Construction of Sensitive and Selective Zirconia-Based CO Sensors

Article pubs.acs.org/Langmuir

Construction of Sensitive and Selective Zirconia-Based CO Sensors Using ZnCr2O4-Based Sensing Electrodes Yuki Fujio,†,‡ Vladimir V. Plashnitsa,§,∥ Michael Breedon,‡,⊥ and Norio Miura*,⊥ †

Interdisciplinary Graduate School of Engineering Sciences, Kyushu University, Kasuga-shi, Fukuoka 816-8580, Japan Japan Society for the Promotion of Science, Tokyo 102-8471, Japan § Research and Education Center of Carbon Resources, Kyushu University, Kasuga-shi, Fukuoka 816-8580, Japan ⊥ Art, Science and Technology Center for Cooperative Research, Kyushu University, Kasuga-shi, Fukuoka 816-8580, Japan ‡

S Supporting Information *

ABSTRACT: The carbon monoxide (CO) sensitivity of a mixed-potential-type yttria-stabilized zirconia (YSZ)-based tubular-type sensor utilizing a ZnCr2O4 sensing electrode (SE) was tuned by the addition of different precious metal nanoparticles (Ag, Au, Ir, Pd, Pt, Ru and Rh; 1 wt % each) into the sensing layer. After measuring the electromotive force (emf) response of the fabricated SEs to 100 ppm of CO against a Pt/air−reference electrode (RE), the ZnCr2O4−Au nanoparticle composite electrode (ZnCr2O4(+Au)−SE) was found to give the highest response to CO. A linear dependence on the logarithm of CO concentration in the range of 20−800 ppm at an operational temperature of 550 °C under humid conditions (5 vol % water vapor) was observed. From the characterization of the ZnCr2O4(+Au)−SE, we can conclude that the engineered high response toward CO originated from the specific properties of submicrometer sized Au particles, formed via the coalescence of nanosized Au particles located on ZnCr2O4 grains, during the calcining process at 1100 °C for 2 h. These particles augmented the catalytic activities of the gas-phase CO oxidation reaction in the SE layer, as well as to the anodic reaction of CO at the interface; while suppressing the cathodic reaction of O2 at the interface. In addition, the response of the ZnCr2O4(+Au)−SE sensor toward 100 ppm of CO gradually increased throughout the 10 days of operation, and plateaued for the remainder of the month that the sensor was examined. Correlations between SEM observations and the CO sensing characteristics of the present sensor were suggestive that the sensitivity was mostly affected by the morphology of the Au particles and their catalytic activities, which were in close proximity to the ZnCr2O4 grains. Furthermore, by measuring the potential difference (emf) between the ZnCr2O4(+Au) and a ZnCr2O4 electrode, sensitivities to typical exhaust component gases other than CO were found to be negligible at 550 °C.

■

(NMHCs) at 550 °C in the presence of 5 vol % water vapor.11,15 Here, it should be emphasized that the sensor utilizing a ZnCr2O4−SE gave a negligible response to CO. From this result and considering the mixed-potential theory,23,24 we can consider two potential reasons as to why the sensors using ZnCr2O4−SE exhibited no response to CO under these operating conditions. The first possibility is that most of the CO was oxidized when passing through the ZnCr2O4 sensing layer, due to its high catalytic activity to gas-phase CO oxidation, reducing the proportion of CO that could reach the interface between ZnCr2O4−SE and YSZ solid electrolyte. It is also possible that ZnCr2O4 has poor electrochemical activity to the CO anodic reaction at the interface. Considering the later rationale, if we could enhance the CO sensitivity via the addition of a secondary material into ZnCr2O4−SE, it may be possible to obtain a selective CO sensor by connecting the

INTRODUCTION Since the successful implementation of electrochemical zirconia-based gas sensors to monitor exhaust gas combustion efficiency in automotive engines (λ-sensor),1−4 particular attention has been paid to the development of complementary devices for the detection of residual concentrations of hazardous exhaust gases, such as nitrogen oxides (NO and NO2),5−8 hydrocarbons (HCs)9−15 and carbon monoxide (CO).16−22 Among these reports, there are very few reports of CO sensitive and selective sensors operating at high operating temperatures, such as those found in exhaust gases. The addition of precious metal particles (e.g., Au, Pt, etc.) into the sensing electrode (SE) material was found to be a pertinent method for augmenting gas sensing characteristics of electrochemical zirconia-based sensors.7,12−14,22 Nevertheless, the development of highly sensitive, selective and reliable CO sensors remains challenging. Quite recently, we reported that a mixed-potential-type yttria-stabilized zirconia (YSZ)-based sensor utilizing ZnCr2O4−SE exhibited highly selective electromotive force (emf) responses to nonmethane hydrocarbons © 2011 American Chemical Society

Received: October 7, 2011 Revised: November 28, 2011 Published: December 5, 2011 1638

dx.doi.org/10.1021/la203935w | Langmuir 2012, 28, 1638−1645

Langmuir

Article

previously reported.13,25,26 The separation of CO from other gases was carried out on a RT-Msieve5A capillary column (30 m in length, 0.32 mm in inner diameter, Restek Corp., US) with a linear temperature gradient from 35 to 65 at 5 °C/min. Helium was used as a carrier gas at a column-head pressure of 100 kPa. The mass-to-charge (m/z) ratio was monitored over the range of 10−40. The ion peaks with an m/z ratio of 12 (12C), 16 (16O), and 28 (12C16O) were recorded via selective-ion monitoring. The percentage of CO converted to CO2 was estimated from the peak areas of m/z values at 12, 16, and 28. The current−voltage (polarization) curves were measured by the potentiodynamic method at a constant scan-rate of 5 mV min−1 by using an automatic polarization system (HZ-3000, Hokuto Denko, Japan). The two-probe complex-impedance measurements were performed over the frequency range of 0.1 Hz to 1 MHz with an impedance analyzer (1255WB, Solartron, U.K.), when SE was exposed to the base gas or 100−800 ppm of CO (+synthetic air). The amplitude of the applied alternating-current (AC) potential and the applied direct-current (DC) potential was fixed at 50 mV and 0 mV, respectively, during all measurements. To obtain complex-impedance plots (Nyquist plots), the real component (Z′) of the measured total impedance value (|Z|) was plotted versus the imaginary component (Z″). The crystal structure and microstructure of the SEs formed on YSZ plates (10 × 10 mm, and 0.2 mm in thickness) were elucidated via X-ray diffraction (XRD, 2100VLR/PC, Rigaku, Japan) analysis and field-emission scanning electron microscopy (FE−SEM, JSM-6340F, JEOL, Japan), respectively.

original SE and the modified SE in a similar manner to combined-type planar sensors; making it possible to cancel out the emf responses toward interfering gases, as has already been reported elsewhere.16 In this study, we investigate the CO sensing characteristics of mixed-potential-type YSZ-based sensors utilizing a ZnCr2O4 and precious metal nanoparticle composite electrode (ZnCr2O4(+M)−SE, M: precious metal). After optimizing the SE material, CO selectivity was achieved by combining a ZnCr2O4 electrode with the modified ZnCr2O4 electrode. The stability of the CO sensing characteristics as well as the CO sensing mechanism for the present sensor will be described here, with particular attention paid to correlations between CO sensing characteristics and the morphology of Au nanoparticles in the ZnCr2O4−SE.

■

EXPERIMENTAL SECTION

SE Material Preparation. The SE materials were prepared by mixing commercial ZnCr2O4 powder (Kojundo Chemical Lab. Co., Ltd., Japan) and one of the following suspensions containing dispersed precious metal nanoparticles (Ag, Au, Ir, Pd, Pt, Ru, and Rh; 1 wt % metal content each; Kyoto Nano Chemical Co. Ltd., Japan) in an agate mortar, which was subsequently dried to evaporate the dispersant solution. The original sizes of all of the as-received precious metal particles in a colloidal solution were in the range of 2−6 nm in diameter, as specified by the manufacturer. Sensing Device Fabrication. The sensors were assembled in a tubular configuration using a commercial hemispherically terminated YSZ tube (8 mol % Y2O3-doped ZrO2, Nikkato Corp., Japan). The physical dimensions of the YSZ tube were 300 mm in length; 5 and 8 mm inner and outer diameters, respectively. First, an intermediate YSZ layer was formed on the outer surface of the YSZ tube using a YSZ paste, which was made by mixing a commercial YSZ powder (Tosoh Corp., Japan) and α-terpineol, to improve the mechanical and electrochemical stability of the interface between the SE and YSZ tube, as reported elsewhere.15 Secondly, the prepared SE material pastes (which were also obtained by mixing with α-terpineol) were applied on the prefabricated YSZ layer to form the SE. A commercial Pt paste was applied on the inner surface of the closed end of the YSZ tube to form the Pt/air-RE. Finally, the painted and assembled YSZ tube was sintered at 1100 °C for 2 h in air to fabricate the final sensing device. The schematic view of the final sensing device in the present study is similar in construction to that described previously.15 Measurement of Sensing Performances. Gas sensing measurements were carried out in a conventional gas-flow apparatus equipped with a furnace operating at 550 °C. The SE was exposed to the base gas (synthetic air + 5 vol % H2O) or the sample gas containing representative components of exhaust gases (CO, NO, NO2, CH4, and C3H8; 100 ppm each) or different CO concentrations in the range of 20−800 ppm with a flow rate of 100 cm3 min−1. The Pt/air−RE was always exposed to atmospheric air. Water vapor was introduced along with the base gas (or the sample gas) by using a water-vapor generator consisting of a small in-line evaporator and a microflow pump (L-2100, Hitachi, Japan). The difference in potential (emf) between the two electrodes was measured as the sensing signal by means of a digital electrometer (R8240, Advantest, Japan). The sensitivity (Δemf) to each gas was defined as the difference between the emf value of the sensor in the sample gas and that in the base gas, prior to sample gas exposure. Characterization of SE Materials. The catalytic activity to the gas-phase CO oxidation (700 ppm of CO + synthetic air) to CO2 was separately evaluated for ZnCr2O4 and ZnCr2O4(+Au) powders (50 mg each), calcined at 1100 °C for 2 h in air, using a gas chromatograph mass spectrometer system (GC−MS; GC instrument, GC-17A, Shimadzu, Japan; MS instrument, MS-QP5050A, Shimadzu, Japan) over a temperature range of 25−800 °C. The measurement conditions, such as the capillary column used, operating temperature of the different GC−MS components and gas flow rate, are the same as those

■

RESULTS AND DISCUSSION CO Sensing Characteristics of the Sensors Using ZnCr2O4(+M)−SE. The effect of the addition of different precious metal nanoparticles into the ZnCr2O4−SE on the CO sensing characteristics of the mixed-potential-type YSZ-based sensors, was investigated by measuring the emf response of a ZnCr2O4− or different ZnCr2O4(+M)−SE against an inner Pt/ air−RE, after exposure to 100 ppm of CO. Figure 1 shows the

Figure 1. Comparison of the sensitivities to 100 ppm of CO for the sensors using ZnCr2O4 or ZnCr2O4 (+1 wt % precious metal nanoparticles)−SEs at 550 °C under humid operating conditions (+ 5 vol % H2O).

comparison of the sensitivity to 100 ppm of CO for the sensors using either the ZnCr2O4− or ZnCr2O4(+M)−SEs at an operational temperature of 550 °C under humid operating conditions (5 vol % H2O). It can be seen that the CO sensitivity manifests after the addition of Pt, Pd, and Au nanoparticles. Among them, the ZnCr2O4(+Au)−SE was found to give the highest CO sensitivity. In addition, as has already been reported,11,15 the sensor utilizing a ZnCr2O4−SE exhibited negligible sensitivity to 100 ppm of CO under the 1639


Langmuir

Article

sensors.23 The sensing characteristics of mixed-potential-type YSZ-based sensor are influenced by three different reactions, when the sensing device is exposed to the sample gas.23,24 The first reaction which must be considered is the oxidation of CO on the surface of the SE material, which occurs during the diffusion process as CO approaches the triple phase boundary:

operating conditions described in this manuscript. Hereafter, we used ZnCr2O4(+Au) as a suitable SE material for the sensitive detection of CO, and further examined the sensing characteristics of a sensor consisting of a ZnCr2O4(+Au)−SE and Pt/air−RE. Figure 2 depicts the response transient to different concentrations of CO and the dependence of sensitivity on the logarithm

CO + (1/2)O2 → CO2

(1)

The others are anodic and cathodic reactions at the interface between the SE and YSZ:

(anodic) 2CO + 2O2 − → 2CO2 + 4e−

(2)

(cathodic) O2 + 4e− → 2O2 −

(3)

The obtained mixed-potential arises from the competing anodic reaction of CO (eq 2) and the cathodic reaction of O2 (eq 3). However, the total CO concentration reaching at the interface between SE and YSZ is governed by the catalytic CO oxidation, as described in the gas-phase reaction (eq 1). Thus, we can speculate that both the negligible and enhanced emf responses toward CO for the sensors using the ZnCr2O4− SE and ZnCr2O4(+Au)−SE, respectively, can be attributed to the contributions of the catalytic activities of these three reactions (eqs 1-3). Characterization of ZnCr2O4(+Au)−SE. As mentioned above, the gas-phase CO oxidation catalytic activity (eq 1) in the SE layer is an important factor affecting the sensing characteristics of mixed-potential-type YSZ-based sensor. To quantify this contributing factor, the catalytic activity of ZnCr2O4 and ZnCr2O4(+Au) samples to the gas-phase oxidation of CO to CO2 was estimated via GC−MS measurements. The CO conversion was calculated by the peak area at an m/z values of 12, 16, and 28 after passing the sample gas containing 700 ppm of CO (+ synthetic air) through each sample powder (50.0 mg each), calcined at 1100 °C for 2 h in air. Figure 3

Figure 2. (a) Response transients toward different concentrations of CO in the range of 20−800 ppm and (b) sensitivity dependence on the logarithm of CO concentration for the sensors using either ZnCr2O4−SE or ZnCr2O4(+Au)−SE at 550 °C under humid operating conditions.

of CO concentration over the examined range of 20−800 ppm for the sensor using a ZnCr2O4(+Au)−SE. For comparison, the results of the sensor using a ZnCr2O4−SE were also shown in this figure. It is clear that the sensor using a ZnCr2O4(+Au)−SE exhibited high sensitivity to different concentrations of CO, with good repeatable responses toward 800 ppm of CO, while the ZnCr2O4−SE gave negligible responses to CO within the whole examined concentration range. The Δemf values to a high concentration of CO (800 ppm of CO) at 550 °C for the sensors using either a ZnCr2O4−SE or ZnCr2O4(+Au)−SE were −2 mV and −90 mV, respectively. The 90% response and recovery times against 800 ppm of CO at 550 °C for the sensor utilizing a ZnCr2O4(+Au)−SE were approximately 30 and 100 s, respectively. The CO sensitivity was linearly dependent on the logarithm of CO concentration (51 mV/decade) in the examined range of 20−800 ppm. Such a linear dependence is usually observed in mixed-potential-type

Figure 3. Temperature dependence of CO conversion to CO2, over ZnCr2O4 and ZnCr2O4(+Au) powders (50.0 mg each).

shows the temperature dependence of CO conversion to CO2 for powder samples, representative of the electrode materials. It can be seen that the catalytic activity of ZnCr2O4 to gas-phase CO oxidation increases gradually with increasing temperature in the range of 200−500 °C. Following this, CO oxidation increased drastically, displaying a similar temperature dependent CO conversion of the blank, at temperatures exceeding 500 °C. 1640


Langmuir

Article

Interestingly, 50 mg of ZnCr2O4 powder at an identical operational temperature to the ZnCr2O4−SE (550 °C) had a CO conversion efficiency of approximately 20%. The SE of the fabricated sensor was formed by using a fraction of the prepared electrode paste formulated by mixing either 30.0 mg of ZnCr2O4 or ZnCr2O4(+Au) powder with 30.0 mg of α-terpineol. These GC−MS results indicate that ZnCr2O4−SE of the fabricated sensor has poor catalytic activity to gas-phase CO oxidation even at 550 °C, therefore most of the CO can reach the interface, as the mass of the electrode was significantly lower than the mass of ZnCr2O4 used in GC−MS measurements. On the basis of these results, we believe that the poor response of the ZnCr2O4−SE toward CO is a result of its poor catalytic activity to the anodic reaction of CO at the interface. While, the CO conversion of the ZnCr2O4(+Au) sample increased rapidly above 200 °C and was consistently higher than that of ZnCr2O4. Maintaining its superior conversion efficiency in the temperature range of 300−500 °C, the conversion efficiency increased drastically above 500 °C. It should be noted that 50.0 mg of ZnCr2O4(+Au) powder at 550 °C showed a CO conversion efficiency of about 30%, hence, most of the CO is expected to reach the interface of ZnCr2O4(+Au)−SE/YSZ. It is surprising that even such a small difference in CO conversion efficiency compared ZnCr2O4 (about 20%) with ZnCr2O4(+Au) (about 30%) powders induces the dramatic difference in sensor response (emf) against 800 ppm of CO, utilizing ZnCr2O4− (−2 mV) and ZnCr2O4(+Au)−SEs (−90 mV). To investigate the effect of the addition of Au nanoparticles into ZnCr2O4−SE on the electrochemical characteristics of the fabricated sensor, the polarization (I−V) curve measurements and the AC complex-impedance measurements (Nyquist plots) for the sensors utilizing either a ZnCr 2 O 4 −SE or ZnCr2O4(+Au)−SE were carried out. The modified polarization (I−V) curves of both sensors in base gas and in base gas spiked with 100 ppm of CO at an operational temperature of 550 °C under humid operating conditions are presented in Figure 4. The modified polarization curves for the cathodic reaction of O2 were obtained by plotting the absolute current values measured in base gas, while for the anodic reaction of CO was determined from the absolute values of the difference between the current measured in the base gas and in the sample gas (100 ppm of CO + base gas). As can be seen from Figure 4, the current generated by the anodic reaction of 100 ppm of CO for the ZnCr2O4−SE (Figure 4a) was less than 100 nA over the examined applied potential range from −50 to 0 mV. This was negligibly small when compared with that of the ZnCr2O4(+Au)−SE (several μA) over the same range, indicating that the anodic reaction of CO barely proceeds at the interface between ZnCr2O4−SE and YSZ under these operating conditions. While, the anodic reaction current in 100 ppm of CO for the ZnCr2O4(+Au)−SE (Figure 4a) was found to increase, the cathodic reaction current in the base gas (0.02 μA/mV) decreased drastically (Figure 4b), in comparison with that of the ZnCr2O4−SE (0.16 μA/mV). This may indicate that the anodic reaction of CO was enhanced and the cathodic reaction of O2 was inhibited, by the addition of Au nanoparticles into ZnCr2O4−SE. Here, a mixed-potential in a sample gas can be estimated from the intersection between the modified anodic and cathodic polarization curves, because a mixed-potential arises when the rates of the anodic and cathodic reactions become equal.23 The estimated values under 100 ppm of CO of the ZnCr2O4−SE and ZnCr2O4(+Au)−SE

Figure 4. Modified polarization curves recorded (a) in 100 ppm of CO (+synthetic air) and (b) in synthetic air (21 vol % O2) at 550 °C under humid operating conditions for the sensors utilizing either a ZnCr2O4−SE or ZnCr2O4(+Au)−SE.

were observed to be close to zero and approximately −50 mV, respectively. These values are in good agreement with those experimentally measured (obtained from Figures 1 and 2), confirming that the fabricated sensor follows the trends previously identified as being in good accordance with mixedpotential theory.23,24 The obtained Nyquist plots in the base gas and in the sample gas of varying CO concentrations are presented in Figure 5.

Figure 5. Nyquist plots in the base gas and in sample gas containing different concentrations of CO for YSZ-based sensors using (a) ZnCr2O4−SE and (b) ZnCr2O4(+Au)−SE at 550 °C. (Inset: enlarged view of Nyquist plots for ZnCr2O4−SE.)

The Nyquist plots of the sensor using ZnCr2O4−SE in both base gas and 100−800 ppm of CO (Figure 5a) were observed 1641


Langmuir

Article

positions which were in excellent agreement with those of the standard ZnCr2O4 [JCPDS no. 22−1107] and Au [JCPDS no. 04−0784], respectively. Indicating the formation of a ZnCr2O4 and Au composite, with no formation of solid solution between these two constituents. Figure 6 shows the representative cross-

to have two semicircles; the smaller one was located at the higher frequency range (around 1 MHz) and the larger of the two occurring in the lower frequency range (around 0.1 Hz). It can also be seen that the Nyquist plots at both the higher frequency range and lower frequency range with respect to different concentrations of CO, overlap exactly with Nyquist plots in base gas. We have already reported that the impedancebased device using an YSZ tube and ZnCr2O4−SE could measure the total NOx (NO and NO2) concentration at 700 °C.27,28 Here, we also report the Nyquist plots of the ZnCr2O4−SE in base gas showed two semicircles, and the resistance (Z′) estimated from the intersection of the smaller semicircle at higher frequency range (around 10 kHz ∼ 1 MHz) was mainly due to the ZnCr2O4-bulk resistance (including the YSZ-bulk resistance) and the diameter of the larger incomplete semicircle at lower frequency range (from 0.1 Hz to approximately 10 kHz) seems to be due to the reaction resistance at the interface between YSZ and the ZnCr2O4− SE.27,28 Thus, the unchanging Nyquist plots of ZnCr2O4−SE, when the gas flow was changed from base gas to different concentrations of CO, reinforces the assumption that there is a negligible anodic reaction of CO at the interface of ZnCr2O4− SE and YSZ. Similarly, the Nyquist plots of ZnCr2O4(+Au)−SE (Figure 5b) was also observed to exhibit two semicircles at the higher and lower frequency range. At the higher frequency range, the difference caused by exposure to increasing CO concentrations was insignificant, similar to the ZnCr2O4−SE. It is noteworthy that the semicircle obtained in lower frequency range of the ZnCr2O4(+Au)−SE was found to be much larger than that of the ZnCr2O4−SE, at around 6 kΩ there appears to be a subtle gradient change which may be indicative of the mixing two semicircles, if compared to those in base gas. Thus, we can rationalize the drastic decrease (almost an order of magnitude lower) in cathodic reaction current of O2 for the ZnCr2O4(+Au)−SE compared with ZnCr2O4−SE (Figure 4b) was caused by resistance increasing at lower frequencies, which can be attributed to the interfacial reaction resistance, caused by the addition of Au nanoparticles into ZnCr2O4−SE. Furthermore, the resistance value in the lower frequency range decreased upon the introduction of CO. It is clear that the resistance at the lower frequency range decreases with increasing CO concentration; this phenomenon has also been observed for the similar impedance-based CO sensor.29 Therefore, from the results of the GC−MS, polarization curve and Nyquist plots measurements, we can conclude that the Au nanoparticles added into ZnCr2O4−SE of the mixed-potential-type sensor play an important role in promoting the gas-phase CO oxidation in the SE layer (eq 1) and the anodic reaction of CO at the interface (eq 2), as well as suppressing the cathodic reaction of O2 (eq 3) under our established operating conditions. The morphology of the SE layer is known to affect the sensing characteristics of mixed-potential-type gas sensors,7,8,12,18,30 due to alteration of the rates for the gas-phase CO oxidation reaction in SE layer, anodic reaction of CO and cathodic reaction of O2 at the interface between SE and YSZ. Additionally, the morphology of Au nanoparticles is also an important factor, not only for the sensing characteristics of the mixed-potential-type YSZ-based sensor,7,12,31,32 but also for the catalytic activity of the CO oxidation catalysts.33−40 Therefore, the crystal structure and morphology of the fabricated ZnCr2O4(+Au)−SE were investigated. XRD measurements (Supporting Information, Figure S1) of ZnCr2O4(+Au)−SE calcined at 1100 °C for 2 h in air, revealed Bragg reflection

Figure 6. Representative cross-sectional BS−SEM images for ZnCr2O4(+Au)−SE after calcining at 1100 °C for 2 h: (a) low and (b) high magnification.

sectional backscattering (BS)−SEM images of a ZnCr2O4(+Au)−SE calcined at 1100 °C for 2 h. There is a clear boundary between the ZnCr2O4(+Au)−SE and the underlying YSZ particle layer. The use of an underlying YSZ layer and ZnCr2O4(+Au)−SE is expected to improve the mechanical and long-term stability, when compared with ZnCr2O4(+Au)−SE applied directly to the YSZ tube, as has already been reported elsewhere.15 The white dots in Figure 6a (or white particles in Figure 6b) represent the agglomerated Au nanoparticles added into ZnCr2O4−SE. The Au nanoparticles were found to be dispersed throughout the ZnCr2O4−SE, in close contact with ZnCr2O4 grains, without the formation of an interconnected network of Au nanoparticles. It should be noted that the Au nanoparticles had an original particle size of 2−4 nm in size, which increased to diameters of approximately 50−200 nm after calcination at 1100 °C for 2 h. We can speculate that the Au nanoparticles located on ZnCr2O4 grains agglomerate, forming submicrometer sized Au particles. It has been reported that massive bulk Au is rather inert for the adsorption and activation of oxygen species,7,38 when compared to nanodimensional Au. Thus, it seems that the drastic decrease in catalytic activity to the cathodic reaction of O2 by the addition of Au into ZnCr2O4−SE may be a result of blocking the active cathodic reaction sites of O2 on ZnCr2O4 grains, as submicrometer sized Au have reportedly inert properties for the adsorption and activation of oxygen species. Furthermore, the enhancement of catalytic activity to gas-phase oxidation reaction and anodic reaction of CO as well as the obtained high emf response to CO for the ZnCr2O4(+Au)−SE have been attributed to the adsorption of CO on submicrometer sized Au particles. Thus, the possible 1642


Langmuir

Article

Figure 7. (a) Time dependent emf response to the base gas (synthetic air) and 100 ppm of CO at 550 °C under humid operating conditions for the sensor using ZnCr2O4(+Au)−SE; (b) thermal history of the developed sensor, and representative BS−SEM images of ZnCr2O4(+Au)−SEs (c) asfabricated, (d) after 10 days of operation, and (e) after 33 days of operation. (Each inset: enlarged SEM images of Au particle on ZnCr2O4 grain.)

are representative BS−SEM images of the same electrode material examined just after fabrication and after prolonged intervals at elevated temperature. As noted above, the Au nanoparticles, which were added into ZnCr2O4−SE coalesced, forming submicrometer sized Au particles with diameters of 50−200 nm. After 10 days operation at 550 °C, it was difficult to observe Au particles with diameters of below 100 nm, and these submicrometer sized Au particles were found to have further coalesced to diameters of 100−300 nm. However, there was almost no change in Au particle size from the 10th day to the 33rd day. While, the morphology of the ZnCr2O4 grains remained unchanged throughout the examined period. When Au particles coalesce, it is anticipated that the dispersion of Au particles in ZnCr2O4−SE will become sparser due to the decrease in the number of Au particles as well as the decrease in availability as active sites for gas-phase CO oxidation, resulting in a decrease in the catalytic activity toward the CO oxidation reaction. This means only a fraction CO was oxidized to CO2 in SE layer, and an adequate amount of CO can reach the interface. In addition, the shape of the submicrometer sized Au particles was also found to deviate from a roughly spherical shape observed initially to a polygonal particles with a hexagonal motif, observed on the 10th day and the 33rd day, as presented in inset of Figure 7, parts c−e. It has been reported that the catalytic activity to CO oxidation for crystalline Au particles was higher than that for amorphous Au particles,41 and that a strong contact between the Au particle and its support material leads to higher catalytic activity.35 However, in fact, the sensitivity of the sensor using ZnCr2O4(+Au)−SE toward CO increased, when the catalytic activity to CO oxidation became higher with the crystallization of Au particles and its strong contact with ZnCr2O4 grain. Indicating that the sensing characteristics of the present sensor utilizing ZnCr2O4(+Au)−SE seems to be primarily affected the catalytic activity to the electrochemical reactions (anodic and cathodic reactions) at the interface, i.e., the charge transfer process may be the ratedetermining step. Logically, it is reasonable to expect that the

CO sensing mechanism of the mixed-potential-type YSZ-based sensor utilizing ZnCr2O4(+Au)−SE can be proposed on the basis of the aforementioned results and considerations. First, CO was captured on the submicrometer Au particles with inert properties for the adsorption and activation of oxygen species, in close proximity to the ZnCr2O4 after passing through the SE layer. Secondly, the adsorbed CO on submicrometer sized Au particles spill over onto the SE/YSZ interface. During which, some CO is oxidized to CO2 via the gas-phase CO oxidation reaction (eq 1). Finally, the enhanced anodic reaction of CO (eq 2) and the hindered cathodic reaction of O2 (eq 3) occur simultaneously, resulting in the drastically augmented emf response to CO in the mixed-potential-type YSZ-based sensor using ZnCr2O4(+Au)−SE. However, the behavior of CO migration on submicrometer sized Au particles should be investigated further. Stability Test of ZnCr2O4(+Au)−SE. The stability of the developed sensor in either base gas or 100 ppm of CO was examined at a constant operational temperature of 550 °C in the presence of 5 vol % H2O, for over a month. The summary of the obtained results is depicted in Figure 7a. It can be seen that the emf response to base gas is stable for the whole 33 days examined. However, the emf response to 100 ppm of CO was found to increase gradually before the 10th day operation, and stabilized after 10 days. Here, it has already been reported that the mixed-potential-type YSZ-based sensor using a ZnCr2O4− SE had highly stable sensing characteristics for the 33 days examined.15 Thus, we can consider that the change in sensing characteristics for the sensor using the ZnCr2O4(+Au)−SE was influenced by a change in submicrometer sized Au particles adjacent to neighboring ZnCr2O4 grains. To elucidate the changing emf response during long-term operation, the morphologies of the as-fabricated ZnCr2O4(+Au)−SE, as well as the electrodes operated at 550 °C for 10 days, and for 33 days were examined via BS−SEM. Figure 7b represents the thermal history of the fabricated sensor utilizing a ZnCr2O4(+Au)−SE, and parts c−e of Figures 7 1643


Langmuir

Article

rate-determining step is the charge-transfer process, due to the Butler−Volmer type behavior observed in the polarization curves of ZnCr2O4(+Au)−SE. Thus, we conclude that the engineered high CO sensitivity of ZnCr2O4(+Au)−SE during the long-term operation at 550 °C, was directly attributed to the augmented catalytic activity to the anodic reaction of CO at the interface, which is dependent on the number of Au particles adjacent to ZnCr2O4 grains, as well as their crystallinity and proximity to the SE/YSZ interface during long-term operation. Construction of a Combined-Type CO Selective Sensor. The sensing characteristics of the sensor using ZnCr2O4(+Au)−SE against representative components of exhaust gases (100 ppm each) were investigated at 550 °C under humid operating conditions and depicted in Figure 8.

Figure 9. (a) Schematic view of the sensor structure and (b) its response transients to representative components of exhaust gases (100 ppm each) at 550 °C under humid operating conditions, for a combined-type planar sensor comprised of ZnCr2O4(+Au) and ZnCr2O4 electrodes formed on the same surface of a YSZ tube.

Figure 8. Comparison of cross sensitivities to representative components of exhaust gases (100 ppm each) at 550 °C under humid conditions for the sensors using ZnCr 2 O 4 −SE and ZnCr2O4(+Au)−SE.

utilizing a ZnCr2O4(+Au)−SE exhibited the highest CO sensitivity, with good repeatable responses to 800 ppm of CO. The CO sensitivity was found to vary almost linearly on a logarithmic scale of gas concentration (51 mV/decade) in the examined range of 20−800 ppm. The characterization of ZnCr2O4(+Au)−SE revealed that the SE was a composite consisting of ZnCr2O4 grains and submicrometer sized Au particles 50−200 nm in diameter. The addition of Au nanoparticles into the ZnCr2O4−SE was found to enhance the catalytic activity to CO oxidation as well as to the anodic reaction of CO, while inhibiting the catalytic activity to cathodic reaction of O2; responsible for the high emf response to CO under our examined operational parameters. Furthermore, we confirmed that the sensing mechanism of the sensor using ZnCr2O4(+Au)−SE obeys the mixed-potential theory, and postulated why the present sensor showed high emf response to CO at 550 °C. In addition, the emf response of the ZnCr2O4(+Au)−SE to 100 ppm of CO increased gradually before the 10th day of operation at 550 °C, and stabilized at around 60 mV shortly thereafter. It was revealed via SEM observations that this increase in emf response to CO appears to be influenced by changes in the morphology and crystallinity of Au particles adjacent to ZnCr2O4 grains, as well as the alteration of catalytic activity to both gas-phase CO oxidation in SE layer and the anodic reaction of CO at the interface between SE and YSZ. By connecting the ZnCr2O4(+Au) and ZnCr2O4 electrodes, the present sensor demonstrated highly sensitive and selective response toward CO, due to the electrical cancellation of responses to all examined gases, excluding CO.

For comparison, results for the sensor using a ZnCr2O4− SE were also shown in Figure 8. It was found that ZnCr2O4(+Au)−SE also gave relatively high sensitivity to 100 ppm of C3H8 (−30 mV) and low responses to NO, NO2 and CH4 (|Δemf| < 7 mV), similar to the ZnCr2O4−SE, reported elsewhere.14 Such a similar sensitivity to 100 ppm of C3H8 suggests that the C3H8 response can be canceled out if the ZnCr2O4(+Au) electrode are combined with the ZnCr2O4 electrode, and both electrodes are exposed simultaneously to the sample gas. To validate this approach, we fabricated a combined-type planar sensor comprised of both electrodes on the outer surface of a YSZ tube, as schematically depicted in Figure 9a, and the potential difference between both electrodes was measured as a sensor response in various sample gases at 550 °C under humid operating conditions. The cross sensitivity comparison of this electrode arrangement is presented in Figure 9b. The combined-type sensor showed selective response to CO, and negligible responses to NO, NO2, CH4, and C3H8 by measuring the sensing signal between two electrodes. Here, C3H8 is used as a representative of typical HCs in automotive exhausts, as the sensitivity toward other HCs have been examined with the sensor using ZnCr2O4−SE, as reported previously.15

■

CONCLUSIONS

The CO sensing characteristics of the mixed-potential-type YSZ-based sensor utilizing ZnCr2O4(+M)−SE were examined at an operational temperature of 550 °C under humid operating conditions. Among the examined combinations, the sensor 1644


Langmuir

■

Article

Sberveglieri, G., Eds.; Springer Science+Business Media LCC: New York, 2009; pp 181−207. (24) Fergus, J. W. J. Solid State Electrochem. 2011, 15, 971−984. (25) Lefller, C. W.; Balabanova, L.; Sullivan, D.; Wang, X.; Fedinec, A. L.; Parfenova, H. Am. J. Physiol. Heart Circ. Physiol. 2003, 285, 292− 297. (26) Anderson, C. R.; Wu, W. H. J. Agric. Food Chem. 2005, 53, 7019−7023. (27) Miura, N.; Nakatou, M.; Zhuiykov, S. Electrochem. Commun. 2002, 4, 284−287. (28) Miura, N.; Nakatou, M.; Zhuiykov, S. Sens. Actuators, B 2003, 93, 221−228. (29) Wu, N.; Chen, Z.; Xu, J.; Chyu, M.; Mao, S. X. Sens. Actuators, B 2005, 110, 49−53. (30) Fujio, Y.; Plashnitsa, V. V.; Miura, N. J. Ceram. Soc. Jpn. 2010, 118, 197−201. (31) Plashnitsa, V. V.; Elumalai, P.; Miura, N. J. Electrochem. Soc. 2008, 155, J301−J306. (32) Plashnitsa, V. V.; Elumalai, P.; Fujio, Y.; Kawaguchi, T.; Miura, N. Nanoscale 2011, 3, 2286−2293. (33) Grunwaldt, J. D.; Maciejewski, M.; Becker, O. S.; Fabrizioli, P.; Baiker, A. J. Catal. 1999, 186, 458−469. (34) Boccuzzi, F.; Chiorino, A.; Manzoli, M.; Lu, P.; Akita, T.; Ichikawa, S.; Haruta, M. J. Catal. 2001, 202, 256−267. (35) Haruta, M. CATTECH 2002, 6, 102−115. (36) Guczi, L.; Horváth, D.; Pászti, Z.; Petõ, G. Catal. Today 2002, 72, 101−105. (37) Ruszel, M.; Grzybowska, B.; Samson, K.; Gressel, I.; Klisińska, A. Catal. Today 2006, 112, 126−129. ́ (38) Grzybowska-Swierkosz, B. Catal. Today 2006, 112, 3−7. (39) Sardar, R.; Funston, A. M.; Mulvaney, P.; Murray, R. W. Langmuir 2009, 25, 13840−13851. (40) Xia, X.; Strunk, J.; Busser, W.; Comotti, M.; Schüth, F.; Muhler, M. J. Phys. Chem. C 2009, 113, 9328−9335. (41) Shen, Y.; Yang, X.; Wang, Y.; Zhang, Y.; Zhu, H.; Gao, L.; Jia, M. Appl. Catal., B 2008, 79, 142−148.

ASSOCIATED CONTENT S Supporting Information * XRD patterns of ZnCr2O4−SE and ZnCr2O4(+1 wt % Au)−SE after calcining at 1100 °C for 2 h. This material is available free of charge via the Internet at http://pubs.acs.org/

■

AUTHOR INFORMATION Corresponding Author *Telephone: +81 92 583 8852. Fax: +81 92 583 8976. E-mail: [email protected]. Present Address ∥ Department of Chemistry and Biochemistry, University of Notre Dame, Notre Dame, IN, 46545.

■

ACKNOWLEDGMENTS This work was partially supported by “Grant-in-Aid for JSPS fellows (Nos. 222597 and 2200353)” and “Grant-in-Aid for Scientific Research (B) (No. 22350095)” from the Japan Society for the Promotion of Science as well as Kyushu University programs on “Novel Carbon Resource Sciences”.

■

REFERENCES

(1) Riegel, J.; Neumann, H.; Wiedenmann, H. Solid State Ionics 2002, 152, 783−800. (2) Lee, J. J. Mater. Sci. 2003, 38, 4247−4257. (3) Park, C. O.; Fergus, J. W.; Miura, N.; Park, J.; V, A. Ionics 2009, 15, 261−284. (4) Miura, N.; Jin, H.; Wama, R.; Nakakubo, S.; Elumalai, P.; Plashnitsa, V. V. Sens. Actuators, B 2011, 152, 261−266. (5) Miura, N.; Lu, G.; Yamazoe, N.; Kurosawa, H.; Hasei, M. J. Electrochem. Soc. 1996, 143, L33−L35. (6) Zosel, J.; Franke, D.; Ahlborn, K.; Gerlach, F.; Vashook, V.; Guth, U. Solid State Ionics 2008, 179, 1628−1631. (7) Plashnitsa, V. V.; Elumalai, P.; Fujio, Y.; Miura, N. Electrochim. Acta 2009, 54, 6099−6106. (8) Park, J.; Yoon, B. Y.; Park, C. O.; Lee, W. J.; Lee, C. B. Sens. Actuators B 2009, 135, 516−523. (9) Hibino, T.; Kuwahara, Y.; Kuroki, Y.; Oshima, T.; Inoue, R.; Kitanoya, S.; Fuma, T. Solid State Ionics 1997, 104, 163−166. (10) Zosel, J.; Müller, R.; Vashook, V.; Guth, U. Solid State Ionics 2004, 175, 531−533. (11) Fujio, Y.; Plashnitsa, V. V.; Elumalai, P.; Miura, N. Electrochem. Solid State Lett. 2008, 11, J73−J75. (12) Plashnitsa, V. V.; Elumalai, P.; Kawaguchi, T.; Fujio, Y.; Miura, N. J. Phys. Chem. C 2009, 113, 7857−7862. (13) Wama, R.; Plashnitsa, V. V.; Elumalai, P.; Kawaguchi, T.; Fujio, Y.; Utiyama, M.; Miura, N. J. Electrochem. Soc. 2009, 156, J102−J107. (14) Elumalai, P.; Plashnitsa, V. V.; Fujio, Y.; Miura, N. Sens. Actuators, B 2010, 144, 215−219. (15) Fujio, Y.; Plashnitsa, V. V.; Elumalai, P.; Miura, N. Talanta 2011, 85, 575−581. (16) Miura, N.; Raisen, T.; Lu, G.; Yamazoe, N. Sens. Actuators, B 1998, 47, 84−91. (17) Wu, R. J.; Hu, C. H.; Yeh, C. T.; Su, P. G. Sens. Actuators, B 2003, 96, 596−601. (18) Li, X.; Kale, G. M. Sens. Actuators, B 2007, 123, 254−261. (19) Chevallier, L.; Bartolomeo, E. D.; Grilli, M. L.; Mainas, M.; White, B.; Wachsman, E. D.; Traversa, E. Sens. Actuators, B 2008, 129, 591−598. (20) Shuk, P.; Bailey, E.; Zosel, J.; Guth, U. Ionics 2009, 15, 131−138. (21) Park, J.; Azad, A.; Song, S.; Wachsman, E. D. J. Am. Ceram. Soc. 2010, 93, 742−749. (22) Plashnitsa, V. V.; Anggraini, S. A.; Miura, N. Electrochem. Commun. 2011, 13, 444−446. (23) Miura, N.; Elumalai, P.; Plashnitsa, V. V.; Ueda, T.; Wama, R.; Utiyama, M. In Solid state gas sensing; Comini, E.; Faglia, G.; 1645


Construction of Sensitive and Selective Zirconia-Based CO Sensors

Recommend Documents