Anal. Chem. 2007, 79, 7669-7675
Simultaneous Electrochemical Detection of Nitric Oxide and Carbon Monoxide Generated from Mouse Kidney Organ Tissues Youngmi Lee* and Jiyeon Kim
Department of Chemistry, The University of Tennessee, Knoxville, Tennessee 37996-1600
A planar-type amperometric dual microsensor for simultaneous detection of nitric oxide and carbon monoxide is presented. The sensor consists of a dual platinum microdisk-based working electrode (WE) and a Ag/AgCl counter/ reference electrode covered with an expanded poly(tetrafluoroethylene) (Tetra-tex) gas-permeable membrane. The dual WE possesses two different platinized platinum disks (WE1 and WE2, 250 and 25 µm in diameter, respectively). The larger WE1 is further modified with electrochemical deposition of tin. Use of two sensing disks different in their size as well as in their surface modification produces apparently different sensitivity ratios of NO to CO at WE1 and at WE2 (∼2 and ∼10, respectively) that are induced by favorable CO oxidation on the surface of tin versus platinum. Anodic currents independently measured at WE1 and at WE2 are successfully converted to the concentrations of NO and CO in the co-presence of these gases using the differentiated sensitivities at each electrode. The sensor is evaluated in terms of its analytical performance: respectable linear dynamic range (sub nM to µM); low detection limit (∼1 nM for NO and 50). The employed internal solution composition was as recommended by Tsceng and Yang to optimize kinetics of CO oxidation at a platinum working electrode of a CO gas sensor.15 All the electrochemical experiments for the sensor fabrications were carried out using a scanning electrochemical microscope, model 900B (CH Instruments Inc.). Surface NO and CO Measurements. Dual NO/CO sensors were calibrated before and after NO/CO measurements using standard NO (1.9 mM) and CO (0.9 mM) saturated PBS solutions.28 Standard NO and CO solutions were prepared by purging PBS solutions (pH 7.4) with argon gas for 30 min to remove oxygen and then purging the solutions with either NO or CO, respectively, for another 30 min. The detailed procedures can be found elsewhere.29,30 Mice (c57) were sacrificed by decapitation and both left and right kidneys were immediately excised and stored in cold Euro-Collins solutions (32.7 g/L D-glucose, 2.05 g/L KH2PO4, 7.40 g/L K2HPO4, 1.12 g/L KCl, 0.84 g/L NaHCO3) on ice until they were used for experiments. For the simultaneous measurements of surface NO/CO concentrations, the kidneys (28) Handbook of Chemistry and Physics; Lide, D. R., Ed.; CRC Press Inc.; Boca Raton, FL, 1990-1991. (29) Feelisch, M. J. Cardiovasc. Pharmacol. 1991, 17, S25-S33. (30) Kozma, F.; Johnson, R. A.; Zhang, F.; Yu, C.; Tong, X.; Nasjletti, A. Am. J. Physiol. Regul. Integr. Comp. Physiol. 1999, 276, R1087-R1094.
were cleaned with PBS buffer solution (pH 7.41) and then immersed in PBS. NO/CO sensors were positioned over the kidney tissues bathed in PBS using a micromanipulator (World Precision Instrumentation Inc. Sarasota, FL). Potentials applied to the inner WE of the sensor were 0.7 V for the platinized/Sn deposited Pt electrode (WE1), and 0.85 V for the platinized Pt electrode (WE2) vs Ag/AgCl electrode. The currents at the dual electrode were concurrently recorded using a bipotentiostat and then converted to the concentration of NO and CO using calibration curves obtained immediately before tissue experiments. RESULTS AND DISCUSSION Responses of Clark-Type Amperometric NO Sensor to NO and CO. Most of current electrochemical methods of NO measurements including commercially available ones (one of the major providers is World Precision Instruments Inc.) are based on the use of Clark-type amperometric NO sensors.31 The Clark sensor consists of a WE covered with a gas-permeable membrane for the selective detection of NO. As an analytical signal, current is measured at a constant WE potential, which is sufficient for the electrochemical oxidation of NO. Thus, the measured current is considered quantitatively proportional to the concentration of NO in a sample. Here we raise a question about the responses of these NO sensors when a sample also contains CO in addition to NO. Indeed, gas-permeable membranes (e.g., PTFE) or other polymer membranes (e.g., Nafion, cellulose acetate, etc.) have been employed for NO sensor fabrications to achieve high selectivity by preventing the interference of charged species (nitrite, ascorbate, etc.). On the other hand, these selective membranes may not provide effective selectivity for NO over CO due to their very similar properties. In addition, the oxidation potentials of these gases are very close to each other, and therefore, the differentiation between them is not practicable by varying applied potentials to WE. With this concern, we tested a typical Clark-type NO sensor in terms of its sensitivity and selectivity for both NO and CO. This sensor was composed of a platinized Pt WE and a Ag/AgCl reference electrode (RE) in an aqueous internal solution (most commonly used 30 mM NaCl and 0.3 mM HCl32) covered with PTFE gas-permeable membrane. It (31) Methods in Nitric Oxide Research; Feelish, M., Stamler, J., Eds.; John Wiley: Chichester, U.K., 1996. (32) Shibuki, K. Neurosci. Res. 1990, 9, 69-76.
Analytical Chemistry, Vol. 79, No. 20, October 15, 2007
7671
Figure 2. Responses of a Clark-type amperometric NO sensor to NO and to CO. The sensor consists of a Pt black WE (76-µm diameter), Ag/AgCl RE/CE, and an internal solution (30 mM NaCl, 0.3 mM HCl). Applied potential to WE was 0.75 V vs Ag/AgCl RE. Note that the use of two different X-axes for NO and CO is attributed to the very slow oxidation kinetics of CO. The CO oxidation kinetics is enhanced in the case of NO/CO dual sensors by employing a more acidic internal solution.
can be seen in Figure 2 that the sensor definitely responded not only to NO but also to CO. Even though the sensor sensitivity for NO is 8-10-fold (cf., ∼20-fold at a bare Pt-based sensor) higher over CO and the response time for CO is relatively slower, copresent CO could affect on the accurate measurements of NO. The inaccuracy induced by CO would be significant especially when a CO concentration is relatively greater than a NO concentration. Possibilities of much higher concentration of CO in biological samples have been suggested for rat olfactory receptor neurons even while the values were approximately estimated via indirect method.33 Indeed, our experimental data exhibited 6-8 times greater CO concentration than NO concentration for mouse kidneys. In this particular case, these Clarktype NO sensors could provide overestimated values for NO concentrations (1.75-2 times at a platinized Pt-based sensor, 1.31.4 times at a bare Pt-based sensor). Note that currently it is not achievable to accurately measure NO quantitatively with conventional amperometric NO sensors under biological conditions (where NO and CO usually coexist). Surface Modification of Dual Microelectrodes. Amperometric detection of NO and CO using a dual sensor is based on the electrochemical oxidation of NO and CO on dual WE surfaces. Because of the similar properties of these two gas molecules, it was not plausible to discriminate NO from CO by adjusting applied WE potentials as discussed in the previous section. However, we were able to make the sensitivities for each gas molecule different at each disk WE of a dual sensor with the use of two differentsized WE disks (250- and 25-µm Pt) as well as the different modification of each WE surface. The dual sensors were optimized by accomplishing the sensitivity ratios of NO over CO at WE1 and at WE2 different as much as possible. Figure 3 exhibits linear sweep voltammograms (33) Ingi, T.; Chiang, G.; Ronnett, G. V. J. Neurosci. 1996, 16, 5621-5628.
7672 Analytical Chemistry, Vol. 79, No. 20, October 15, 2007
Figure 3. Linear sweep voltammograms obtained using variously modified Pt working electrodes: platinized (Pt black deposited) by CV; by CC; additional Sn deposited on Pt black deposited by CV; and by CC. Pt black was electrochemically deposited by CV (in 3% chloroplatinic acid solution, scan rate, 20 mV/s) or CC (in 1% chloroplatinic acid solution). Sn was additionally deposited at a constant potential -0.6 V vs Ag/AgCl for 1 h. Solution contained 65 µM CO witih 30 mM NaCl and 0.5 M H2SO4. Working electrode diameter, 250 µm; quiet time, 30 s; scan rate, 100 mV/s.
(LSV) obtained using variously modified platinized Pt-based electrodes. Because the solutions were degassed by purging with Ar gas, CO can be considered stable with its negligible conversion to CO2. In fact, there was no noticeable change observed for the initial 10 consecutive LSV runs. As shown, the measured currents (at ∼0.8 V) attributed to the electrochemical oxidation of CO were higher for the platinized/Sn deposited electrode compared to the simply platinized ones either (by CC or by CV). In addition, the oxidation currents were dependent on platinization techniques. In fact, the highest currents (6-7-fold enhanced current, reasonably considered proportional to sensitivities for CO) were observed with a Pt electrode platinized by chronocoulometry (CC) and then Sn deposited. Therefore, one Pt disk (250-µm diameter, WE1) of a dual WE was modified in this way to enhance the sensitivity particularly to CO without changing it much to NO. Due to much hindered CO oxidation compared to NO, resulting in the usual low sensitivity to CO, a Pt disk with a larger size was used as WE1 designed for mainly detecting CO, because sensitivity is proportional to the sensing area. The other Pt disk (25-µm diameter, WE2) of the dual WE was optimized for higher NO sensitivity over CO and was simply plantinized by cyclic voltammetry (CV) as previously employed in NO sensor fabrication.25 Polarizing Potentials versus Sensitivities. Sensitivities of a dual electrode for both NO and CO were studied as a function of operating potentials for each WE (Table 1). The sensitivities were obtained by calibrating a sensor with dynamic response curves (Figure 4 and Figure 5) at various WE potentials. Detailed calibration procedure is described in the following section. As potentials become more positive, the sensitivities for NO increase and the ones for CO decrease at both WE1 and WE2. On WE1 designed for mainly detecting CO, the sensitivity for CO was highest at 0.7 V; on the other hand, on WE2 designed for detecting mostly NO, the sensitivity for NO was the best at 0.85 V. These two potentials (0.7 V for WE1 and 0.85 V for WE2 vs Ag/AgCl) were employed for all the following experiments. As shown in Table 1, the sensitivity ratios (NO/CO) were reasonably different
Table 1. Sensitivity Comparison of the Sensors Based on Pt Black or on Pt Black/Sn Working Electrodes for NO and CO Depending on Polarization Potentialsa sensitivity (nA/µM) of Pt black/Sn working electrode (250 µm dia)
sensitivity (nA/µM) of Pt black working electrode (25 µm dia)
polarization potential (V vs Ag/AgCl)
NO
CO
NO
0.70 0.75 0.80 0.85
19.8 ( 3.1 20.7 ( 3.5 50.1 ( 2.8 68.8 ( 3.7
9.6 ( 1.5 7.1 ( 1.6 6.7 ( 1.3 5.0 ( 1.1
0.39 ( 0.03 0.41 ( 0.05 0.80 ( 0.05 1.06 ( 0.11
sensitivity ratio of NO to CO
CO
Pt black/Sn WE (250µm dia)
Pt black WE (25µm dia)
0.23 ( 0.02 0.15 ( 0.01 0.13 ( 0.01 0.10 ( 0.01
2.1 2.9 7.5 13.8
1.7 2.7 6.2 10.6
a The data are given as the mean values ( SD (N ) 12). Only 0.7-0.85 V range was tested because the oxidations of both NO and CO are not favorable at a potential lower than 0.7 V and also WEs are not stable at a potential more positive than 0.85 V.
Figure 5. Typical calibration curves for a dual sensor based on a dual microelectrode ((a) WE1 and (b) WE2) to NO and CO corresponding to the dynamic response curves shown in Figure 4.
Figure 4. Typical dynamic response curves to NO and CO obtained using a dual sensor based on a dual microelectrode consisting of 250-µm-diameter Pt black/Sn WE (WE1) and 25-µm-diameter Pt black WE (WE2). Applied potentials to WE1 and WE2 were 0.7 and 0.85 V vs Ag/AgCl RE, respectively. Currents of (a) WE1 and (b) WE2 were recorded with time. A given amount of standard NO or CO saturated solution was injected into a deaerated PBS buffer solution to change the concentration of NO or CO in the solution.
(∼5 times) at WE1 and at WE2 (2.1 and 10.6, respectively) at these optimized potentials. Dynamic Response Curves of NO/CO Dual Sensors. Figure 4 shows typical dynamic response curves of a NO/CO dual sensor, and Figure 5 exhibits the corresponding calibration curves. The sensor currents were recorded with time in response to varying concentrations with successive injections of NO or CO
standard solutions. When the concentrations change for the ranges shown in Figure 4, response times (the time required to reach 90% of the steady-state current) were 16.2-22.9 and 13.816.8 s for NO and 26.8-37.6 and 25.2-31.3 s for CO at WE1 and at WE2, respectively. The response times for CO were improved to a great extent by using a more acidic solution (compared to the case of Figure 1) though those are observed still relatively slower versus NO, being attributed to the slower oxidation kinetics of CO compared to NO. In the comparison between WE1 versus WE2, the response times were faster at WE2 for NO as well as CO, suggesting that more favored NO and CO oxidations at Pt black over at Sn. Detection limits and linear dynamic ranges of the sensors were also evaluated. Detection limits (at signal-tonoise ) 3) were observed as a few-nanomolar ranges:
