solid contact electrodes with redox buffer doped polymer membranes on SAM-modified gold substrates show no interference from oxygen. They were immersed into 1 mM KCl solutions, and O2 and Ar were bubbled alternatingly into the aqueous sample solution while recording the emf. As Figure S3 shows (see the Supporting Information), no significant potential changes were observed. The stabilizing effect of the redox couple on the phase boundary potential at the interface of the ISM and the underlying electron conductor makes it possible to study the dependence of the emf on water uptake into the ISE membrane. As shown by Figure 2, the emf of K+ ISEs exposed

Figure 3. emf response of a solid contact K+ ISE with a polymeric membrane on a 1-hexanethiol modified gold electrode: The electrode was exposed to high humidity for 24 h prior to its first contact with an aqueous solution (0.1 mM KCl) at t = 0 h.

initial emf drift over 1 h from 44 to 3.3 mV (see the Supporting Information). In consistency with this finding, cyclic voltammetry of [Co(C9,C9-bipy)3]3+ in acetonitrile showed that there was no effect of water (up to 100 mM) and oxygen on the peak current potentials of the [Co(C9,C9-bipy)3]3+/2+ redox couple (see Figure S4 of the Supporting Information). While the new redox couple [Co(C9,C9-bipy)3]3+/2+ made it possible to extend the redox buffer approach to solid contact ionophore-based ISEs and gave excellent initial E° electrode-toelectrode reproducibilities, longer exposure of solid contact ISEs to highly concentrated primary ion solutions still resulted in the loss of redox couple cations into the aqueous sample. This is, for example, shown by a comparison of the emf of K+ ISEs immersed in 0.1 mM and 1 M KCl. After 1 h, the difference in the emf for the two solutions was 204 mV, which is close to the theoretically expected difference for solutions

Figure 2. emf responses of K+-selective electrodes on a 1-hexanethiol modified gold substrate in 1 M and 0.1 mM KCl.

to 0.1 mM KCl and 1 M KCl solutions increased within the first 1 h by 44 and 35 mV, respectively. Because longer term drifts resulted in decreases rather than increases in the emf, we suspected that this initial drift was caused by the uptake of water into the polymeric membranes. This hypothesis was confirmed by exposure of freshly prepared electrodes to high humidity for 24 h, followed by immersion of the ISEs into 0.1 8690

dx.doi.org/10.1021/ac501625z | Anal. Chem. 2014, 86, 8687−8692

Analytical Chemistry

Article

differing by 3.8 orders of magnitude in K+ activity. This indicates that the high KCl concentration initially does not impede the normal response mechanism of the electrode. However, the further decrease of 158 mV of the electrode in 1 M KCl after an additional 45 h was much larger than the additional 9 mV decrease for the electrode stored in 0.1 mM KCl (Figure 2). This finding seems to be consistent with the loss of cobalt complexes into the aqueous electrolyte solution. As follows from the ion exchange equilibria described by eqs 2A and 2B (where L stands for the ionophore), it is evident that a higher concentration of K+ in the aqueous phase would be predicted to result in a faster depletion of the cobalt complex concentration in the membrane phase.

Addition of concentrated aqueous sodium tetraphenylborate produced a precipitate, which was dissolved in diethyl ether. Washing of the organic phase with water and evaporation of the solvent gave a solid that was redissolved in tetrahydrofuran to give a UV−vis spectrum with an absorption maximum at 310 nm. This value is very close to the absorption maxima of 310 and 307 nm for [Co(C9,C9-bipy)3]3+ and [Co(C9,C9-bipy)3]2+, respectively. Neither tetraphenylborat e, t etrakis(pentafluorophenyl)borate, nor the plasticizer show substantial absorption above 300 nm. This is consistent with the conclusion that the exposure of the ion-selective membranes to a high primary ion concentrations resulted in a loss of redoxactive species into the solution. (Details of the extraction process can be found in the Supporting Information.) Not surprisingly, for a set of three K+ ISEs, the increase of the standard deviation of E° as a function of the immersion time into KCl solution is substantial for the more concentrated 1 M KCl but much smaller for the 0.1 mM KCl solution (see Figure 4).

3Lmem + 3K aq + + Ox mem 3 + ⇌ 3LK mem+ + Ox aq 3 + (2A)

2Lmem + 2K aq + + Red mem 2 + ⇌ 2LK mem+ + Redaq 2 + (2B)

The ion exchange reaction in eq 2A for the oxidized redox active complex, Ox3+, assisted by the ionophore, L, is described by the following equilibrium constant: 3

Keq,Ox =

[LK+]mem [Ox 3 +]aq 3

[K+]aq [L]mem 3 [Ox 3 +]mem

= k OxkK 3β 3 (3)

where kOx = [Ox ]aq/[Ox ]mem, kK = [K ]mem/[K ]aq, and β = [LK+]mem/([K+]mem[L]mem). Using eq 3, the leaching process as a function of time can be described using a similar model as described previously for ionophore and ionic site loss out of ISE membranes54−56 3+

In

[Ox 3 +]mem, t = 0 [Ox 3 +]mem, t

3+

=

Daq t k Oxdδ

+

=

+

kK 3Daq β 3t dδKeq,Ox

Figure 4. Standard deviation of E° of K+-selective electrodes on a 1hexanethiol modified gold substrate for sets of three electrodes that were continuously kept for 50 h in 1.0 M or 0.1 mM KCl.

(4)

where Daq is the diffusion coefficient of the cobalt complex in water, d is the thickness of the membrane, δ is the thickness of the Nernstian boundary layer contacting the membrane phase, and t is time. The analogous equation for the loss of the reduced complex species, Red2+, has a right-hand term proportional to (kK2β2/Keq,Red). Equation 4 suggests that the rate of the loss of redox active species into the aqueous solution increases with the stability of the ionophore complex, β, and depends on the energy of phase transfer of the primary ion into the ISE membrane, as described by kK. Indeed, while Na+-selective electrodes continuously immersed into 1.0 mM NaCl solutions only showed increases in the standard deviation of E° from 0.9 mV after 1 h to 2.0 mV after 24 h (and as little as 8.2 mV after 3 days), K+-selective electrodes immersed in a 1.0 mM KCl solution showed increases from 0.7 mV after 1 h to 16.3 mV after 24 h. This is consistent with the very large binding constant 1010.1 of the complexes of K+ and valinomycin in DOS as compared to the substantially smaller stability of the complex of Na+ and sodium ionophore VI of 106.6 in DOS.51 Note also that solving eq 4 for [Ox3+]mem and insertion of the resulting term into eq 1 predicts a linear decrease of the emf with time, as it is indeed observed in the case of 1 M KCl over the first day of measurement (see Figure 2). Loss of cobalt complex into the aqueous phase could be confirmed also by UV−vis spectroscopy (Figure S5 in the Supporting Information). For this purpose, K+-selective membranes were equilibrated with a 1 M KCl solution.

■

CONCLUSIONS The results reported here demonstrate a platform with high redox buffer capacity, generating a highly reproducible phase boundary potential at the interface between an ion-selective membrane and the underlying electron conductor. The thus prepared solid contact K+ ISEs show E° values with an electrode-to-electrode standard deviation as low as 0.7 mV when gold contacts modified with a self-assembled monolayer of 1-hexanethiol are used. This redox buffer was also applied to Na+, Ca2+, H+, and CO32− ISEs, all of which show highly reproducible E° values and selectivities over interfering ions. While the use of redox buffer species with a much improved hydrophobicity extends this approach to ionophore-based ISEs, long-term measurements with very high primary ion activities show that loss of the redox buffer species into aqueous samples cannot be fully prevented even for this redox couple with six nonyl groups per redox center. To address this problem, ongoing efforts in our laboratory target redox couples with smaller charge numbers and polymers modified with covalently attached redox couples.57

■

ASSOCIATED CONTENT

S Supporting Information *

(1) Procedures for the synthesis and 1H and 19F NMR spectra of the metal complexes, (2) structural formulas of all 8691

dx.doi.org/10.1021/ac501625z | Anal. Chem. 2014, 86, 8687−8692

Analytical Chemistry

Article

(25) Konopka, A.; Sokalski, T.; Michalska, A.; Lewenstam, A.; MajZurawska, M. Anal. Chem. 2004, 76, 6410−6418. (26) Lai, C.-Z.; Joyer, M. M.; Fierke, M. A.; Petkovich, N. D.; Stein, A.; Bühlmann, P. J. Solid State Chem. 2009, 13, 123−128. (27) Rius-Ruiz, F. X.; Crespo, G. A.; Bejarano-Nosas, D.; Blondeau, P.; Riu, J.; Rius, F. X. Anal. Chem. 2011, 83, 8810−8815. (28) Crespo, G. A.; Macho, S.; Bobacka, J.; Rius, F. X. Anal. Chem. 2009, 81, 676−681. (29) Crespo, G. A.; Macho, S.; Rius, F. X. Anal. Chem. 2008, 80, 1316−1322. (30) Fouskaki, M.; Chaniotakis, N. Analyst 2008, 133, 1072−1075. (31) Li, F.; Ye, J.; Zhou, M.; Gan, S.; Zhang, Q.; Han, D.; Niu, L. Analyst 2012, 137, 618−623. (32) Ping, J.; Wang, Y.; Wu, J.; Ying, Y. Electrochem. Commun. 2013, 13, 1529−1532. (33) Liu, D.; Meruva, R. K.; Brown, R. B.; Meyerhoff, M. E. Anal. Chim. Acta 1996, 321, 173−183. (34) Lutze, O.; Meruva, R. K.; Frielich, A.; Ramamurthy, N.; Brown, R. B.; Hower, R.; Meyerhoff, M. E. Fresenius. J. Anal. Chem. 1999, 364, 41−47. (35) Veder, J.-P.; De Marco, R.; Patel, K.; Si, P.; Grygolowicz-Pawlak, E.; James, M.; Alam, M. T.; Sohail, M.; Lee, J.; Pretsch, E.; Bakker, E. Anal. Chem. 2013, 85, 10495−10502. (36) Vanamo, U.; Bobacka, J. Electrochim. Acta 2014, 122, 316−321. (37) de Levie, R. J. Chem. Educ. 1999, 76, 574. (38) Fibbioli, M.; Bandyopadhyay, K.; Liu, S.-G.; Echegoyen, L.; Enger, O.; Diederich, F.; Bühlmann, P.; Pretsch, E. Chem. Commun. 2000, 0, 339−340. (39) Zhou, M.; Gan, S.; Cai, B.; Li, F.; Ma, W.; Han, D.; Niu, L. Anal. Chem. 2012, 84, 3480−3483. (40) Zou, X. U.; Cheong, J. H.; Taitt, B. J.; Bühlmann, P. Anal. Chem. 2013, 85, 9350−9355. (41) Zou, X. U.; Chen, L. D.; Lai, C.-Z.; Bühlmann, P. Electroanalysis 2014, in press. (42) Herman, H. B.; Rechnitz, G. A. Science 1974, 184, 1074−1075. (43) Meier, P. C. Anal. Chim. Acta 1982, 136, 363−368. (44) Morf, W. E. The Principles of Ion-Selective Electrode and of Membrane Transport; Elsevier: New York, 1981. (45) Sapp, S. A.; Elliott, C. M.; Contado, C.; Caramori, S.; Bignozzi, C. A. J. Am. Chem. Soc. 2002, 124, 11215−11222. (46) Barton, J. K.; Raphael, A. L. J. Am. Chem. Soc. 1984, 106, 2466− 2468. (47) Chou, J.-C.; Huang, Y.-P.; Chen, C.-C. Method for sodium ion selective electrode, sodium ion selective electrode therefrom and sodium ion sensing device. U.S. Patent 7,994,546 B2, August 19, 2011. (48) Oesch, U.; Brzózka, Z.; Xu, A.; Simon, W. Med. Biol. Eng. Comput. 1987, 25, 414−419. (49) Sokalski, T.; Ceresa, A.; Fibbioli, M.; Zwickl, T.; Bakker, E.; Pretsch, E. Anal. Chem. 1999, 71, 1210−1214. (50) Behringer, C.; Lehmann, B.; Haug, J.-P.; Seiler, K.; Morf, W. E.; Hartman, K.; Simon, W. Anal. Chim. Acta 1990, 233, 41−47. (51) Qin, Y.; Mi, Y.; Bakker, E. Anal. Chim. Acta 2000, 421, 207− 220. (52) Cattrall, R. W.; Drew, D. M.; Hamilton, I. C. Anal. Chim. Acta 1975, 76, 269−277. (53) Meruva, R. K.; Meyerhoff, M. E. Electroanalysis 1995, 7, 1020− 1026. (54) Dinten, O.; Spichiger, U. E.; Chaniotakis, N.; Gehrig, P.; Rusterholz, B.; Morf, W. E.; Simon, W. Anal. Chem. 1991, 63, 596− 603. (55) Bakker, E.; Pretsch, E. Anal. Chim. Acta 1995, 309, 7−17. (56) Bühlmann, P.; Umezawa, Y.; Rondinini, S.; Vertova, A.; Pigliucci, A.; Bertesago, L. Anal. Chem. 2000, 72, 1843. (57) Pawlak, M.; Bakker, E. Electroanalysis 2014, 26, 1121−1131.

ionophores, (3) CV of [Co(C9,C9-bipy)3](TPFB)2, (4) effect of oxygen on the response of a K+-selective ISE, (5) effect of oxygen and water on the CV of [Co(C9,C9-bipy)3](TPFB)3 in acetonitrile, and (6) UV−vis spectrum of [Co(C9,C9-bipy)3]3+, [Co(C9,C9-bipy)3]2+, and a redox active species leached out from a K+-selective membrane. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS

■

REFERENCES

This research was supported by the National Science Foundation (OISE Grant CHE-0809328) and a doctoral dissertation fellowship to X.U.Z. from the graduate school of the University of Minnesota.

(1) Bakker, E.; Bühlmann, P.; Pretsch, E. Chem. Rev. 1997, 97, 3083− 3132. (2) Bobacka, J.; Ivaska, A.; Lewenstam, A. Chem. Rev. 2008, 108, 329−351. (3) Bühlmann, P.; Chen, L. D. Ion-Selective Electrodes With Ionophore-Doped Sensing Membranes. In Supramolecular Chemistry: From Molecules to Nanomaterials; Gale, P. A., Steed, J. W., Eds. John Wiley & Sons, Ltd.: Hoboken, NJ, 2012; Vol. 5. (4) Bühlmann, P.; Pretsch, E.; Bakker, E. Chem. Rev. 1998, 98, 1593− 1688. (5) Johnson, R. D.; Bachas, L. G. Anal. Bioanal. Chem. 2003, 376, 328−341. (6) Yin, T.; Qin, W. Trends Anal. Chem. 2013, 51, 79−86. (7) Cattrall, R. W.; Freiser, H. Anal. Chem. 1971, 43, 1905−1906. (8) Michalska, A. Electroanalysis 2012, 24, 1253−1265. (9) Chumbimuni-Torres, K. Y.; Rubinova, N.; Radu, A.; Kubota, L. T.; Bakker, E. Anal. Chem. 2006, 78, 1318−1322. (10) De Marco, R.; Clarke, G.; Pejcic, B. Electroanalysis 2007, 19, 1987−2001. (11) Bobacka, J. Electroanalysis 2006, 18, 7−18. (12) Cadogan, A.; Gao, Z.; Lewenstam, A.; Ivaska, A.; Diamond, D. Anal. Chem. 1992, 64, 2496−2501. (13) Gyurcsányi, R. E.; Rangisetty, N.; Clifton, S.; Pendley, B. D.; Lindner, E. Talanta 2004, 63, 89−99. (14) Michalska, A. J.; Appaih-Kusi, C.; Heng, L. Y.; Walkiewicz, S.; Hall, E. A. H. Anal. Chem. 2004, 76, 2031−2039. (15) Bobacka, J. Anal. Chem. 1999, 71, 4932−4937. (16) Gorelov, I. P.; Ryasenskii, S. S.; Kartamyshev, S. V.; Fedorova, M. V. J. Anal. Chem. 2005, 60, 65−69. (17) Vázquez, M.; Danielsson, P.; Bobacka, J.; Lewenstam, A.; Ivaska, A. Sens. Actuators, B 2004, 97, 182−189. (18) Sutter, J.; Radu, A.; Peper, S.; Bakker, E.; Pretsch, E. Anal. Chim. Acta 2004, 523, 53−59. (19) Lindfors, T.; Ivaska, A. Anal. Chem. 2004, 76, 4387−4394. (20) Han, W.-S.; Chung, K.-C.; Kim, M.-H.; Ko, H.-B.; Lee, Y.-H.; Hong, T.-K. Anal. Sci. 2004, 20, 1419−1422. (21) Paciorek, R.; van der Wal, P. D.; de Rooij, N. F.; Maj-Ż urawska, M. Electroanalysis 2003, 15, 1314−1318. (22) Michalska, A.; Konopka, A.; Maj-Zurawska, M. Anal. Chem. 2003, 75, 141−144. (23) Michalska, A.; Dumańska, J.; Maksymiuk, K. Anal. Chem. 2003, 75, 4964−4974. (24) Pandey, P. C.; Singh, G.; Srivastava, P. K. Electroanalysis 2002, 14, 427−432. 8692

dx.doi.org/10.1021/ac501625z | Anal. Chem. 2014, 86, 8687−8692