Potentiometric Online Detection of Aromatic Hydrocarbons in Aqueous

Potentiometric Online Detection of Aromatic Hydrocarbons in Aqueous Phase Using Carbon Nanotube-Based Sensors. Alemayehu P. Washe, Santiago Macho, ...
2 downloads 3 Views 1MB Size
Download PDF
Anal. Chem. 2010, 82, 8106–8112

Potentiometric Online Detection of Aromatic Hydrocarbons in Aqueous Phase Using Carbon Nanotube-Based Sensors Alemayehu P. Washe, Santiago Macho, Gasto´n A. Crespo, and F. Xavier Rius* Department of Analytical and Organic Chemistry, Universitat Rovira i Virgili, 43007 Tarragona, Spain Surfaces made of entangled networks of single-walled carbon nanotubes (SWCNTs) display a strong adsorption affinity for aromatic hydrocarbons. Adsorption of these compounds onto the walls of SWCNTs changes the electrical characteristics of the SWCNT-solution interface. Using these features, we have developed a potentiometric sensor to detect neutral aromatic species. Specifically, we can detect online aromatic hydrocarbons in industrial coolant water. Our chromatographic results confirm the adsorption of toluene onto the walls of carbon nanotubes, and our impedance spectroscopy data show the change in the double layer capacitance of the carbon nanotube-solution interface upon addition of toluene, thus confirming the proposed sensing mechanism. The sensor showed a toluene concentration dependent EMF response that follows the shape of an adsorption isotherm and displayed an immediate response to the presence of toluene with a detection limit of 2.1 ppm. The sensor does not respond to other nonaromatic hydrocarbons that may coexist with aromatic hydrocarbons in water. It shows a qualitative sensitivity and selectivity of 100% and 83%, respectively, which confirms its ability to detect aromatic hydrocarbons in aqueous solutions. The sensor showed an excellent ability to immediately detect the presence of toluene in actual coolant water. Its operational characteristics, including its fast response, low cost, portability, and easy use in online industrial applications, improve those of current chromatographic or spectroscopic techniques. In addition to the great advances in ion-selective electrodes,1,2 new potentiometric sensors extended recent detection possibilities to complex organic and biological analytes.3,4 These new sensors have incorporated original recognition layers linked to different transduction elements involving new sensing mechanisms. Introducing nanostructured materials as transducers facilitates the * To whom correspondence should be addressed. Address: Department of Analytical and Organic Chemistry, Universitat Rovira i Virgili, Campus Sescelades, Marcel · lı´ Domingo s/n, 43007 Tarragona, Spain. E-mail: [email protected]. Tel: +34 977 559562. Fax: +34 977 558446. (1) Bakker, E.; Pretsch, E. Angew. Chem., Int. Ed. 2007, 46, 5660–5668. (2) Bakker, E.; Pretsch, E. Trends Anal. Chem. 2008, 27, 612–618. (3) Zhou, Y.; Yu, B.; Guiseppi-Elie, A.; Sergeyev, V.; Levon, K. Biosens. Bioelectron. 2009, 24, 3275–3280. (4) Zelada, G. A.; Riu, J.; Du ¨ zgu ¨ n, A.; Rius, F. X. Angew. Chem., Int. Ed. 2009, 48, 7334–7337.

8106

Analytical Chemistry, Vol. 82, No. 19, October 1, 2010

solid-state signal transduction and confers a high stability to the recorded signal.5,6 Among these nanostructured materials, singlewalled carbon nanotubes (SWCNTs), which have several outstanding properties, have been widely studied as a key element in electrochemical sensors.7 Single-walled carbon nanotubes (SWCNTs), which possess sp2 carbon atoms, are tubular structures with diameters of a few nanometers and lengths of up to millimeters. They have a large specific surface area and an electronic distribution that provides plenty of sites for interaction with adsorbing species. In particular, their electrical properties, derived from π electrons on the outer surface of the one-dimensional structure, make them highly sensitive to changes in their chemical environment. This ability to change their conductivity due to neighboring small chemical events involving a charge transfer with a wide range of species in their vicinity makes them suitable for many sensing applications. SWCNTs have been used in biosensors based on field-effect transistors8,9 where the nanotubes interact directly with the analyte molecules to detect molecules both in gas phase and in solution. The change in electrical properties of the nanotubes upon direct charge transfer of the target molecules to or from the nanotubes or due to chemical interaction with the suitable receptor previously linked to the nanotubes has been used as a sensing mechanism. SWCNTs have also been used as solid contact transducers in ionselective electrodes.10 Recently, it has been reported that, when a SWCNT-modified electrode is brought into direct contact with the ionic solution, the electrical double layer capacitance is revealed as the essential component of the ion-to-electron transduction in this type of ISE.11 The high double layer capacitance observed is due to the large specific surface area of the SWCNT and the porosity of their network.12,13 (5) Lai, C.-Z.; Fierke, M A.; Stein, A.; Buhlmann, P. Anal. Chem. 2007, 79, 4621–4626. (6) Fouskaki, M.; Chaniotakis, N. Analyst 2008, 133, 1072–1075. (7) Avouris, P.; Radosavljevic´, M.; Wind, S. J. Carbon Nanotube Electronics and Optoelectronics in Applied Physics of Carbon Nanotubes; Springer: New York, 2005, p 227-251. (8) Star, A.; Kauffman, D. R. Angew. Chem., Int. Ed. 2008, 47, 6550–6570. (9) Gruner, G. Anal. Bioanal. Chem. 2006, 384, 322–335. (10) Crespo, G. A.; Macho, S.; Rius, F. X. Anal. Chem. 2008, 80, 1316–1322. (11) Crespo, G. A.; Macho, S.; Bobacka, J.; Rius, F. X. Anal. Chem. 2009, 81, 676–681. (12) An, K. H.; Kim, W. S.; Park, Y. S.; Choi, Y. C.; Lee, S. M.; Chung, D. C.; Bae, D. J.; Lim, S. C.; Lee, Y. H. Adv. Mater. 2001, 13, 497–450. (13) Fierke, M. A.; Lai, C. Z.; Buhlmann, P.; Stein, A. Anal. Chem. 2010, 82, 680–688. 10.1021/ac101146k  2010 American Chemical Society Published on Web 09/01/2010

Numerous theoretical and experimental studies suggest that SWCNTs strongly adsorb aromatic compounds via hydrophobic and, most importantly, π-π stacking interaction.14-17 This physisorption of organic matter onto carbon nanotubes has been used in solid phase extraction, water treatment,18,19 and the development of electrochemical sensors where the receptors are linked to the carbon nanotubes by noncovalent functionalization.20 Therefore, whenever aromatic hydrocarbons are introduced into a solution containing a constant electrolyte concentration, at electrochemical equilibrium with the SWCNT-based sensor, the hydrocarbons undergo adsorptive interaction with the SWCNTs on the surface of the sensor. This adsorptive interaction is facilitated by, among other effects, the high electronic polarizability of the SWCNT sidewalls.21,22 Much theoretical and experimental evidence suggests that π-π electron donor-acceptor interaction occurs between aromatic hydrocarbons and carbon nanotubes.23-25 It is, therefore, expected that the adsorption of aromatic hydrocarbons on SWCNTs on the surface of the sensor could alter the electrical double layer and modify the associated capacitance, thus leading to a change in the electrical potential. In this paper, we envisioned using the change in electrical potential as a consequence of interaction between the SWCNTbased sensor and the aromatic hydrocarbons as a basis for detecting the latter in low-conductivity coolant water. In refineries, heat exchangers are very common. Contamination of the coolant water occurs when process fluids (hydrocarbons) leak through small pores in the heat exchanger. The hydrocarbons are a serious threat to the heat transferring capacity of the water, ultimately resulting in a decrease of the condensing ability of the system. Moreover, as it is difficult to treat large volumes of contaminated cooling water and/or recover the hydrocarbons, early detection of a small leak of hydrocarbons in the coolant water is important. The techniques used so far to detect hydrocarbons in aqueous phase are usually based on gas chromatography (GC) or highperformance liquid chromatography (HPLC). However, these techniques need solid phase extraction for the preconcentration of the hydrocarbons from contaminated water, time-consuming steps, or sophisticated instrumentation. These limitations have deterred GC- and HPLC-based methods from being used for online applications. Flame ionization analyzers26 and fluorescence have been used to detect hydrocarbons in industrial coolant water. Although effective in operation, such technologies require expensive instrumentation. Optical fiber chemical sensors have also (14) Gotovac, S.; Honda, H.; Hattori, Y.; Takahashi, K.; Kanoh, H.; Kaneko, K. Nano Lett. 2007, 7, 583–587. (15) Chen, W.; Duan, L.; Zhu, D. Environ. Sci. Technol. 2007, 41, 8295–8300. (16) Yang, K.; Zhu, L.; Xing, B. Environ. Sci. Technol. 2006, 40, 1855–1861. (17) Pan, B.; Xing, B. Environ. Sci. Technol. 2008, 42, 9005–9013. (18) Li, Q.; Mahendra, S.; Lyon, D. Y.; Brunet, L.; Liga, M. V.; Li, D.; Alvarez, P. J. J. Water Res. 2008, 42, 4591–4602. (19) Valca´rcel, M.; Ca´rdenas, S.; Simonet, B. M.; Moliner-Martı´nez, Y.; Lucena, R. Trends Anal. Chem. 2008, 27, 34–43. (20) Zhao, Y.; Stoddart, J. F. Acc. Chem. Res. 2009, 42, 1161–1171. (21) Ma, R. Z.; Liang, J.; Wei, B. Q.; Zhang, B.; Xu, C. L.; Wu, D. H. J. Power Sources 1999, 84, 126–129. (22) Chen, J.; Chen, W.; Zhu, D. Environ. Sci. Technol. 2008, 42, 7225–7230. (23) Star, A.; Han, T.-R.; Gabriel, J.-C. P.; Bradley, K.; Grulner, G. Nano Lett. 2003, 3, 1421–1423. (24) Silva, L. I. B.; Ferreira, F. D. P.; Rocha-Santos, T. A. P.; Duarte, A. C. J. Chromatogr., A 2009, 1216, 6517–6521. (25) Zhu, D.; Pignatello, J. J. Environ. Sci. Technol. 2005, 39, 2033–2041. (26) Re´-Poppi, N.; Almeida, F. F. P.; Cardoso, C. A. L.; Raposo, J. L.; Viana, L. H.; Silva, T. Q.; Souza, J. L. C.; Ferreira, V. S. Fuel 2009, 88, 418–423.

been proposed for determining aromatic hydrocarbons in contaminated waters.27,28 Their advantages include the possibility of miniaturization for field applications. These sensors usually employ a polymer-clad silica fiber, and their measurements are based on the evanescent wave principle. The disadvantage of this type of sensor, however, is related to the strong absorbance of the O-H group of water over C-H absorbance, which affects sensitivity. Recently reported was a carbon nanotube-based field-effect transistor (CNTFET) detector coupled to a gas chromatograph for detecting the aromatic compounds benzene, toluene, ethylbenzene, m-xylene, p-xylene, and o-xylene. Here, the compatibility of the carbon nanotubes with the well-developed microelectronic fabrication techniques is an advantage for the high degree of miniaturization.24 However, this type of CNTFET detector requires coupling to GC and is difficult for online applications. A simple, inexpensive sensor system for the online detection of aromatic hydrocarbons as soon as they appear in coolant water is, therefore, needed. In this paper, we report the development of a SWCNT-based potentiometric sensor for the online early detection of aromatic hydrocarbons in coolant water. The potentiometric detection of the uncharged hydrocarbons can take advantage of two features of SWCNTs: the high affinity between hydrocarbon and nanotubes and the change in the electrical double layer characteristics of carbon nanotubes when hydrocarbons are present in an aqueous solution of low constant ionic strength. Compared to existing methods, this sensor is more sensitive, faster, cheaper, and easier to use in online applications. Moreover, it is simple and can be easily miniaturized. MATERIALS AND METHODS Reagents. All the reagents, including toluene and cyclohexane, were analytical grade and used as purchased from Aldrich. SWCNTs (OD, 1 to 2 nm; length, 50 µm) with 90% purity were obtained from Carbolex Inc. All solutions were prepared using deionized water obtained with a Milli-Q PLUS (Millipore Corporation). Sensor Preparation. The working sensor was prepared by first preprocessing the surface of the distal end of a glassy carbon rod followed by deposition of the SWCNTs. The glassy carbon rod (Sigradur G., length; 50 mm, 3 mm L) was polished to a smooth surface first using a sheet of abrasive paper (Buehler Carbimet 600/P1200) and then to a shiny mirror in a slurry containing alumina of different sizes (25, 1, and 0.03 µm, Buehler). The resulting active surface was about 0.07 cm2. The SWCNTs were subjected to thermal treatment in air in a silica furnace chamber to selectively remove the amorphous carbon under the following conditions: T ) 365 °C, air flow rate ) 100 cm3/ min, and t ) 90 min. The SWCNTs were deposited by spraying 10 mL of aqueous dispersion containing 25 mg of the purified SWCNT (first dried and powdered with marble mill) and 100 mg of sodium dodecyl sulfate (SDS) onto the polished glassy carbon surface. Prior to the deposition, the dispersion was homogenized using a tip sonicator (amplitude 60%, cycle 0.5, Ultraschallprocessor UP200S, Dr. Hielscher) for 30 min. The (27) McCue, R. P.; Walsh, J. E.; Walsh, F.; Regan, F. Sens. Actuators, B 2006, 114, 438–444. (28) Wolfbeis, O. S. Anal. Chem. 2008, 80, 4269–4283.

Analytical Chemistry, Vol. 82, No. 19, October 1, 2010

8107

deposition of SWCNT was achieved in successive steps. After spraying the dispersion of nanotubes twice (for 2 s) on heated surface, the layer of the SWCNT was dried by heating with a high temperature air blower (approximately 200 °C). This step was repeated 3 times. After 2 min of air cooling, the sprayed surface of the sensor was then thoroughly rinsed with water to get rid of SDS, air cooled for 3 min, and dried again with the hot air blower. These processes were performed in six rounds to obtain a homogeneous coverage of entangled SWCNTs on the surface of the sensor. The sensor was then kept in a silica furnace chamber in order to selectively remove residual SDS under the following conditions previously optimized by thermogravimetric analysis: T ) 280 °C; air flow rate ) 100 cm3/min for 1 h. Finally, it was inserted into a Teflon body (length, 35 mm; outer diameter, 6 mm) and used for measurement. Sample Preparation. A 10-5 M NaCl solution was used as a matrix for the analyte and for blank measurements. Toluene was used as a representative analyte of aromatic hydrocarbons. The solubility of toluene in water is around 0.47 g/L (20-25 °C). Using this data, we prepared an aqueous solution of 300 ppm of toluene in 10-5 M NaCl solution as an initial test solution. Measurements were taken by transferring the required volume of this initial solution to the matrix solution (10-5 M NaCl) and recording the change in EMF in two ways: (i) by increasing the additions of toluene to the same test solution (starting with 1 ppm and reaching up to 130 ppm) and (ii) using different test solutions each containing a different concentration of toluene. Thermogravimetric Analysis. Thermogravimetric analysis on two sets of carbon nanotube specimens sampled from the same batch was performed with Model TGA/SDTA851 (METTLER/ TOLEDO) series instrument in O2 and N2 atmosphere at 30-900 °C. One set was the SWCNTs purified by air oxidation. The second sample was oxidized using the above procedure, then dispersed in a water SDS mixture, stirred under sonication, filtered, and washed thoroughly to remove SDS and replicate the spraying process. Potentiometric Measurements. Potentiometric measurements were taken using a Keithley 6514 potentiometer. The potentiometric system consisted of a surface-modified glassy carbon sensor as a working sensor and a double junction Ag/ AgCl reference electrode with saturated KCl (3 M) as inner filling solution and a 1 M LiOAc as a bridge electrolyte. EMF was recorded under stirring conditions at room temperature (20-23 °C). Toluene and cyclohexane were selected as model analytes for aromatic and aliphatic hydrocarbons that could exist in the coolant water and as a source of useful mechanistic information. All the potentiometric measurements involved a prior conditioning of the sensor in the background solution overnight. Electrochemical Impedance Spectroscopy (EIS). Electrochemical impedance measurements were taken using a onecompartment, three-electrode electrochemical cell. The glassy carbon/SWCNTs (3 mm L, area 0.07 cm2) acted as the working electrode, and the auxiliary electrode was a glassy carbon rod. The reference electrode was an Ag/AgCl/KCl (3M) single junction electrode (Model 6.0733.100, Metrohm). All dc potentials (Edc) were referred to this reference 8108

Analytical Chemistry, Vol. 82, No. 19, October 1, 2010

electrode. A general purpose electrochemical system, the Autolab PGSTAT128N (Ecochemie) with frequency response analyzer (FRA2), was used. The impedance spectra were recorded in the 100 kHz to 0.01 Hz frequency range. The amplitude for the sinusoidal excitation signal was 10 mV, and the temperature was 20-23 °C. The spectra were then fitted to an equivalent electrical circuit using the Autolab impedance analysis software. For practical purposes, 10-2 M NaCl solution was used as the background electrolyte in EIS measurements. The capacitive and resistive responses were analyzed to gain insight into the mechanism. GC/MS Analysis. Analysis of the toluene adsorbed onto the surface walls of the SWCNTs was conducted by GC after the thermal extraction step. Toluene was extracted by heating the sensor (with the Teflon removed after recording the EMF) in a sealed vial at 150 °C for 10 min. Then, 500 µL of headspace was injected in a 10:1 split mode into an Agilent HP 6890 Series GC coupled to an HP 5973 Mass Selective Detector. For the analysis, we used an HP 624 column. This was 25 m in length, had internal diameter of 0.2 mm, and was coated with a 1.12 µm thick film of 5% phenyl/95% methylpolysiloxane from Agilent J & W Scientific, Folsom, CA. The initial oven temperature was held at 40 °C for 5 min, then increased to 240 °C at a rate of 20 °C min-1, and finally held at 240 °C for 5 min, with a total time of 20 min. Helium was used as a carrier gas at a constant flow rate of 1 mL min-1 through the column with a total flow of 13.3 mL/min. The initial inlet gas pressure was 115.8 kPa. The temperatures of the injector and the ion source were 250 and 230 °C, respectively. The electron energy was 70 eV. Mass data were collected in full-scan mode (m/z 35-280). The qualifier ion was 91 (m/z). The retention time of toluene was 7.87 min. RESULTS AND DISCUSSION Taking advantage of the material compatibility between the glassy carbon and the carbon nanotubes, a layer of ∼30 µm thick network of SWCNTs was physically deposited onto the distal end of the polished glassy carbon rod following the spraying procedure described above. The nanotube network was characterized by environmental scanning electron microscopy (ESEM). Figure 1a shows the ESEM image of the sensing part of the sensor, where the nanotubes formed a well interconnected spaghetti-like structure. Potentiometric Characteristics. The potentiometric response of the sensor is shown in Figure 1b. The sensor displayed a concentration-dependent response with immediate responses to changes in toluene concentration, though the response time in terms of the IUPAC definition is roughly 4 min irrespective of the concentration of the toluene added. We were using potentiometry to detect neutral species through an adsorption-based mechanism. The detection mechanism was very different from the ion-selective electrodes or other surface-based potentiometric sensors.3,4 We, therefore, expected neither a logarithmic relationship between the EMF and the toluene concentration nor a Nernstian response. The higher the toluene concentration, the higher the increase in EMF that spans a few tens of milivolts. However, the calibration slope was similar to an adsorption isotherm, with a decreasing slope in a nonlinear way, which indicates a gradual saturation of the adsorption sites available on the SWCNTs (Figure 1c).

Figure 1. Potentiometric results. (a) ESEM image of the surface of the sensor, showing an entangled network of SWCNTs. (b) Potentiometric response of the glassy carbon/SWCNT sensor against time for successive additions of increasing concentrations of toluene, 1 ppm-130 ppm. (c) ∆EMF versus concentration of toluene obtained by successive additions of toluene. The dots represent measurements obtained using a sensor of a different size (10 mm L). (d) ∆EMF versus initial concentration of toluene obtained for different toluene test solutions under the same experimental conditions. EMF is calculated with respect to the potential of the 10-5 M NaCl solution in the absence of toluene. Error bars represent the range calculated using three different sensors in all cases.

Sensing was based on a superficial phenomenon. Replicated devices with the same diameter display slightly different superficial areas due to the distinct deposition of the nanotube network. Consequently, replicated sensors show a range of EMF values at a zero level of toluene characterized by a standard deviation of 10.4 mV (N ) 10). Sensors made of glassy carbon rods with different diameters (3, 7, and 10 mm) display different initial EMFs in the absence of the analyte. We observed a very minor enhancement of sensitivity when EMF was recorded against a different concentration of toluene with electrodes of 7 or 10 mm diameter compared to the 3 mm diameter (dotted curve in Figure 1c). The slight increment of sensitivity when electrodes of different surface area were used could be due to the simultaneous change of the surface area, the amount of CNT sprayed on it, and the amount of toluene adsorbed. Consequently, the ratio of amount of CNT over amount of toluene adsorbed could remain nearly constant. To mimic the real world, where coolant water could become contaminated by a sudden or successively increasing leakage of hydrocarbons and to study the differences in the sensor’s response to contamination, we tested different initial concentrations of aromatic hydrocarbon in addition to the successive additions tested earlier. Before each measurement, the sensor was immersed in a solution of 10-5 M NaCl overnight to condition the

sensor and ensure the same starting conditions. As Figure 1d shows, the sensor displayed a consistent nonlinear response to the presence of different initial concentrations of toluene and the same trend as in the successive additions mode but with a slightly enhanced slope. The EMF values correspond to the observed shift in potential from the baseline corresponding to the blank measurement (10-5 M NaCl). The error bars represent the range of potential values obtained for three sensors. We studied the reusability of the sensor device, which is important for its practical use, by recording EMF after regeneration. The sensor was regenerated by heating the active surface gently (T ≈ 100 °C; t ) 3 min) after each round of measurement to eliminate surface-bound target species and solvent molecules. The sensor showed a similar response to that represented in Figure 1d but lost a little sensitivity, which was expected because heating the surface of the sensor, though effective in removing most surface-bound species, does not totally recover the original surface. The limit of detection, calculated as the analyte concentration corresponding to the mean blank responses plus three standard deviations, was 2.1 ppm. We conducted a control experiment to assess the importance of the interaction between nanotubes and aromatic hydrocarbons. The glassy carbon substrate, after pretreatment of its surface to a shiny mirror, was used without carbon nanotubes. Its potential Analytical Chemistry, Vol. 82, No. 19, October 1, 2010

8109

was measured for different toluene concentrations under the same experimental conditions as for the glassy carbon/SWCNT sensor. When carbon nanotubes were not present, the potential increase from the blank line was only 1.5 mV in the presence of 60 ppm of toluene. Aliphatic compounds are part of another major class of components that are generated in the fuel refining process. It would be interesting, therefore, to determine the possibility of detecting them in the coolant water. Cyclohexane is a typical aliphatic compound whose molecules are made of carbon hexagons. The C-C bonds display sp3 hybridization and do not have a π system. Cyclohexane, which has an even higher hydrophobic partitioning coefficient15 than aromatic hydrocarbons, may adsorb onto the walls of SWCNTs by hydrophobic forces that provide a shift in the potential of the glassy carbon/ SWCNT sensor. Our results, however, showed that there is practically no response up to 30 ppm of cyclohexane. These are consistent with those in the literature which show that the adsorption affinity of organic compounds with SWCNTs correlates poorly with chemical hydrophobicity.15 Our sensor is rather insensitive to slight changes of ionic strength (2-10%), the situation that could probably happen in the real world. However, in the presence of ionic species at high levels, the sensitivity depends not only on the absolute value of the ionic strength but also on the nature of the ion. For instance, the sensor is more sensitive to Ca(NO3)2 than to NaCl. The selectivity factors, following the matched potential method (MPM)29 against toluene, were estimated as logkIJMPM ) -1.5 for NaCl and logkIJMPM ) -0.65 for Ca(NO3)2 using ionic strength changes from 10-5 to 10-3 M. Therefore, the sensor is much more sensitive to the interactions with aromatic hydrocarbons than to changes of ionic strength that could result from the sudden changes of ionic species in the monitored sample. It is also interesting to assess the influence of redox couples in the test solution that might be originated from changes of oxygen content in the solution or by interferences due to corrosion effects. To test the effect of O2/H2O/H+, we recorded the EMF of the CNT sensor in the presence and absence of dissolved O2 in the test solution. We did not observe any change in the instrumental response except for the perturbations resulting from purging the solution. Additionally, we checked the influence of a fast charge transfer redox couple such as Fe3+/Fe2+ that might be originated from corrosion effects. To this effect, we recorded the EMF of our CNT sensor in the presence of the redox couple Fe(CN)63-/ Fe(CN)64- at molar ratios: 10.0, 1.0, and 0.1 while maintaining the ionic strength constant. We compared the results with the ones recorded with a Pt electrode. The results show that both the CNT sensor and the Pt electrode are sensitive to changes of the ratio of these redox species when they are at high concentration level (total concentration [Fe(CN)63-] + [Fe(CN)64-] ) 10-2 M). However, when the total concentration of [Fe(CN)63-] + [Fe(CN)64-] decreases to 10-4 M (a much more realistic situation when considering an interference in the actual cooling process), the Pt electrode still shows a response while the CNT sensor displays total insensitivity. (29) Bakker, E.; Pretsch, E.; Bu ¨ hlmann, P. Anal. Chem. 2000, 72, 1127–1133.

8110

Analytical Chemistry, Vol. 82, No. 19, October 1, 2010

Therefore, CNTs are insensitive to this type of fast redox system when they are present at low concentration levels. In our case, the change in recorded potential could result from the change in electrical charges in the interfacial diffuse layer upon change in ionic concentration. We also observed a very small drift (