Spectroelectrochemical Sensing Based on Multimode Selectivity

Nov 5, 2009 - Figure 1. Spectroelectrochemical sensor consisting of an optically ..... For an ideal situation, the ΔEp should be zero, and the fwhm s...
0 downloads 0 Views 634KB Size
Download PDF
Anal. Chem. 2009, 81, 9599–9606

Spectroelectrochemical Sensing Based on Multimode Selectivity Simultaneously Achievable in a Single Device. 21. Selective Chemical Sensing Using Sulfonated Polystyrene-blockpoly(ethylene-ran-butylene)block-polystyrene Thin Films Sara E. Andria, Carl J. Seliskar, and William R. Heineman* Department of Chemistry, University of Cincinnati, 301 Clifton Court, Cincinnati, Ohio 45221-0172 Spectroelectrochemical sensors combine three modes of selectivity in a single device (electrochemistry, spectroscopy, and selective partitioning). A thin polymer film is coated onto the sensing platform in order to facilitate chemically selective transport to the electrode. The film is an essential part of the sensor because it provides an increase in selectivity and sensitivity by selectively preconcentrating the analyte. Here, we report the next step in the characterization of partially sulfonated polystyreneblock-poly(ethylene-ran-butylene)block-polystyrene(SSEBS) films for the purpose of chemical sensing by examining the selectivity of the sensor fabricated with this novel thin film material. Binary mixtures using model analytes were used to demonstrate the sensor’s ability to detect an analyte in the presence of a direct interference. The binary mixtures consisted of Ru(bpy)32+/Fe(CN)63-, Ru(bpy)32+/ Fe(bpy)32+, and Ru(bpy)32+/Cu(bpy)22+. Demonstration of the selective partitioning mode using the Ru(bpy)32+/Fe(CN)63- mixture and absorption detection showed the SSEBS film’s preference for Ru(bpy)32+ over Fe(CN)63-, and therefore, Fe(CN)63- did not interfere with the sensor’s response to Ru(bpy)32+. Furthermore, the importance of the use of three modes together was demonstrated by analysis of the Ru(bpy)32+/Fe(bpy)32+ and the Ru(bpy)32+/Cu(bpy)22+ test mixtures, where both selection of a specific wavelength for absorption and selection of a specific potential window were required to reduce or eliminate the signal from the interference. Finally, analysis of the Ru(bpy)32+/Fe(bpy)32+ test mixture was also demonstrated using fluorescence detection. Spectroelectrochemical sensors are unique because they employ three modes of selectivity:1 selective partitioning (via a chemically selective film), spectroscopy, and electrochemistry. In order for the analyte to be detected, it must (1) partition into the film, (2) be electrochemically oxidized or reduced at a particular * To whom correspondence should be addressed. (1) Shi, Y.; Slaterbeck, A. F.; Seliskar, C. J.; Heineman, W. R. Anal. Chem. 1997, 69, 3679–3686. 10.1021/ac901595b CCC: $40.75  2009 American Chemical Society Published on Web 11/05/2009

Figure 1. Spectroelectrochemical sensor consisting of an optically transparent electrode (glass coated with indium tin oxide) modified with an ion-exchange film. The pentagon symbol represents the analyte in the sample because it meets the requirements for detection: (1) it partitions into the film, (2) it has the appropriate electrochemistry, and (3) it absorbs light at the selected wavelength in one oxidation state but not in the other. The interferences in the sample are excluded from the sensor response because they do not have the appropriate charge (circle), optical properties (hexagon), or electrochemical properties (star).

potential, and (3) absorb or emit light in either its oxidized or its reduced form at the observed wavelength. The spectroelectrochemical sensor integrates the modes of selectivity at an optically transparent electrode (OTE), which consists of a glass substrate coated with indium-tin oxide (ITO). This platform simultaneously serves as the working electrode, the optical waveguide, and the solid support for the selective film, which is illustrated in Figure 1. The three modes work together to exclude interferences based on their ionic, electrochemical, and/or optical properties. The first mode, the chemically selective film, is typically an ion-exchange polymer. The advantage of the selective film is 2-fold: potential interferences of opposite charge are excluded or preferentially repelled and sensitivity of the sensor is improved through preconcentration. Unfortunately, finding materials that can be used as a selective film is a difficult task because it must meet several requirements.2-6 Recently, we have demonstrated that the cation exchange polymer, sulfonated polystyrene-block(2) Dasenbrock, C. O.; Ridgway, T. H.; Seliskar, C. J.; Heineman, W. R. Electrochim. Acta 1998, 43, 3497–3502. (3) Gao, L.; Seliskar, C. J. Chem. Mater. 1998, 10, 2481–2489. (4) Gao, L.; Seliskar, C. J.; Milstein, L. Appl. Spectrosc. 1997, 51, 1745–1752. (5) Hu, Z.; Seliskar, C. J.; Heineman, W. R. Anal. Chem. 1998, 70, 5230–5236. (6) Paddock, J. R.; Zudans, I.; Heineman, W. R.; Seliskar, C. J. Appl. Spectrosc. 2004, 58, 608–612.

Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

9599

poly(ethylene-ran-butylene)-block-polystyrene (SSEBS), can be used for this purpose.7 SSEBS is a triblock polymer that consists of a polystyrene unit, a sulfonated polystyrene unit, and an ethylene/butylene unit. When exposed to a sample, the sulfonated polystyrene groups exchange their counterions for cations and prevent the partitioning of anions, which is illustrated by the oval symbol in Figure 1. SSEBS was easily able to form thin, uniform films by spin-coating. The sulfonated polystyrene groups allowed for the preconcentration of positively charged chromophores. The SSEBS film proved to not only be electrochemically inactive but also did not interfere with either the electrochemistry of the preconcentrated chromophores or the optical measurements. In addition, the film proved to be both rugged and stable. The second mode of selectivity, the spectroscopy mode, exists on two levels: choice of the optical detection method and choice of the wavelength at which the optical response is monitored. The optical response of the sensor is measured using attenuated total reflectance spectroscopy. Total reflectance is achieved when the light source is propagated into the glass substrate at an angle greater or equal to the critical angle. At each of the reflection points, an electromagnetic component of the light source, known as the evanescent wave, penetrates past the ITO layer into the film. Currently, the spectroelectrochemical sensor can be operated using absorbance or fluorescence to detect the analyte. When using the sensor in absorbance mode, the evanescent wave can be absorbed by species that have partitioned into the film; however, in fluorescence mode, the evanescent wave can excite the fluorescence of these same species. Using the fluorescence-based sensor is preferable because this technique is generally much more sensitive compared to its absorbance-based counterpart. Furthermore, fluorescence offers more selectivity compared to absorbance because (a) two wavelengths are required for the analyte’s detection (i.e., excitation and emission) and (b) very few analytes fluoresce. Unfortunately, this means limited applications for the fluorescence-based sensor. Absorbance or fluorescence is monitored at a wavelength that is specific to the analyte. Interfering species will be excluded from the sensor response if they do not absorb or fluoresce at the selected wavelength, which is illustrated by the hexagon symbol in Figure 1. The third mode of selectivity is provided by the choice of the applied potential window. If there are interfering species in the sample that both partition into the film and respond optically at the same wavelength(s) appropriate for the detection of the analyte, the sensor can exclude these interferences based on their redox potentials. The potential applied to the working electrode can be scanned in a window that is specific to the analyte, thus excluding interferences that do not have the appropriate electrochemistry or do not produce a change in optical signal when electrochemically cycled, as illustrated by the star symbol in Figure 1. In an ideal situation (pentagon symbol, Figure 1), the resulting sensor response is a change in optical signal (∆A or ∆F), which corresponds to the concentration of analyte in the film and is used to calculate the concentration of the analyte in the sample. (7) Pantelic, N.; Andria, S. E.; Heineman, W. R.; Seliskar, C. J. Anal. Chem. 2009, 81, 6756–6764.

9600

Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

Many of the previous studies from our group have involved demonstration of the spectroelectrochemical sensor using relatively pure solutions that contain only the analyte of interest.8-13 Only three studies have examined the performance of the spectroelectrochemical sensor using complex samples. In the study by Shi et al., binary mixtures of metal complexes were used to investigate the performance and selectivity of PDMDAAC-SiO2, where PDMDAAC is poly(dimethyldiallylammonium chloride) and Nafion-SiO2 films with an absorbance-based sensor.14 In an effort to further demonstrate the abilities of this device, we wanted to evaluate the performance and selectivity of the spectroelectrochemical sensor using the new SSEBS material. Both the absorbance- and fluorescence-based sensors were employed to detect an analyte in the presence of either a direct optical or an electrochemical interference or in the presence of species of differing charges. Although the absorbance-based sensor has been previously examined with other films, there have been no studies reporting the performance of the fluorescence-based sensor using complex samples. In this study, binary test mixtures were analyzed using a spectroelectrochemical sensor fabricated with a 24 or 50 nm thick SSEBS film on an OTE made of ITO. Each test mixture consisted of two metal complexes, where one complex was arbitrarily designated as the analyte and the other was designated as the interference. The metal complexes used were Ru(bpy)32+, Fe(bpy)32+, Cu(bpy)22+, and Fe(CN)63-, which provided electrochemistry over a range of reduction potentials and varying spectroscopic properties. Three different test mixtures were used to demonstrate how the modes work to exclude or reduce interferences from the sensor response and allow for the detection of the analyte of interest. EXPERIMENTAL SECTION Chemical Reagents. The following reagents were used tris(2,2′-bipyridyl) ruthenium(II) chloride hexahydrate (Ru(bpy)3Cl2 · 6H2O, Aldrich), potassium ferricyanide (K3Fe(CN)6, Fisher), cupric sulfate (CuSO4, Mallinckrodt), ferrous ammonium sulfate (Fe(NH4)2(SO4)2 · 6H2O, Mallinckrodt), 2,2′-dipyridyl (bpy, Aldrich), potassium nitrate (Fisher), and polystyreneblock-poly(ethylene-ran-butlyene)-block-polystyrene (sulfonated, 5% solution in 1-propanol and dichloroethane, Aldrich). All reagents were used without further purification. A 0.1 M KNO3 supporting electrolyte solution was prepared with deionized water from a Barnstead water purification system. Ru(bpy)32+ and Fe(CN)63- stock solutions were prepared by dissolving the appropriate amounts in the supporting electrolyte solution. Fe(bpy)32+ and Cu(bpy)22+ stock solutions were prepared by (8) DiVirgilio-Thomas, J. M.; Heineman, W. R.; Seliskar, C. J. Anal. Chem. 2000, 72, 3461–3467. (9) Richardson, J. N.; Dyer, A. L.; Stegemiller, M. L.; Zudans, I.; Seliskar, C. J.; Heineman, W. R. Anal. Chem. 2002, 74, 3330–3335. (10) Kaval, N.; Seliskar, C. J.; Heineman, W. R. Anal. Chem. 2003, 75, 6334– 6340. (11) Andria, S. E.; Richardson, J. N.; Kaval, N.; Zudans, I.; Seliskar, C. J.; Heineman, W. R. Anal. Chem. 2004, 76, 3139–3144. (12) Maizels, M.; Seliskar, C. J.; Heineman, W. R. Electroanalysis 2000, 12, 1356–1362. (13) Wansapura, C. M.; Seliskar, C. J.; Heineman, W. R. Anal. Chem. 2007, 79, 5594–5600. (14) Shi, Y.; Seliskar, C. J.; Heineman, W. R. Anal. Chem. 1997, 69, 4819– 4827.

Table 1. Recipes for Preparation of SSEBS Film of Varying Film Thickness film thickness (nm)

% vol SSEBS

spin rate (rpm)

24 50 103 340

1 1 2 5

6000 3000 3000 3000

dissolving the appropriate amount of Fe(NH4)2(SO4)2 · 6H2O or CuSO4 with a 4-fold excess of ligand in the supporting electrolyte solution. The test mixtures were prepared by mixing the appropriate volume from each metal complex stock solution in the supporting electrolyte solution. Indium tin oxide (ITO) coated glass slides (Corning 1737F glass, 11-50 Ω/sq, 135 nm thick film on 1.1 mm glass, Thin Film Devices, Anaheim, CA) with the dimensions 10 × 40 mm were used as OTEs. Preparation of SSEBS Films on ITO Slides. The ITO slides were first scrubbed with Neutrad glass cleaner (Decon Laboratories, Inc.) and rinsed thoroughly with water. The cleaned slides were rinsed with ethanol and then deionized water and allowed to dry prior to use. The SSEBS films were prepared by the use of a Model 1-PM101DT-R485 spin coater from Headway Research, Inc. A 100 µL aliquot of the SSEBS solution was deposited directly onto the slide and then spun at the appropriate rate. Table 1 gives the SSEBS solution concentrations and spin rates required to achieve the different film thicknesses. These film thickness values were determined using spectroscopic ellipsometry, where only one measurement was completed for each thickness. The experimentally measured data were fitted to theoretical models, and the errors for these fittings, expressed as mean squared errors (MSE), ranged from 1 to 3.7 The 1% and 2% SSEBS solutions were diluted from a 5% stock solution with the appropriate amount of 2-propanol. The spin-coated slides were stored in air until use. Spectroelectrochemical Measurements. A detailed description of the experimental setup has been described previously.10 Briefly, incident radiation was directed through a multimode optical fiber (Romack, 400 µm core step index, NA ) 0.22) to a collimating objective (Newport, M-10×, NA ) 0.25). For the absorbance measurements, a xenon arc lamp (model ILC302UV, ILC Technology, Inc.) was used as the light source. The collimated light was then passed in and out of the ATR cell by way of two Schott SF6 coupling prisms (Karl Lambrecht). The prisms were attached to the rear of the slide by a high refractive index mounting compound (Cargille Meltmount, n ) 1.704). The attenuated light was then focused by a microscope objective lens (Newport, 5×, NA ) 0.12) into another optical fiber, which directed the light into a monochromator (SpectraPro 300i, Acton Research Corp., 0.3 m focal length) outfitted with a photon counting photomultiplier tube (PMT). For the fluorescence measurements, the experimental arrangement was similar except for the type of light source and the position of the collection fiber. A 405 nm laser (model LDCU12/6228 Power Technology Inc.) was used as the excitation source. The emission was collected with an optical fiber located perpendicular to the backside of the OTE at the first reflection point. Data collection was achieved using Acton Research NCL monochromator controller electronics and Spectrasense software.

The flow cell used for the ATR measurements was slightly different from that previously described. The body of the flow cell was made of Delrin and had a volume of ∼0.4 mL. A Pt wire was affixed inside of the cell, which served as the auxiliary electrode. The solutions were pumped through the sensor cell at 0.2 mL/ min using a syringe pump (model NE-1000, New Era Pump Systems, Inc.). For all experiments, the 0.1 M KNO3 supporting electrolyte solution was pumped through the sensor cell for 10 min. Optical measurements commenced while the supporting electrolyte was still pumping through the cell to obtain a baseline. The test mixtures began pumping through the sensor cell at 250 s. The electrochemical modulations and cyclic voltammograms were acquired simultaneously using a potentiostat (Epsilon, BAS) to scan the applied potential at 5 mV/s. The three-electrode cell assembly included a miniature Ag/ AgCl, 3 M KCl reference electrode (Cypress Systems), a Pt wire auxiliary electrode, and the SSEBS-modified ITO working electrode. All data analysis and manipulations were carried out using commercial spreadsheet and graphics algorithms. Optical and Electrochemical Measurements of Each Metal Complex. A visible spectrum and a cyclic voltammogram were acquired for each complex used in this study. An appropriate amount of each metal complex was dissolved in the 0.1 M KNO3 supporting electrolyte solution giving a 0.1 mM metal solution. The visible spectra were obtained using a spectrophotometer (Cary 50, Varian) for all solutions except for the Cu(bpy)22+. For the Cu(bpy)22+ solution, the spectroelectrochemical cell was used to acquire the spectrum of the reduced form (Cu(bpy)21+) under ATR conditions at the ITO electrode, because it is the reduced form of the copper complex that absorbs at the desired wavelength (400 nm). The optical configuration used for this measurement is described in the previous section. Cyclic voltammograms were acquired using a standard threeelectrode cell under quiescent conditions. The cell configuration included a bare ITO working electrode, a Ag/AgCl reference electrode, and a Pt mesh auxiliary electrode. The potential was scanned in the appropriate windows at 0.05 V/s. Because the Cu(bpy)22+/1+ couple is located in a relatively negative potential window, this solution was deoxygenated for ∼20 min with nitrogen prior to running the electrochemical experiment. RESULTS AND DISCUSSION Selective Partitioning Mode. The SSEBS film should serve to preferentially repel anions present in a sample that could interfere with the sensor’s response to the analyte. To demonstrate this, a test mixture consisting of Ru(bpy)32+ (designated the analyte) and Fe(CN)63- (designated the interference) was used. Since these species have opposite net charges, but absorb in the same region of the visible spectrum, selective partitioning provided by the SSEBS film should prohibit a sensor response from the Fe(CN)63-, while providing an enhanced response for Ru(bpy)32+. The molar absorptivity of Fe(CN)63- (ε ) 1040 M-1 cm-1 at 420 nm15) is small compared to Ru(bpy)32+ (ε ) 14 600 M-1 cm-1 at 454 nm16), therefore, adjustments were made (15) Waltz, W. L.; Akhtar, S. S.; Eager, R. L. Can. J. Chem. 1973, 51, 2525– 2529. (16) Fabian, R. H.; Classen, D. M.; Sonntag, R. W. Inorg. Chem. 1980, 19, 1977.

Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

9601

Figure 2. Visible spectra (A) and cyclic voltammograms (B) for each of the metal complexes: Ru(bpy)32+ (s), Fe(bpy)32+ (---), Cu(bpy)22+ (s s s), and Fe(CN)63- ( · · · ). The concentration of each metal complex solution was 0.1 mM in 0.1 M KNO3 supporting electrolyte. The cyclic voltammograms were obtained at a bare ITO working electrode while employing a Ag/AgCl reference electrode and Pt mesh auxiliary electrode. Scan rate ) 50 mV/s. Table 2. Electrochemical and Optical Properties of the Metal Complexes metal complex 2+

Ru(bpy)3 Fe(bpy)32+ Fe(CN)63Cu(bpy)22+

E0′ (V, bare ITO vs Ag/AgCl)

λmax (nm)

1.08 0.89 0.20 -0.11

450 520 420 400

in order to make Fe(CN)63- a more potent interference. Specifically, a 10-fold higher concentration of Fe(CN)63- than Ru(bpy)32+ was used, and the optical response was monitored at 420 nm, which is the λmax for Fe(CN)63-. Figure 2 shows an absorbance spectrum (panel A) and cyclic voltammogram (panel B) for each complex. In addition, Table 2 provides the optical and electrochemical information. Figure 3A shows the sensor response to the test mixture consisting of 0.1 mM Ru(bpy)32+ and 1 mM Fe(CN)63- when using 24, 103, and 340 nm SSEBS films (numbered 1, 2, and 3, respectively). The initial baseline seen corresponds to only supporting electrolyte exposed to the sensor. Exposing the test mixture to the sensor causes an immediate increase in the absorbance due to the partitioning of Ru(bpy)32+ into the SSEBS film by ion exchange at open circuit. The absorbance versus time response reaches a plateau when equilibrium has been reached between the Ru(bpy)32+ concentration in the solution and Ru(bpy)32+ concentration in the film. The amount of Ru(bpy)32+ able to partition into the film is related to the thickness of the film; thicker films possess more ion-exchange sites for the Ru(bpy)32+ to occupy than do the thinner films. 9602

Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

Figure 3. (A) Sensor response to a mixture of 0.1 mM Ru(bpy)32+ and 1 mM Fe(CN)63- in 0.1 M KNO3 for three different film thicknesses: 24 nm (1), 103 nm (2), and 340 nm (3). The potential versus time plot above the absorbance data shows how the potential was scanned to achieve the optical modulations. (Note: the potential-time trace does not correctly correlate to all of the absorbance traces as the electrochemical excitation step was started at different times for each film thickness.) All modulations were obtained by scanning the potential at 5 mV/s. (B) One cyclic voltammogram that was acquired simultaneously with the first set of optical modulations in the above figure. The potential was scanned from 0.7 to 1.3 to -0.3 to 0.7 V at 5 mV/s.

For this reason, the absorbance magnitude is greater with the thicker films. The optical modulations in Figure 3A result from scanning the potential in three different potential windows. The electrochemical excitation used to produce each set of optical modulations is shown above the absorbance vs time plot. The modulations of the different films do not occur at the same time because the electrochemical excitation was started at different times for each experiment. The first set of optical modulations, labeled (a), is the result of scanning the potential positively from 0.7 to 1.3 V, then negatively to -0.3 V, and then back to 0.7 V. The change in absorbance only occurs as the potential is scanned between 0.7 and 1.3 V. These modulations correspond to the electrolysis of only Ru(bpy)32+. This observation was verified by cycling each complex separately. For the optical modulations labeled (b), the potential was scanned between 0.7 and 1.3 V. In this potential window, the electrolysis of Ru(bpy)32+ occurs, and

the resulting modulations are the same magnitude as those in (a). The modulations labeled (a) and (b) are sharper at the bottom compared to when they return to the original absorbance value. This is due to the range in which the potential was scanned relative to the E0′ of the Ru(bpy)32+ couple. For the modulations (a), the potential window was between -0.3 and 1.3 V. The E0′ for Ru(bpy)32+ is 1.06 V, therefore, the colored, reduced form is present for a much longer time compared to the colorless, oxidized form. The same is true for the (b) modulations because the E0′ of the Ru(bpy)32+ couple is still located relatively positive in the window used (0.7 to 1.3 V). The modulations in (b) are also spaced closer together compared to those in (a), which is due to the potential window for (b) being smaller compared to that for (a). For the optical response labeled (c), the potential was scanned between 0.5 and -0.3 V. This results in the electrolysis of the Fe(CN)63(vide infra); however, no optical modulations were observed, which demonstrates the optical/film selectivity against Fe(CN)63-. As the change in optical signal was produced, the electrochemical response was simultaneously acquired. One cyclic voltammogram for each film thickness is shown in Figure 3B. Like the optical response, the current response for the Ru(bpy)32+ increases with the film thickness because more Ru(bpy)32+ was available in the thicker films for electrolysis. The voltammograms are sharp, symmetrical, and nearly return to baseline. The peak separation (∆Ep) and the full width at half-maximum (fwhm) were calculated for the Ru(bpy)32+ couple for all film thicknesses. The ∆Ep ranged from 24 to 28 mV, and the fwhm ranged from 117 to 134 mV. For an ideal situation, the ∆Ep should be zero, and the fwhm should be 90.6/n mV; however, experimentally these values are rarely seen.17,18 It is more typical to see ∆Ep values