sensors Article
Selective Sensing of Gas Mixture via a Temperature Modulation Approach: New Strategy for Potentiometric Gas Sensor Obtaining Satisfactory Discriminating Features Fu-an Li 1,2,† , Han Jin 1,† , Jinxia Wang 3 , Jie Zou 1, * and Jiawen Jian 1, * 1 2 3
* †
Gas Sensors & Sensing Technology Lab., Faculty of Electrical Engineering and Computer Science, Ningbo University, Ningbo 315211, China; [email protected] (F.L.); [email protected] (H.J.) School of Information Engineering, Huangshan University, Huangshan 245021, China School of Electronic and Information Engineering, Ningbo University of Technology, Ningbo 315211, China; [email protected] Correspondence: [email protected] (J.Z.); [email protected] (J.J.); Tel.: +86-574-8760-9489 (J.Z.) These authors contributed equally to this work.
Academic Editor: Andrzej Lewenstam Received: 3 January 2017; Accepted: 28 February 2017; Published: 12 March 2017
Abstract: A new strategy to discriminate four types of hazardous gases is proposed in this research. Through modulating the operating temperature and the processing response signal with a pattern recognition algorithm, a gas sensor consisting of a single sensing electrode, i.e., ZnO/In2 O3 composite, is designed to differentiate NO2 , NH3 , C3 H6 , CO within the level of 50–400 ppm. Results indicate that with adding 15 wt.% ZnO to In2 O3 , the sensor fabricated at 900 ◦ C shows optimal sensing characteristics in detecting all the studied gases. Moreover, with the aid of the principle component analysis (PCA) algorithm, the sensor operating in the temperature modulation mode demonstrates acceptable discrimination features. The satisfactory discrimination features disclose the future that it is possible to differentiate gas mixture efficiently through operating a single electrode sensor at temperature modulation mode. Keywords: Temperature modulation; PCA; discriminating features; In2 O3 ; ZnO
1. Introduction Currently, air pollutants such as nitrogen oxides (NO, NO2 ), carbon monoxide (CO) and hydrocarbons (HCs) derived from vehicles and industrial factories are a concerning issue, due to their effect of generating photochemical smog and airpocalypse [1]. Hence, strict regulations to reduce the emissions of these air pollutants have been implemented worldwide [2,3]. For the purpose of reducing the air pollutants efficiently and/or monitoring the air quality, high-performance gas sensors have to be developed and are considered crucial for improving the performance of combustion controllers and three-way catalysts [4,5]. To date, numerous types of gas sensors, which include but are not limited to the sensing principles of chemiresistive, amperometric, potentiometric and impedancimetric, have been reported [6–10]. Among them, yttria-stabilized zirconia (YSZ)-based electrochemical sensors consisting of a In2 O3 or ZnO sensing electrode (SE) exhibit great potential to sense NOx , CO, HCs and NH3 because of their reliable performance in harsh conditions [11–23]. However, overcoming their poor selectivity is still a challenging issue for these sensors. To promote selectivity, tuning the microstructure of the SEs, e.g., the morphology, thickness, and crystal phase, is usually considered, although this approach relies on sophisticated materials synthesis techniques [24,25]. Meanwhile, the strategy to design the Sensors 2017, 17, 573; doi:10.3390/s17030573
www.mdpi.com/journal/sensors
Sensors 2017, 17, 573
2 of 9
sensor array and combine it with specific data processing algorithms also shows promising results in differentiating gas mixtures [6,26,27]. Some examples of data processing algorithms include but are not limited to Fisher discriminant analysis, hierarchical clustering analysis, artificial neural networks, simple fuzzy logic and principal component analysis (PCA) [27–31]. PCA in multivariate statistical analysis is a method of compressing the dimension of vector data by extracting principal typical components from sample data set. It is a custom program, proposed for analyzing sensor output data [6,26–29]. Nevertheless, complex preparation techniques remarkably raise the cost of sensor arrays, thereby restraining the manufacturing of such devices on a large scale. Herein, we report a novel strategy to discriminate gas mixtures through modulating the operating temperature. A sensor comprised of a single SE is designed and its sensing characteristics are further investigated to verify the applicability of using the proposed discrimination method in future applications. 2. Materials and Methods 2.1. Sample Preparation and Fabrication of the Sensor Commercially available 8YSZ powder (8 % mol Y2 O3 added zirconia, Tosoh, Japan) were used as electrolyte substrate by tape casting. The slurry was prepared by a two-step ball milling procedure. In the first step, the ceramic powder was homogeneously dispersant in a roller ball mill for 24 h using zirconia balls, acrylic resin, xylene, and butyl acetate as milling medium, dispersant, and solvents, respectively. The polyvinyl butyral (PVB) was added as binder and polyethylene glycol (PEG) was used as plasticizer, followed by roller ball mill for another 6 h. The formula of slurry is showed in Table 1. Before tape casting, the slurries were vacuum pumped for 10 min to remove air. The 8YSZ film was casted on Mylar substrate at the blade height of 300 µm. After drying, the thickness of the green 8YSZ tape was maintained at 125 µm. In order to fabricate the thick electrolyte supported layer, six sheets of 8YSZ green tapes were laminated. The 8YSZ green layer was calcined at 1400 ◦ C for 4 h in air to form the 8YSZ ceramic substrate. Finally, the Length, width and thickness were 16, 9 and 0.6 mm, respectively. Table 1. The formula of electrolyte substrate slurry. Mill Media Mat. ZrO2
Initial wt. 800 g
Formula (g) acrylic resin 16
Butyl acetate 25.4
Xylene 25.4
8YSZ 200
PVB 20
PEG 10
To fabricate the sensor device, commercially available ZnO and In2 O3 (Sinopharm, Shanghai, China) powers were used to prepare the sensing electrodes. ZnO and In2 O3 powders were mixed with the weight ratio of 0, 5, 10, 15, 20, 30 and 100 wt.%, respectively. Afterwards, these powder mixtures were mixed with α-terpineol and screen-printed on the surface of the 8YSZ ceramic plate. These semi-manufactured samples were calcined at 800, 900, 1000 and 1100 ◦ C for 2 h in air, respectively to form the SE. Finally, MnO2 paste were screen-printed and calcined in the same way to form the Mn-based reference electrodes (RE). The photographic image of the fabricated sensor is shown in Figure 1a.
Sensors 2017, 17, 573 Sensors 2017, 17, 573
3 of 9 3 of 9 (a)
(b)
Figure (a)Photograph Photograph of fabricated the fabricated comprised In2Ocomposite 3/ZnO composite SE and Figure 1.1.(a) of the sensorsensor comprised of In2 O3of /ZnO SE and Mn-based Mn-based RE; (b) schematic view of the gas sensing characteristics evaluation apparatus. RE; (b) schematic view of the gas sensing characteristics evaluation apparatus.
2.2. 2.2. Characterization Characterization of of Sensing Sensing Characteristics Characteristics The sensing characteristics The gas gas sensing characteristics of of these these sensors sensors were were measured measured in in aa conventional conventional gas gas flow flow ◦ apparatus equipped with an electric furnace in the temperature range of 450–700 °C. Both SE and apparatus equipped with an electric furnace in the temperature range of 450–700 C. Both SE and RE RE are simultaneously exposed %O O2 in in aa balance balance of of N N2)) or or sample sample gas gas (NO (NO2,, CO, are simultaneously exposed to to the the base base gas gas (5 (5 vol. vol.% CO, 2 2 2 C reference C33H H66 and and NH NH33at at50–400 50–400ppm), ppm),respectively. respectively. The The voltages voltages between between sensing sensing electrode electrode and and reference electrode multifunction data electrode was was measured measured by by aa multifunction data acquisition acquisition card card (HP (HP 34970A, 34970A, Agilent, Agilent, Santa Santa Clara, Clara, CA, USA). The potential difference ΔV was subtracted V base gas from Vsample gas, where the Vsample gas CA, USA). The potential difference ∆V was subtracted V base gas from V sample gas , where the V sample gas represents thesensing sensingresponse response of the sensor toward sample gas the Vrepresents base gas represents the represents the of the sensor toward sample gas and theand V base the sensing gas sensing response of the sensor to base gas. Figure 1b demonstrates the schematic of sensing the gas response of the sensor to base gas. Figure 1b demonstrates the schematic view of view the gas sensing characteristics evaluation apparatus. PCA algorithm was writtenvia in-house via MATLAB® characteristics evaluation apparatus. PCA algorithm was written in-house MATLAB®R2009a ver. R2009a ver. 7.8.0 software (MathWorks© Inc., Natick, MA, USA). 7.8.0 software (MathWorks© Inc., Natick, MA, USA). 3. Results and Discussion In order ordertoto obtain sensitivity a desirable discriminatory a obtain highhigh sensitivity and aand desirable discriminatory capabilitycapability through athrough simplified simplified sensor configuration, namelycomposed a sensor composed single SE, thesensing following sensing sensor configuration, namely a sensor of a single of SE,a the following strategy is strategy is proposed: 2O3 composites in different ratios were synthesized and utilized as the proposed: ZnO/In2 OZnO/In composites in different ratios were synthesized and utilized as the SEs to 3 SEs to enhance the sensing magnitude. the performance and patterns responsetopatterns various enhance the sensing magnitude. Then, theThen, performance and response variousto examined ◦ examined gases for the sensor were evaluated and recorded in the temperature of 450–700 °C. gases for the sensor were evaluated and recorded in the temperature range ofrange 450–700 C. Finally, Finally, the obtained response patterns the sample processed PCA algorithm the obtained response patterns to the to sample gasesgases werewere processed withwith the the PCA algorithm to to investigate the differentiation efficiency. It was expected that due to the distinct difference in the investigate the differentiation efficiency. It was expected that due to the distinct difference gas-phase catalytic catalytic and electrocatalytic electrocatalytic activity activity of of the the SE to each target gas, response patterns to these four gases would vary with the change of the operating temperature. Since the PCA algorithm is a powerful discriminant analysis tool, particularly when dealing with response patterns with distinctive differences, these four gases were expected to be identified even if the examination was implemented through a sensor consisting of a single SE. Initially, the sensitivity and response magnitude of the sensor were optimized. Figure 2 shows the response signal (∆V) (ΔV) to 200 ppm sample gases of the sensors using various composite SEs. It can be concluded that the sensing behavior of the sensors operated at 600 ◦°C C is remarkably tuned by the amount of ZnO. For instance, after incorporating 5 wt. % ZnO ZnO with with In In22O33,, the wt.% the response response signal signal to to CO, CO, C33H 6 and NH 3 suddenly dropped, whereas the addition of 5 wt. % ZnO had a minor effect on the H6 and NH3 suddenly dropped, whereas the addition of 5 wt.% response value of NO22. Afterwards, Afterwards, with with continuously continuously increasing increasing the the ZnO amount, the response signal of the sensors to CO, C C33H66 and and NH33 increased increased gradually gradually and and reached reached the the maximum value at 15 wt.% wt. %ZnO. ZnO.Then, Then,the theresponse responsevalue valueto to these these three three target target gases gases started started to to decrease if the ratio of ZnO in the composite was further increased. Note that the sensing performance of the the sensor sensor to to NO NO22 was relatively less affected by incorporating ZnO when compared with its behavior with other three gases. Hence, in view of the highest sensing signal being observed for most of the examined gases by the addition of 15 wt. % ZnO, the sensor comprised of (In2O3 + 15 wt. % ZnO) composite SE was used for the following examination.
Sensors 2017, 17, 573
4 of 9
the addition of 15 wt.% ZnO, the sensor comprised of (In2 O3 + 15 wt.% ZnO) composite SE was used Sensors 2017, 17, 573 4 of 9 for the following examination.
Sensors 2017, 17, 573
4 of 9
Figure 2. Comparison of the characteristics forthe thesensors sensors comprised of various composite Figure 2. Comparison of sensing the sensing characteristics for comprised of various composite SEs (calcined at ◦1000 °C 2for h) 200 to 200 ppmsample sample gases, at 600 °C. ◦ C. SEs (calcined at 1000 C for h)2 to ppm gases,operated operated at 600 Then, the influence of the calcination temperature was investigated to further enhance the
Figure 2. Comparison of the sensing characteristics for the sensors comprised of various composite Then, the influence of the being calcination temperature was to further enhance the sensing sensing After calcined at temperatures in investigated the range 800–1100 °C (with intervals SEsmagnitude. (calcined at 1000 °C for 2 h) to 200 ppm sample gases, operated at 600of°C. ◦ (with intervals of 100 ◦ C), magnitude. being calcined at of temperatures in the range of 800–1100 of 100After °C), the cross-sensitivity the sensors with a SE containing 15 wt. % C ZnO was examined at ◦ 600 °C and the is shown Figure Interestingly, the response magnitude all further the examined the cross-sensitivity of theinsensors a SE containing 15 ZnOtowas examined atgases 600 Then, influence of the3.with calcination temperature waswt.% investigated to enhance the C and reached the maximum value when the sensor was fabricated at 900 °C. The ΔV measured at 600 °C sensing magnitude. After being calcined at temperatures in the to range 800–1100 °C (with intervals is shown in Figure 3. Interestingly, the response magnitude all of the examined gases reached the for 200°C), ppm NH3, C3H6 and 2sensors was about mV, −62.0215mV, mV and 30.08 mV, ◦ C for of 100 theCO, cross-sensitivity of NO thefabricated with a900 SE ◦containing wt. −97.4 % ZnO was examined at maximum value when the sensor was at−40.64 C. The ∆V measured at 600 200 ppm respectively. the3.optimal calcination temperature for thetosensor selected as 600 °C and isConsequently, shown in Figure Interestingly, the response magnitude all the was examined gases CO, NH900 about −40.64 mV, −ZnO 62.02and mV, −internal 97.4 mVconnection and 30.08between mV, respectively. 3, C 3 HRegarding 6 and NOthe 2 was °C. tuning mechanism of was the fabricated the900 reached the maximum value when the sensor at °C. The ΔV measured at◦ 600the °C Consequently, the optimal calcination temperature for the sensor was selected as 900 C. mV, Regarding calcination temperature and behavior, it is probably due to the−97.4 gas-phase catalytic and for 200 ppm CO, NH3, C 3H6 the andsensing NO2 was about −40.64 mV, −62.02 mV, mV and 30.08 the tuning mechanism of the and the electrode internal connection between calcination temperature electrocatalytic activity of ZnO thethe composite being closely dependent on its composition and respectively. Consequently, optimal calcination temperature for the the sensor was selected as microstructure. This will be thoroughly studied in the near future and will not be discussed in and the900 sensing behavior,the it is probably due toofthe catalytic electrocatalytic activity of °C. Regarding tuning mechanism thegas-phase ZnO and the internaland connection between this the work, since the main focus ofthe thissensing research is to confirm the feasibility themicrostructure. strategy incatalytic identifying calcination temperature and behavior, is probably due toof the gas-phase and will be the composite electrode being closely dependent onitits composition and This gas mixtures with a single SE. Afterwards, the dependence of the response value on the and gas electrocatalytic activity of the composite electrode being closely dependent on its composition thoroughly studied in the near future and will not be discussed in this work, since the main focus concentration also For brevity, list near the sensing behavior to be C3H 6 herein in as this the microstructure.was This willexamined. be thoroughly studiedwe in the future and will not discussed of this research is to confirmAs the feasibility of 4, the strategy in identifying gasvalid mixtures with representative shown in Figure response value of the sensor linearly witha single work, since theexample. main focus of this research is tothe confirm the feasibility of the strategy in identifying SE. Afterwards, the of dependence of the response value on sensor the gas concentration was also examined. the theaCsingle 3H6 concentration; sensitivity of the roughly −30.21 gas logarithm mixtures with SE. Afterwards, the dependence of was the responseestimated value onas the gas mV/Dec. Furthermore, can be seen from inserted in Figure that the sensor For brevity, we list the behavior C3 Hthe herein asillustration the representative As shown in concentration wassensing also itexamined. Forto brevity, list the sensing behavior to C4example. 3H 6 herein as the 6 we showed a quickexample. response rate, with a Figure 90% response time to 200 3Hsensor 6 of of about 17 s. However, representative As shown in 4,linearly the response value of C the valid with Figure 4, the response value of the sensor valid with theppm logarithm the Clinearly 3 H6 concentration; the sensor relatively low recovery rate even when operated theroughly temperature of 600as°C, e.g., logarithm ofa the C3Hroughly 6 concentration; sensitivity of the sensor at was estimated −30.21 sensitivity of the gave sensor was estimated as −30.21 mV/Dec. Furthermore, it can be seen from the 90% recovery time to 200 ppm C 3 H 6 was about 34 s, which needs further improvement in the mV/Dec. Furthermore, it can be seen from the inserted illustration in Figure 4 that the sensor the inserted illustration in Figure 4 that the sensor showed a quick response rate, with a 90% response future. showed a quick response rate, with a 90% response time to 200 ppm C3H6 of about 17 s. However, time to the 200sensor ppmgave C3 Ha6relatively of aboutlow 17 recovery s. However, the sensor gave a relatively low recovery rate even rate even when operated at the temperature of 600 °C, e.g., ◦ C, e.g., the 90% recovery time to 200 ppm C H was about when operated at the temperature of 600 3 in 6 the the 90% recovery time to 200 ppm C3H6 was about 34 s, which needs further improvement 34 s, which needs further improvement in the future. future.
Figure 3. Influence of the calcination temperature (800–1100 °C, with intervals of 100 °C) on the sensing behavior of the sensor comprised of (In2O3 + 15 wt. % ZnO)-SE, operated at 600 °C. Figure 3. Influence of the calcination temperature (800–1100 with intervals of 100◦ °C) on the Figure 3. Influence of the calcination temperature (800–1100 ◦ C,°C, with intervals of 100 C) on the sensing sensing behavior of the sensor comprised of (In2O3 + 15 wt. % ZnO)-SE, operated at 600 °C. behavior of the sensor comprised of (In2 O3 + 15 wt.% ZnO)-SE, operated at 600 ◦ C.
Sensors 2017, 17, 573 Sensors Sensors 2017, 2017, 17, 17, 573 573
5 of 9 55 of of 99
Figure 4. Dependence sensing signal signal (∆V) (ΔV) on the the logarithm logarithm of C H66 concentrations; Dependence of of the concentrations; Insert: Insert: Figure 4. Dependence of the sensing (ΔV) on logarithm of of C C333H concentrations; Insert: response transitions to C 3 H 6 in the concentration range of 50–400 ppm. H66 in inthe theconcentration concentrationrange range of of 50–400 50–400 ppm. ppm. response transitions to C3H
Theoretically, high discrimination efficiency can obtained through operating the sensor in Theoretically, highdiscrimination discriminationefficiency efficiencycan canbebe be obtained through operating sensor in Theoretically, high obtained through operating thethe sensor in the the temperature modulation mode, if distinct differences in the response patterns are observed. the temperature modulation mode, if distinct differences the response patterns are observed. temperature modulation mode, if distinct differences in theinresponse patterns are observed. Thus, Thus, the response patterns to the 50 ppm sample gas were recorded within the operational Thus, the response patterns to the 50 ppm sample gas were recorded within the operational the response patterns to the 50 ppm sample gas were recorded within the operational temperatures of temperatures of 450–700 °C for sensor using the 2O3 + 15 wt. % ZnO) SE. It can be concluded temperatures 450–700 for the the the (In (In 2O3 +SE. 15 It wt. % be ZnO) SE. It can be concluded 450–700 ◦ C forofthe sensor°C using thesensor (In2 O3using + 15 wt.% ZnO) can concluded from the results from the results summarized in Figure 5 that the sensor demonstrated apparent differences in from the results summarized in Figure 5 that the sensor demonstrated apparent differences in the the summarized in Figure 5 that the sensor demonstrated apparent differences in the response patterns to response patterns to each target gas within the examined temperature range. With the increasing response patterns to each target gas within the examined temperature range. With the increasing each target gas within the examined temperature range. With the increasing operational temperature, operational temperature, the response operational response value value of of NO NO22 decreased, decreased, whereas whereas the the other other three three sample sample gases gases the responsetemperature, value of NOthe 2 decreased, whereas the other three sample gases gave a more complex gave a more complex variation in their sensing behavior. Based on these results, it is reasonable to gave a more complex variation in their sensing behavior. Based on these results, it is reasonable to variation in their sensing behavior. Based on these results, it is reasonable to believe that desirable believe that desirable discriminating features will be obtained if processing these response patterns believe that desirable discriminating features will be obtained if processing these response patterns discriminating features will be obtained if processing these response patterns with the PCA algorithm. with the Consequently, a 20 × column) feature vector (shown in Table 2) with the PCA PCA algorithm. algorithm. Consequently, 20 × × 66 (row (row feature Table 2) Consequently, a 20 × 6 (row × column)afeature vector× column) (shown in Table vector 2) was(shown createdin and input was created and input into the PCA software. Each row represents a single measurement at aa was created and input into the PCA software. Each row represents a single measurement at into the PCA software. Each row represents a single measurement at a specific gas concentration specific gas concentration within examined temperature ranges each represents aa specificall gas concentration within all all the the examined temperature ranges and and each column column represents within the examined temperature ranges and each column represents a feature calculated for that feature calculated for that particular measurement at a specific temperature. All data obtained from feature calculated for thatatparticular a specific temperature. obtained from particular measurement a specificmeasurement temperature.atAll data obtained from All the data sensor operated at the sensor operated at six different temperatures is equivalent to that given by six different sensors the sensor operated at six different temperatures is equivalent to that given by six different sensors six different temperatures is equivalent to that given by six different sensors that were operated at that were operated the temperature. As aa result, aa virtual sensor array that wereoperational operated at at the same same operational operational Asarray result, virtual the same temperature. As a result, atemperature. virtual sensor composed of sixsensor sensorsarray was composed of six sensors was established and applied for sensing these four gases at the level of composed of six sensors was established and applied for sensing these four gases at the level of established and applied for sensing these four gases at the level of 50–400 ppm. A flowchart that 50–400 ppm. A flowchart that summarizes the procedure is shown in Figure 6. 50–400 ppm. A flowchart that summarizes the procedure is shown in Figure 6. summarizes the procedure is shown in Figure 6.
Figure 5. Response patterns of the the (50 ppm) at operating Figure 5. 5. Response Responsepatterns patterns of sensor the sensor sensor to the examined examined gases (50 ppm)operating at various various operating of the to the to examined gases (50gases ppm) at various temperatures. temperatures. temperatures.
Sensors 2017, 17, 573
6 of 9
Sensors 2017, 17, 573
6 of 9
Table 2. The 20 × 6 (row × column) feature vectors being created for PCA analysis.
Table 2. The 20 × 6 (row × column) feature vectors being created for PCA analysis.
Response at Various Operating Temperatures/mV
Sample Gas
Concentration/ppm
Sample Gas
CO CO
C3 H6 C3H6
NH3
NH3
Concentration/ppm
50 50 90 150 90 250150 400250
NO2
−28.80 −−28.80 35.55 −−35.55 49.86 −−49.86 59.84
50 400 90 50 150 90 250150 400250
−−59.84 64.00 −−64.00 85.79 −−85.79 95.33 −−95.33 101.95 −−101.95 113.98
400
−−113.98 11.85 −−11.85 20.89 −−20.89 27.35 −−27.35 40.60 −−40.60 50.66
50 90 50 150 90 250150 400250 400
NO2
Operating Temperature Response at Various Operating Temperatures/mV 500 ◦ C Operating 550 ◦ CTemperature 600 ◦ C 650 ◦ C 700 ◦ C 450 °C 500 °C 550 °C 600 °C 650 °C 700 °C −17.86 −17.58 −13.98 −19.32 −15.64 −3.36 −17.86 −17.58 −13.98 −19.32 −15.64 −3.36 450 ◦ C
50 50 90 90 150 150 250 250 400 400
−50.66
68.49 68.49 81.07 81.07 84.90 84.90 89.57 89.57 105.53
105.53
Sensor Sensor11
−29.52
−30.65
−32.47
−25.07
−8.19
−29.52 −38.18
−30.65 −38.89
−32.47 −42.59 −25.07 −29.41 −8.19 −13.59
−38.18 −48.76
−38.89 −47.51
−42.59 −49.79 −29.41 −37.41−13.59−20.62
−57.19 −48.76
−55.35 −47.51
−60.04 −37.41 −44.59−20.62−28.60 −49.79
−57.19 −55.35 −73.63 −56.55 −60.04 −72.98 −44.59 −44.70−28.60−18.36 −73.63 −56.55 −90.31 −69.14 −72.98 −88.84 −44.70 −81.54−18.36−49.13 −90.31 −69.14 −87.33 −74.55 −88.84 −88.39 −81.54 −85.39−49.13−56.04 −87.33 −74.55 −92.90 −83.55 −88.39 −98.40 −85.39 −94.02−56.04−65.11 −92.90 − 106.38 −83.55 −90.32 −98.40 −97.68 −94.02 −101.02−65.11−73.50 −106.38 −90.32 −22.04 −15.27 −97.68 −25.23 −101.02 −28.34−73.50−15.47 −22.04 −15.27 −25.23 −28.34 −31.12 −21.69 −32.46 −46.30−15.47−29.93 −31.12 −21.69 −32.46 −46.30 −37.71 −31.58 −50.86 −59.13−29.93−44.26 −37.71 −31.58 −47.75 −39.50 −50.86 −57.74 −59.13 −72.44−44.26−57.34 −47.75 −39.50 −59.53 −45.50 −57.74 −65.48 −72.44 −84.14−57.34−68.49 −59.53 −45.50 −65.48 −84.14 −68.49 69.45 54.76 48.34 27.39 14.27 69.45 54.76 48.34 27.39 14.27 80.73 57.42 58.15 41.94 27.01 80.73 57.42 58.15 41.94 27.01 86.86 56.96 60.37 49.97 39.47 86.86 56.96 60.37 49.97 39.47 88.81 54.94 61.28 56.48 50.57 88.81 54.94 61.28 56.48 50.57 73.11 49.51 56.35 61.79 63.01 73.11 49.51 56.35 61.79 63.01 Sensor22 Sensor Sensor Sensor Sensor 5 Sensor 6 Sensor 3 3 Sensor 4 4Sensor 5 Sensor 6 Virtual Sensor Array Virtual Sensor Array
Figure 6. Flowchart temperature Figure 6. Flowchart of of processing processing the the response response pattern pattern that that was was obtained obtained from from the the temperature modulation mode with PCA software in detail. modulation mode with PCA software in detail.
The PCA transformation for all response patterns is shown in Figure 7, with the same colored The PCA transformation for all response patterns is shown in Figure 7, with the same colored symbol corresponding to each target gas. The spatial distance between each target gas represents the symbol corresponding to each target gas. The spatial distance between each target gas represents discrimination capability of the sensor. The overlap of these symbols shown in the PCA scheme the discrimination capability of the sensor. The overlap of these symbols shown in the PCA scheme implies low differentiating characteristics of the sensor when sensing these symbols representing implies low differentiating characteristics of the sensor when sensing these symbols representing gas. In contrast, a large spatial distance between these symbols suggests a desirable discrimination gas. In contrast, a large spatial distance between these symbols suggests a desirable discrimination feature detected by the sensor array. The results demonstrated in the PCA scheme suggest that the feature detected by the sensor array. The results demonstrated in the PCA scheme suggest that the sensor exhibits the best discrimination capability to NO2 and acceptable differentiating features to sensor exhibits the best discrimination capability to NO2 and acceptable differentiating features to CO, CO, NH3 and C3H6. This is in line with the sensing characteristics shown in Figure 5. As shown in Figure 5, the sensor gave relatively similar response patterns to CO, NH3 and C3H6 within the examined temperature range, whereas distinctive sensing behavior was observed for NO2, thereby leading to the satisfactory differentiating feature for NO2 and less discriminating capability for CO,
Sensors 2017, 17, 573
7 of 9
NH3 and C3 H6 . This is in line with the sensing characteristics shown in Figure 5. As shown in Figure 5, the sensor gave relatively similar response patterns to CO, NH3 and C3 H6 within the examined temperature range, Sensors 2017, 17, 573 whereas distinctive sensing behavior was observed for NO2 , thereby leading 7 of 9 to the satisfactory differentiating feature for NO2 and less discriminating capability for CO, NH3 and C3 H6 . 3 and C3H6. Note that the preliminary success of applying the temperature modulation approach NoteNH that the preliminary success of applying the temperature modulation approach together with together with the PCA algorithm in discriminating gas mixtures gives a potential method to replace the PCA algorithm in discriminating gas mixtures gives a potential method to replace the complex the complex sensor array with a simplified sensor configuration. In particular, it is reasonable to sensor array with a simplified sensor configuration. In particular, it is reasonable to expect that the expect that the discrimination features will be further improved if the SE exhibits distinctive discrimination features will further improved the efforts SE exhibits distinctive response patterns to all response patterns to all thebe target gases, implyingifmore are needed to develop an appropriate the target gases, implying more efforts are needed to develop an appropriate SE in future research. SE in future research.
Figure 7. PCA schemes ofof the sensorutilizing utilizing O 15+wt. 15 % wt.% ZnO)-SE, operated in Figure 7. PCA schemes thedata dataset setfor for the the sensor (In(In 2O23 + 3 ZnO)-SE, operated in the temperature modulation mode. the temperature modulation mode.
4. Conclusions 4. Conclusions Through fabricatingaacomposite composite SE operating it initthe modulation mode, themode, Through fabricating SEand and operating intemperature the temperature modulation sensor consisting of the single SE exhibited an acceptable performance in sensing four types of the sensor consisting of the single SE exhibited an acceptable performance in sensing four types of hazardous gases. With the addition of 15 wt. % ZnO to In2O3 and being calcined at 900 °C for 2 h, the hazardous gases. With the addition of 15 wt.% ZnO to In2 O3 and being calcined at 900 ◦ C for 2 h, sensor gave its maximum sensing magnitude to all the examined gases. Moreover, with the aid of the sensor gave its maximum sensing magnitude to all the examined Moreover, with the aid the PCA algorithm, acceptable differentiating features were observedgases. through the temperature of the PCA algorithm, differentiating features were observed through the temperature modulation approach.acceptable Such promising results indicate a novel strategy to sense gas mixtures with a modulation Suchsensor promising results indicate a novel strategy to sense gas mixtures with simplifiedapproach. potentiometric configuration.
a simplified potentiometric sensor configuration. Acknowledgments: This work was financially sponsored by Kuan Cheng Wong Magna Fund at Ningbo University and supported by the Research Foundation for Startup Researcher (ZX2016000095), Acknowledgments: This work wasScientific financially sponsored by Kuan Cheng Wong Magna Fund atthe Ningbo Nationaland Natural Science Foundation of China Research (61301050, 61471210 and 51472126). University supported by the Scientific Foundation for Startup Researcher (ZX2016000095), the National Natural Science Foundation of China (61301050, 61471210 and 51472126). Author Contributions: Jie Zou and Han Jin conceived and designed the experiments; Fu-an Li performed the
Author Contributions: JieJinxia Zou Wang and Han Jin conceived and designed the experiments; Fu-an performed the experiments; Fu-an Li, and Jiawen Jian analyzed the data; Jinxia Wang and Jiawen JianLi contributed experiments; Fu-an Li, Jinxia Wang and Jiawen JianJin analyzed data; Jinxia Wang and Jiawen Jian contributed reagents/materials/analysis tools; Jie Zou and Han wrote thethe paper. reagents/materials/analysis tools; Jie Zou and Han Jin wrote the paper. Conflicts of Interest: The authors declare no conflict of interest.
Conflicts of Interest: The authors declare no conflict of interest.
Abbreviations
Abbreviations
The following abbreviations are used in this manuscript:
The following abbreviations are used in this manuscript: YSZ
Yttria-Stabilized Zirconia
Sensing Electrode YSZ SE Yttria-Stabilized Zirconia Reference Electrode SE RE Sensing Electrode principleElectrode component analysis RE PCAReference PCA principle component analysis
Sensors 2017, 17, 573
8 of 9
References 1. 2. 3.
4. 5. 6. 7. 8. 9. 10.
11. 12. 13. 14. 15. 16.
17. 18. 19. 20. 21. 22. 23.
Karthikeyan, S.; Prathima, A.; Sabariswaran, K. An environmental effect of nano additive on performance and emission in a biofuel fuelled marine engine. Indian J. Geo. Mar. Sci. 2015, 44, 896–901. Tzamkiozis, T.; Ntziachristos, L.; Samaras, Z. Diesel passenger car pm emissions: From euro 1 to euro 4 with particle filter. Atmos. Environ. 2010, 44, 909–916. [CrossRef] O’Driscoll, R.; ApSimon, H.M.; Oxley, T.; Molden, N.; Stettler, M.E.J.; Thiyagarajah, A. A portable emissions measurement system (pems) study of nox and primary NO2 emissions from euro 6 diesel passenger cars and comparison with copert emission factors. Atmos. Environ. 2016, 145, 81–91. [CrossRef] Zhang, H.; Wang, J.M.; Wang, Y.Y. Sensor reduction in diesel engine two-cell selective catalytic reduction (scr) systems for automotive applications. IEEE-ASME Trans. Mechatron. 2015, 20, 2222–2233. [CrossRef] Simons, H.; Simon, U. Zeolites as nanoporous, gas-sensitive materials for in situ monitoring of denox-scr. Beilstein J. Nanotechnol. 2012, 3, 667–673. [CrossRef] [PubMed] Hannon, A.; Lu, Y.; Li, J.; Meyyappan, M. A sensor array for the detection and discrimination of methane and other environmental pollutant gases. Sensors 2016, 16, 1163. [CrossRef] [PubMed] Chiu, S.W.; Tang, K.T. Towards a chemiresistive sensor-integrated electronic nose: A review. Sensors 2013, 13, 14214–14247. [CrossRef] [PubMed] Miura, N.; Sato, T.; Anggraini, S.A.; Ikeda, H.; Zhuiykov, S. A review of mixed-potential type zirconia-based gas sensors. Ionics 2014, 20, 901–925. [CrossRef] Rheaume, J.M.; Pisano, A.P. A review of recent progress in sensing of gas concentration by impedance change. Ionics 2011, 17, 99–108. [CrossRef] Zou, J.; Zheng, Y.; Li, J.; Zhan, Z.; Jian, J. Potentiometric NO2 sensors based on thin stabilized zirconia electrolytes and asymmetric (La0.8 Sr0.2 )0.95 MnO3 electrodes. Sensors 2015, 15, 17558–17571. [CrossRef] [PubMed] Lu, G.; Miura, N.; Yamazoe, N. High-temperature hydrogen sensor based on stabilized zirconia and a metal oxide electrode. Sens. Actuators B Chem. 1996, 35–36, 130–135. [CrossRef] Yamaura, H.; Jinkawa, T.; Tamaki, J.; Moriya, K.; Miura, N.; Yamazoe, N. Indium oxide-based gas sensor for selective detection of CO. Sens. Actuators B Chem. 1996, 36, 325–332. [CrossRef] Ueda, T.; Plashnitsa, V.V.; Nakatou, M.; Miura, N. Zirconia-based amperometric sensor using zno sensing-electrode for selective detection of propene. Electrochem. Commun. 2007, 9, 197–200. [CrossRef] Jin, H.; Breedon, M.; Miura, N. Sensing behavior of YSZ-based amperometric NO2 sensors consisting of Mn-based reference-electrode and In2 O3 sensing-electrode. Talanta 2012, 88, 318–323. [CrossRef] [PubMed] Lee, I.; Jung, B.; Park, J.; Lee, C.; Hwang, J.; Park, C.O. Mixed potential NH3 sensor with LaCoO3 reference electrode. Sens. Actuators B Chem. 2013, 176, 966–970. [CrossRef] Dai, L.; Liu, Y.; Meng, W.; Yang, G.; Zhou, H.; He, Z.; Li, Y.; Wang, L. Ammonia sensing characteristics of La10 Si5 MgO26 -based sensors using In2 O3 sensing electrode with different morphologies and cuo reference electrode. Sens. Actuators B Chem. 2016, 228, 716–724. [CrossRef] Liu, F.; Guan, Y.; Sun, H.; Xu, X.; Sun, R.; Liang, X.; Sun, P.; Gao, Y.; Lu, G. Ysz-based NO2 sensor utilizing hierarchical In2 O3 electrode. Sens. Actuators B Chem. 2016, 222, 698–706. [CrossRef] Miura, N.; Shiraishi, T.; Shimanoe, K.; Yamazoe, N. Mixed-potential-type propylene sensor based on stabilized zirconia and oxide electrode. Electrochem. Commun. 2000, 2, 77–80. [CrossRef] Miura, N.; Raisen, T.; Lu, G.; Yamazoe, N. Highly selective CO sensor using stabilized zirconia and a couple of oxide electrodes. Sens. Actuators B Chem. 1998, 47, 84–91. [CrossRef] Sun, F.; Li, X.; Liu, L.; Wang, J. Novel Zn–M–O (M = Sn, Co) sensing electrodes for selective mixed potential CO/C3 H8 sensors. Sens. Actuators B Chem. 2013, 184, 220–227. [CrossRef] Li, X.; Kale, G.M. Influence of thickness of ito sensing electrode film on sensing performance of planar mixed potential CO sensor. Sens. Actuators B Chem. 2006, 120, 150–155. [CrossRef] Chen, Y.; Xiao, J. Effects of ysz addition on the response of La2 CuO4 sensing electrode for a potentiometric NOx sensor. Sens. Actuators B Chem. 2014, 192, 730–736. Li, X.; Wang, C.; Wang, B.; Yuan, Y.; Huang, J.; Zhang, H.; Xia, F.; Xiao, J. Effects of sintering temperature on the NH3 sensing properties of Mg2 Cu0.25 Fe1 O3.75 electrode for ysz-based potentiometric NH3 sensor. Ceram. Int. 2016, 42, 2214–2220. [CrossRef]
Sensors 2017, 17, 573
24. 25. 26. 27.
28.
29. 30.
31.
9 of 9
Hu, X.; Tian, L.; Sun, H.; Wang, B.; Gao, Y.; Sun, P.; Liu, F.; Lu, G. Highly enhanced NO2 sensing performances of Cu-doped In2 O3 hierarchical flowers. Sens. Actuators B Chem. 2015, 221, 297–304. [CrossRef] Zhang, D.; Liu, Z.; Li, C.; Tang, T.; Liu, X.; Han, S.; Lei, B.; Zhou, C. Detection of NO2 down to ppb levels using individual and multiple In2 O3 nanowire devices. Nano Lett. 2004, 4, 1919–1924. [CrossRef] Tomchenko, A.A.; Harmer, G.P.; Marquis, B.T.; Allen, J.W. Semiconducting metal oxide sensor array for the selective detection of combustion gases. Sens. Actuators B Chem. 2003, 93, 126–134. [CrossRef] Natale, C.D.; Davide, F.A.M.; Amico, A.D.; Sberveglieri, G.; Nelli, P.; Faglia, G.; Perego, C. Complex chemical pattern recognition with sensor array the discrimination of vintage years of wine. Sens. Actuators B Chem. 1995, 25, 801–804. [CrossRef] Sheini, A.; Khajehsharifi, H.; Shahbazy, M.; Kompany-Zareh, M. A chemosensor array for the colorimetric identification of some carboxylic acids in human urine samples. Sens. Actuators B Chem. 2017, 242, 288–298. [CrossRef] Li, D.; Lei, T.; Zhang, S.; Shao, X.; Xie, C. A novel headspace integrated E-nose and its application in discrimination of chinese medical herbs. Sens. Actuators B Chem. 2015, 221, 556–563. [CrossRef] Zhang, L.; Tian, F.; Peng, X.; Dang, L.; Li, G.; Liu, S.; Kadri, C. Standardization of metal oxide sensor array using artificial neural networks through experimental design. Sens. Actuators B Chem. 2013, 177, 947–955. [CrossRef] Zampolli, S.; Elmi, I.; Ahmed, F.; Passini, M.; Cardinali, G.C.; Nicoletti, S.; Dori, L. An electronic nose based on solid state sensor arrays for low-cost indoor air quality monitoring applications. Sens. Actuators B Chem. 2004, 101, 39–46. [CrossRef] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Selective Sensing of Gas Mixture via a Temperature Modulation Approach: New Strategy for Potentiometric Gas Sensor Obtaining Satisfactory Discriminating Features Fu-an Li 1,2,† , Han Jin 1,† , Jinxia Wang 3 , Jie Zou 1, * and Jiawen Jian 1, * 1 2 3
* †
Gas Sensors & Sensing Technology Lab., Faculty of Electrical Engineering and Computer Science, Ningbo University, Ningbo 315211, China; [email protected] (F.L.); [email protected] (H.J.) School of Information Engineering, Huangshan University, Huangshan 245021, China School of Electronic and Information Engineering, Ningbo University of Technology, Ningbo 315211, China; [email protected] Correspondence: [email protected] (J.Z.); [email protected] (J.J.); Tel.: +86-574-8760-9489 (J.Z.) These authors contributed equally to this work.
Academic Editor: Andrzej Lewenstam Received: 3 January 2017; Accepted: 28 February 2017; Published: 12 March 2017
Abstract: A new strategy to discriminate four types of hazardous gases is proposed in this research. Through modulating the operating temperature and the processing response signal with a pattern recognition algorithm, a gas sensor consisting of a single sensing electrode, i.e., ZnO/In2 O3 composite, is designed to differentiate NO2 , NH3 , C3 H6 , CO within the level of 50–400 ppm. Results indicate that with adding 15 wt.% ZnO to In2 O3 , the sensor fabricated at 900 ◦ C shows optimal sensing characteristics in detecting all the studied gases. Moreover, with the aid of the principle component analysis (PCA) algorithm, the sensor operating in the temperature modulation mode demonstrates acceptable discrimination features. The satisfactory discrimination features disclose the future that it is possible to differentiate gas mixture efficiently through operating a single electrode sensor at temperature modulation mode. Keywords: Temperature modulation; PCA; discriminating features; In2 O3 ; ZnO
1. Introduction Currently, air pollutants such as nitrogen oxides (NO, NO2 ), carbon monoxide (CO) and hydrocarbons (HCs) derived from vehicles and industrial factories are a concerning issue, due to their effect of generating photochemical smog and airpocalypse [1]. Hence, strict regulations to reduce the emissions of these air pollutants have been implemented worldwide [2,3]. For the purpose of reducing the air pollutants efficiently and/or monitoring the air quality, high-performance gas sensors have to be developed and are considered crucial for improving the performance of combustion controllers and three-way catalysts [4,5]. To date, numerous types of gas sensors, which include but are not limited to the sensing principles of chemiresistive, amperometric, potentiometric and impedancimetric, have been reported [6–10]. Among them, yttria-stabilized zirconia (YSZ)-based electrochemical sensors consisting of a In2 O3 or ZnO sensing electrode (SE) exhibit great potential to sense NOx , CO, HCs and NH3 because of their reliable performance in harsh conditions [11–23]. However, overcoming their poor selectivity is still a challenging issue for these sensors. To promote selectivity, tuning the microstructure of the SEs, e.g., the morphology, thickness, and crystal phase, is usually considered, although this approach relies on sophisticated materials synthesis techniques [24,25]. Meanwhile, the strategy to design the Sensors 2017, 17, 573; doi:10.3390/s17030573
www.mdpi.com/journal/sensors
Sensors 2017, 17, 573
2 of 9
sensor array and combine it with specific data processing algorithms also shows promising results in differentiating gas mixtures [6,26,27]. Some examples of data processing algorithms include but are not limited to Fisher discriminant analysis, hierarchical clustering analysis, artificial neural networks, simple fuzzy logic and principal component analysis (PCA) [27–31]. PCA in multivariate statistical analysis is a method of compressing the dimension of vector data by extracting principal typical components from sample data set. It is a custom program, proposed for analyzing sensor output data [6,26–29]. Nevertheless, complex preparation techniques remarkably raise the cost of sensor arrays, thereby restraining the manufacturing of such devices on a large scale. Herein, we report a novel strategy to discriminate gas mixtures through modulating the operating temperature. A sensor comprised of a single SE is designed and its sensing characteristics are further investigated to verify the applicability of using the proposed discrimination method in future applications. 2. Materials and Methods 2.1. Sample Preparation and Fabrication of the Sensor Commercially available 8YSZ powder (8 % mol Y2 O3 added zirconia, Tosoh, Japan) were used as electrolyte substrate by tape casting. The slurry was prepared by a two-step ball milling procedure. In the first step, the ceramic powder was homogeneously dispersant in a roller ball mill for 24 h using zirconia balls, acrylic resin, xylene, and butyl acetate as milling medium, dispersant, and solvents, respectively. The polyvinyl butyral (PVB) was added as binder and polyethylene glycol (PEG) was used as plasticizer, followed by roller ball mill for another 6 h. The formula of slurry is showed in Table 1. Before tape casting, the slurries were vacuum pumped for 10 min to remove air. The 8YSZ film was casted on Mylar substrate at the blade height of 300 µm. After drying, the thickness of the green 8YSZ tape was maintained at 125 µm. In order to fabricate the thick electrolyte supported layer, six sheets of 8YSZ green tapes were laminated. The 8YSZ green layer was calcined at 1400 ◦ C for 4 h in air to form the 8YSZ ceramic substrate. Finally, the Length, width and thickness were 16, 9 and 0.6 mm, respectively. Table 1. The formula of electrolyte substrate slurry. Mill Media Mat. ZrO2
Initial wt. 800 g
Formula (g) acrylic resin 16
Butyl acetate 25.4
Xylene 25.4
8YSZ 200
PVB 20
PEG 10
To fabricate the sensor device, commercially available ZnO and In2 O3 (Sinopharm, Shanghai, China) powers were used to prepare the sensing electrodes. ZnO and In2 O3 powders were mixed with the weight ratio of 0, 5, 10, 15, 20, 30 and 100 wt.%, respectively. Afterwards, these powder mixtures were mixed with α-terpineol and screen-printed on the surface of the 8YSZ ceramic plate. These semi-manufactured samples were calcined at 800, 900, 1000 and 1100 ◦ C for 2 h in air, respectively to form the SE. Finally, MnO2 paste were screen-printed and calcined in the same way to form the Mn-based reference electrodes (RE). The photographic image of the fabricated sensor is shown in Figure 1a.
Sensors 2017, 17, 573 Sensors 2017, 17, 573
3 of 9 3 of 9 (a)
(b)
Figure (a)Photograph Photograph of fabricated the fabricated comprised In2Ocomposite 3/ZnO composite SE and Figure 1.1.(a) of the sensorsensor comprised of In2 O3of /ZnO SE and Mn-based Mn-based RE; (b) schematic view of the gas sensing characteristics evaluation apparatus. RE; (b) schematic view of the gas sensing characteristics evaluation apparatus.
2.2. 2.2. Characterization Characterization of of Sensing Sensing Characteristics Characteristics The sensing characteristics The gas gas sensing characteristics of of these these sensors sensors were were measured measured in in aa conventional conventional gas gas flow flow ◦ apparatus equipped with an electric furnace in the temperature range of 450–700 °C. Both SE and apparatus equipped with an electric furnace in the temperature range of 450–700 C. Both SE and RE RE are simultaneously exposed %O O2 in in aa balance balance of of N N2)) or or sample sample gas gas (NO (NO2,, CO, are simultaneously exposed to to the the base base gas gas (5 (5 vol. vol.% CO, 2 2 2 C reference C33H H66 and and NH NH33at at50–400 50–400ppm), ppm),respectively. respectively. The The voltages voltages between between sensing sensing electrode electrode and and reference electrode multifunction data electrode was was measured measured by by aa multifunction data acquisition acquisition card card (HP (HP 34970A, 34970A, Agilent, Agilent, Santa Santa Clara, Clara, CA, USA). The potential difference ΔV was subtracted V base gas from Vsample gas, where the Vsample gas CA, USA). The potential difference ∆V was subtracted V base gas from V sample gas , where the V sample gas represents thesensing sensingresponse response of the sensor toward sample gas the Vrepresents base gas represents the represents the of the sensor toward sample gas and theand V base the sensing gas sensing response of the sensor to base gas. Figure 1b demonstrates the schematic of sensing the gas response of the sensor to base gas. Figure 1b demonstrates the schematic view of view the gas sensing characteristics evaluation apparatus. PCA algorithm was writtenvia in-house via MATLAB® characteristics evaluation apparatus. PCA algorithm was written in-house MATLAB®R2009a ver. R2009a ver. 7.8.0 software (MathWorks© Inc., Natick, MA, USA). 7.8.0 software (MathWorks© Inc., Natick, MA, USA). 3. Results and Discussion In order ordertoto obtain sensitivity a desirable discriminatory a obtain highhigh sensitivity and aand desirable discriminatory capabilitycapability through athrough simplified simplified sensor configuration, namelycomposed a sensor composed single SE, thesensing following sensing sensor configuration, namely a sensor of a single of SE,a the following strategy is strategy is proposed: 2O3 composites in different ratios were synthesized and utilized as the proposed: ZnO/In2 OZnO/In composites in different ratios were synthesized and utilized as the SEs to 3 SEs to enhance the sensing magnitude. the performance and patterns responsetopatterns various enhance the sensing magnitude. Then, theThen, performance and response variousto examined ◦ examined gases for the sensor were evaluated and recorded in the temperature of 450–700 °C. gases for the sensor were evaluated and recorded in the temperature range ofrange 450–700 C. Finally, Finally, the obtained response patterns the sample processed PCA algorithm the obtained response patterns to the to sample gasesgases werewere processed withwith the the PCA algorithm to to investigate the differentiation efficiency. It was expected that due to the distinct difference in the investigate the differentiation efficiency. It was expected that due to the distinct difference gas-phase catalytic catalytic and electrocatalytic electrocatalytic activity activity of of the the SE to each target gas, response patterns to these four gases would vary with the change of the operating temperature. Since the PCA algorithm is a powerful discriminant analysis tool, particularly when dealing with response patterns with distinctive differences, these four gases were expected to be identified even if the examination was implemented through a sensor consisting of a single SE. Initially, the sensitivity and response magnitude of the sensor were optimized. Figure 2 shows the response signal (∆V) (ΔV) to 200 ppm sample gases of the sensors using various composite SEs. It can be concluded that the sensing behavior of the sensors operated at 600 ◦°C C is remarkably tuned by the amount of ZnO. For instance, after incorporating 5 wt. % ZnO ZnO with with In In22O33,, the wt.% the response response signal signal to to CO, CO, C33H 6 and NH 3 suddenly dropped, whereas the addition of 5 wt. % ZnO had a minor effect on the H6 and NH3 suddenly dropped, whereas the addition of 5 wt.% response value of NO22. Afterwards, Afterwards, with with continuously continuously increasing increasing the the ZnO amount, the response signal of the sensors to CO, C C33H66 and and NH33 increased increased gradually gradually and and reached reached the the maximum value at 15 wt.% wt. %ZnO. ZnO.Then, Then,the theresponse responsevalue valueto to these these three three target target gases gases started started to to decrease if the ratio of ZnO in the composite was further increased. Note that the sensing performance of the the sensor sensor to to NO NO22 was relatively less affected by incorporating ZnO when compared with its behavior with other three gases. Hence, in view of the highest sensing signal being observed for most of the examined gases by the addition of 15 wt. % ZnO, the sensor comprised of (In2O3 + 15 wt. % ZnO) composite SE was used for the following examination.
Sensors 2017, 17, 573
4 of 9
the addition of 15 wt.% ZnO, the sensor comprised of (In2 O3 + 15 wt.% ZnO) composite SE was used Sensors 2017, 17, 573 4 of 9 for the following examination.
Sensors 2017, 17, 573
4 of 9
Figure 2. Comparison of the characteristics forthe thesensors sensors comprised of various composite Figure 2. Comparison of sensing the sensing characteristics for comprised of various composite SEs (calcined at ◦1000 °C 2for h) 200 to 200 ppmsample sample gases, at 600 °C. ◦ C. SEs (calcined at 1000 C for h)2 to ppm gases,operated operated at 600 Then, the influence of the calcination temperature was investigated to further enhance the
Figure 2. Comparison of the sensing characteristics for the sensors comprised of various composite Then, the influence of the being calcination temperature was to further enhance the sensing sensing After calcined at temperatures in investigated the range 800–1100 °C (with intervals SEsmagnitude. (calcined at 1000 °C for 2 h) to 200 ppm sample gases, operated at 600of°C. ◦ (with intervals of 100 ◦ C), magnitude. being calcined at of temperatures in the range of 800–1100 of 100After °C), the cross-sensitivity the sensors with a SE containing 15 wt. % C ZnO was examined at ◦ 600 °C and the is shown Figure Interestingly, the response magnitude all further the examined the cross-sensitivity of theinsensors a SE containing 15 ZnOtowas examined atgases 600 Then, influence of the3.with calcination temperature waswt.% investigated to enhance the C and reached the maximum value when the sensor was fabricated at 900 °C. The ΔV measured at 600 °C sensing magnitude. After being calcined at temperatures in the to range 800–1100 °C (with intervals is shown in Figure 3. Interestingly, the response magnitude all of the examined gases reached the for 200°C), ppm NH3, C3H6 and 2sensors was about mV, −62.0215mV, mV and 30.08 mV, ◦ C for of 100 theCO, cross-sensitivity of NO thefabricated with a900 SE ◦containing wt. −97.4 % ZnO was examined at maximum value when the sensor was at−40.64 C. The ∆V measured at 600 200 ppm respectively. the3.optimal calcination temperature for thetosensor selected as 600 °C and isConsequently, shown in Figure Interestingly, the response magnitude all the was examined gases CO, NH900 about −40.64 mV, −ZnO 62.02and mV, −internal 97.4 mVconnection and 30.08between mV, respectively. 3, C 3 HRegarding 6 and NOthe 2 was °C. tuning mechanism of was the fabricated the900 reached the maximum value when the sensor at °C. The ΔV measured at◦ 600the °C Consequently, the optimal calcination temperature for the sensor was selected as 900 C. mV, Regarding calcination temperature and behavior, it is probably due to the−97.4 gas-phase catalytic and for 200 ppm CO, NH3, C 3H6 the andsensing NO2 was about −40.64 mV, −62.02 mV, mV and 30.08 the tuning mechanism of the and the electrode internal connection between calcination temperature electrocatalytic activity of ZnO thethe composite being closely dependent on its composition and respectively. Consequently, optimal calcination temperature for the the sensor was selected as microstructure. This will be thoroughly studied in the near future and will not be discussed in and the900 sensing behavior,the it is probably due toofthe catalytic electrocatalytic activity of °C. Regarding tuning mechanism thegas-phase ZnO and the internaland connection between this the work, since the main focus ofthe thissensing research is to confirm the feasibility themicrostructure. strategy incatalytic identifying calcination temperature and behavior, is probably due toof the gas-phase and will be the composite electrode being closely dependent onitits composition and This gas mixtures with a single SE. Afterwards, the dependence of the response value on the and gas electrocatalytic activity of the composite electrode being closely dependent on its composition thoroughly studied in the near future and will not be discussed in this work, since the main focus concentration also For brevity, list near the sensing behavior to be C3H 6 herein in as this the microstructure.was This willexamined. be thoroughly studiedwe in the future and will not discussed of this research is to confirmAs the feasibility of 4, the strategy in identifying gasvalid mixtures with representative shown in Figure response value of the sensor linearly witha single work, since theexample. main focus of this research is tothe confirm the feasibility of the strategy in identifying SE. Afterwards, the of dependence of the response value on sensor the gas concentration was also examined. the theaCsingle 3H6 concentration; sensitivity of the roughly −30.21 gas logarithm mixtures with SE. Afterwards, the dependence of was the responseestimated value onas the gas mV/Dec. Furthermore, can be seen from inserted in Figure that the sensor For brevity, we list the behavior C3 Hthe herein asillustration the representative As shown in concentration wassensing also itexamined. Forto brevity, list the sensing behavior to C4example. 3H 6 herein as the 6 we showed a quickexample. response rate, with a Figure 90% response time to 200 3Hsensor 6 of of about 17 s. However, representative As shown in 4,linearly the response value of C the valid with Figure 4, the response value of the sensor valid with theppm logarithm the Clinearly 3 H6 concentration; the sensor relatively low recovery rate even when operated theroughly temperature of 600as°C, e.g., logarithm ofa the C3Hroughly 6 concentration; sensitivity of the sensor at was estimated −30.21 sensitivity of the gave sensor was estimated as −30.21 mV/Dec. Furthermore, it can be seen from the 90% recovery time to 200 ppm C 3 H 6 was about 34 s, which needs further improvement in the mV/Dec. Furthermore, it can be seen from the inserted illustration in Figure 4 that the sensor the inserted illustration in Figure 4 that the sensor showed a quick response rate, with a 90% response future. showed a quick response rate, with a 90% response time to 200 ppm C3H6 of about 17 s. However, time to the 200sensor ppmgave C3 Ha6relatively of aboutlow 17 recovery s. However, the sensor gave a relatively low recovery rate even rate even when operated at the temperature of 600 °C, e.g., ◦ C, e.g., the 90% recovery time to 200 ppm C H was about when operated at the temperature of 600 3 in 6 the the 90% recovery time to 200 ppm C3H6 was about 34 s, which needs further improvement 34 s, which needs further improvement in the future. future.
Figure 3. Influence of the calcination temperature (800–1100 °C, with intervals of 100 °C) on the sensing behavior of the sensor comprised of (In2O3 + 15 wt. % ZnO)-SE, operated at 600 °C. Figure 3. Influence of the calcination temperature (800–1100 with intervals of 100◦ °C) on the Figure 3. Influence of the calcination temperature (800–1100 ◦ C,°C, with intervals of 100 C) on the sensing sensing behavior of the sensor comprised of (In2O3 + 15 wt. % ZnO)-SE, operated at 600 °C. behavior of the sensor comprised of (In2 O3 + 15 wt.% ZnO)-SE, operated at 600 ◦ C.
Sensors 2017, 17, 573 Sensors Sensors 2017, 2017, 17, 17, 573 573
5 of 9 55 of of 99
Figure 4. Dependence sensing signal signal (∆V) (ΔV) on the the logarithm logarithm of C H66 concentrations; Dependence of of the concentrations; Insert: Insert: Figure 4. Dependence of the sensing (ΔV) on logarithm of of C C333H concentrations; Insert: response transitions to C 3 H 6 in the concentration range of 50–400 ppm. H66 in inthe theconcentration concentrationrange range of of 50–400 50–400 ppm. ppm. response transitions to C3H
Theoretically, high discrimination efficiency can obtained through operating the sensor in Theoretically, highdiscrimination discriminationefficiency efficiencycan canbebe be obtained through operating sensor in Theoretically, high obtained through operating thethe sensor in the the temperature modulation mode, if distinct differences in the response patterns are observed. the temperature modulation mode, if distinct differences the response patterns are observed. temperature modulation mode, if distinct differences in theinresponse patterns are observed. Thus, Thus, the response patterns to the 50 ppm sample gas were recorded within the operational Thus, the response patterns to the 50 ppm sample gas were recorded within the operational the response patterns to the 50 ppm sample gas were recorded within the operational temperatures of temperatures of 450–700 °C for sensor using the 2O3 + 15 wt. % ZnO) SE. It can be concluded temperatures 450–700 for the the the (In (In 2O3 +SE. 15 It wt. % be ZnO) SE. It can be concluded 450–700 ◦ C forofthe sensor°C using thesensor (In2 O3using + 15 wt.% ZnO) can concluded from the results from the results summarized in Figure 5 that the sensor demonstrated apparent differences in from the results summarized in Figure 5 that the sensor demonstrated apparent differences in the the summarized in Figure 5 that the sensor demonstrated apparent differences in the response patterns to response patterns to each target gas within the examined temperature range. With the increasing response patterns to each target gas within the examined temperature range. With the increasing each target gas within the examined temperature range. With the increasing operational temperature, operational temperature, the response operational response value value of of NO NO22 decreased, decreased, whereas whereas the the other other three three sample sample gases gases the responsetemperature, value of NOthe 2 decreased, whereas the other three sample gases gave a more complex gave a more complex variation in their sensing behavior. Based on these results, it is reasonable to gave a more complex variation in their sensing behavior. Based on these results, it is reasonable to variation in their sensing behavior. Based on these results, it is reasonable to believe that desirable believe that desirable discriminating features will be obtained if processing these response patterns believe that desirable discriminating features will be obtained if processing these response patterns discriminating features will be obtained if processing these response patterns with the PCA algorithm. with the Consequently, a 20 × column) feature vector (shown in Table 2) with the PCA PCA algorithm. algorithm. Consequently, 20 × × 66 (row (row feature Table 2) Consequently, a 20 × 6 (row × column)afeature vector× column) (shown in Table vector 2) was(shown createdin and input was created and input into the PCA software. Each row represents a single measurement at aa was created and input into the PCA software. Each row represents a single measurement at into the PCA software. Each row represents a single measurement at a specific gas concentration specific gas concentration within examined temperature ranges each represents aa specificall gas concentration within all all the the examined temperature ranges and and each column column represents within the examined temperature ranges and each column represents a feature calculated for that feature calculated for that particular measurement at a specific temperature. All data obtained from feature calculated for thatatparticular a specific temperature. obtained from particular measurement a specificmeasurement temperature.atAll data obtained from All the data sensor operated at the sensor operated at six different temperatures is equivalent to that given by six different sensors the sensor operated at six different temperatures is equivalent to that given by six different sensors six different temperatures is equivalent to that given by six different sensors that were operated at that were operated the temperature. As aa result, aa virtual sensor array that wereoperational operated at at the same same operational operational Asarray result, virtual the same temperature. As a result, atemperature. virtual sensor composed of sixsensor sensorsarray was composed of six sensors was established and applied for sensing these four gases at the level of composed of six sensors was established and applied for sensing these four gases at the level of established and applied for sensing these four gases at the level of 50–400 ppm. A flowchart that 50–400 ppm. A flowchart that summarizes the procedure is shown in Figure 6. 50–400 ppm. A flowchart that summarizes the procedure is shown in Figure 6. summarizes the procedure is shown in Figure 6.
Figure 5. Response patterns of the the (50 ppm) at operating Figure 5. 5. Response Responsepatterns patterns of sensor the sensor sensor to the examined examined gases (50 ppm)operating at various various operating of the to the to examined gases (50gases ppm) at various temperatures. temperatures. temperatures.
Sensors 2017, 17, 573
6 of 9
Sensors 2017, 17, 573
6 of 9
Table 2. The 20 × 6 (row × column) feature vectors being created for PCA analysis.
Table 2. The 20 × 6 (row × column) feature vectors being created for PCA analysis.
Response at Various Operating Temperatures/mV
Sample Gas
Concentration/ppm
Sample Gas
CO CO
C3 H6 C3H6
NH3
NH3
Concentration/ppm
50 50 90 150 90 250150 400250
NO2
−28.80 −−28.80 35.55 −−35.55 49.86 −−49.86 59.84
50 400 90 50 150 90 250150 400250
−−59.84 64.00 −−64.00 85.79 −−85.79 95.33 −−95.33 101.95 −−101.95 113.98
400
−−113.98 11.85 −−11.85 20.89 −−20.89 27.35 −−27.35 40.60 −−40.60 50.66
50 90 50 150 90 250150 400250 400
NO2
Operating Temperature Response at Various Operating Temperatures/mV 500 ◦ C Operating 550 ◦ CTemperature 600 ◦ C 650 ◦ C 700 ◦ C 450 °C 500 °C 550 °C 600 °C 650 °C 700 °C −17.86 −17.58 −13.98 −19.32 −15.64 −3.36 −17.86 −17.58 −13.98 −19.32 −15.64 −3.36 450 ◦ C
50 50 90 90 150 150 250 250 400 400
−50.66
68.49 68.49 81.07 81.07 84.90 84.90 89.57 89.57 105.53
105.53
Sensor Sensor11
−29.52
−30.65
−32.47
−25.07
−8.19
−29.52 −38.18
−30.65 −38.89
−32.47 −42.59 −25.07 −29.41 −8.19 −13.59
−38.18 −48.76
−38.89 −47.51
−42.59 −49.79 −29.41 −37.41−13.59−20.62
−57.19 −48.76
−55.35 −47.51
−60.04 −37.41 −44.59−20.62−28.60 −49.79
−57.19 −55.35 −73.63 −56.55 −60.04 −72.98 −44.59 −44.70−28.60−18.36 −73.63 −56.55 −90.31 −69.14 −72.98 −88.84 −44.70 −81.54−18.36−49.13 −90.31 −69.14 −87.33 −74.55 −88.84 −88.39 −81.54 −85.39−49.13−56.04 −87.33 −74.55 −92.90 −83.55 −88.39 −98.40 −85.39 −94.02−56.04−65.11 −92.90 − 106.38 −83.55 −90.32 −98.40 −97.68 −94.02 −101.02−65.11−73.50 −106.38 −90.32 −22.04 −15.27 −97.68 −25.23 −101.02 −28.34−73.50−15.47 −22.04 −15.27 −25.23 −28.34 −31.12 −21.69 −32.46 −46.30−15.47−29.93 −31.12 −21.69 −32.46 −46.30 −37.71 −31.58 −50.86 −59.13−29.93−44.26 −37.71 −31.58 −47.75 −39.50 −50.86 −57.74 −59.13 −72.44−44.26−57.34 −47.75 −39.50 −59.53 −45.50 −57.74 −65.48 −72.44 −84.14−57.34−68.49 −59.53 −45.50 −65.48 −84.14 −68.49 69.45 54.76 48.34 27.39 14.27 69.45 54.76 48.34 27.39 14.27 80.73 57.42 58.15 41.94 27.01 80.73 57.42 58.15 41.94 27.01 86.86 56.96 60.37 49.97 39.47 86.86 56.96 60.37 49.97 39.47 88.81 54.94 61.28 56.48 50.57 88.81 54.94 61.28 56.48 50.57 73.11 49.51 56.35 61.79 63.01 73.11 49.51 56.35 61.79 63.01 Sensor22 Sensor Sensor Sensor Sensor 5 Sensor 6 Sensor 3 3 Sensor 4 4Sensor 5 Sensor 6 Virtual Sensor Array Virtual Sensor Array
Figure 6. Flowchart temperature Figure 6. Flowchart of of processing processing the the response response pattern pattern that that was was obtained obtained from from the the temperature modulation mode with PCA software in detail. modulation mode with PCA software in detail.
The PCA transformation for all response patterns is shown in Figure 7, with the same colored The PCA transformation for all response patterns is shown in Figure 7, with the same colored symbol corresponding to each target gas. The spatial distance between each target gas represents the symbol corresponding to each target gas. The spatial distance between each target gas represents discrimination capability of the sensor. The overlap of these symbols shown in the PCA scheme the discrimination capability of the sensor. The overlap of these symbols shown in the PCA scheme implies low differentiating characteristics of the sensor when sensing these symbols representing implies low differentiating characteristics of the sensor when sensing these symbols representing gas. In contrast, a large spatial distance between these symbols suggests a desirable discrimination gas. In contrast, a large spatial distance between these symbols suggests a desirable discrimination feature detected by the sensor array. The results demonstrated in the PCA scheme suggest that the feature detected by the sensor array. The results demonstrated in the PCA scheme suggest that the sensor exhibits the best discrimination capability to NO2 and acceptable differentiating features to sensor exhibits the best discrimination capability to NO2 and acceptable differentiating features to CO, CO, NH3 and C3H6. This is in line with the sensing characteristics shown in Figure 5. As shown in Figure 5, the sensor gave relatively similar response patterns to CO, NH3 and C3H6 within the examined temperature range, whereas distinctive sensing behavior was observed for NO2, thereby leading to the satisfactory differentiating feature for NO2 and less discriminating capability for CO,
Sensors 2017, 17, 573
7 of 9
NH3 and C3 H6 . This is in line with the sensing characteristics shown in Figure 5. As shown in Figure 5, the sensor gave relatively similar response patterns to CO, NH3 and C3 H6 within the examined temperature range, Sensors 2017, 17, 573 whereas distinctive sensing behavior was observed for NO2 , thereby leading 7 of 9 to the satisfactory differentiating feature for NO2 and less discriminating capability for CO, NH3 and C3 H6 . 3 and C3H6. Note that the preliminary success of applying the temperature modulation approach NoteNH that the preliminary success of applying the temperature modulation approach together with together with the PCA algorithm in discriminating gas mixtures gives a potential method to replace the PCA algorithm in discriminating gas mixtures gives a potential method to replace the complex the complex sensor array with a simplified sensor configuration. In particular, it is reasonable to sensor array with a simplified sensor configuration. In particular, it is reasonable to expect that the expect that the discrimination features will be further improved if the SE exhibits distinctive discrimination features will further improved the efforts SE exhibits distinctive response patterns to all response patterns to all thebe target gases, implyingifmore are needed to develop an appropriate the target gases, implying more efforts are needed to develop an appropriate SE in future research. SE in future research.
Figure 7. PCA schemes ofof the sensorutilizing utilizing O 15+wt. 15 % wt.% ZnO)-SE, operated in Figure 7. PCA schemes thedata dataset setfor for the the sensor (In(In 2O23 + 3 ZnO)-SE, operated in the temperature modulation mode. the temperature modulation mode.
4. Conclusions 4. Conclusions Through fabricatingaacomposite composite SE operating it initthe modulation mode, themode, Through fabricating SEand and operating intemperature the temperature modulation sensor consisting of the single SE exhibited an acceptable performance in sensing four types of the sensor consisting of the single SE exhibited an acceptable performance in sensing four types of hazardous gases. With the addition of 15 wt. % ZnO to In2O3 and being calcined at 900 °C for 2 h, the hazardous gases. With the addition of 15 wt.% ZnO to In2 O3 and being calcined at 900 ◦ C for 2 h, sensor gave its maximum sensing magnitude to all the examined gases. Moreover, with the aid of the sensor gave its maximum sensing magnitude to all the examined Moreover, with the aid the PCA algorithm, acceptable differentiating features were observedgases. through the temperature of the PCA algorithm, differentiating features were observed through the temperature modulation approach.acceptable Such promising results indicate a novel strategy to sense gas mixtures with a modulation Suchsensor promising results indicate a novel strategy to sense gas mixtures with simplifiedapproach. potentiometric configuration.
a simplified potentiometric sensor configuration. Acknowledgments: This work was financially sponsored by Kuan Cheng Wong Magna Fund at Ningbo University and supported by the Research Foundation for Startup Researcher (ZX2016000095), Acknowledgments: This work wasScientific financially sponsored by Kuan Cheng Wong Magna Fund atthe Ningbo Nationaland Natural Science Foundation of China Research (61301050, 61471210 and 51472126). University supported by the Scientific Foundation for Startup Researcher (ZX2016000095), the National Natural Science Foundation of China (61301050, 61471210 and 51472126). Author Contributions: Jie Zou and Han Jin conceived and designed the experiments; Fu-an Li performed the
Author Contributions: JieJinxia Zou Wang and Han Jin conceived and designed the experiments; Fu-an performed the experiments; Fu-an Li, and Jiawen Jian analyzed the data; Jinxia Wang and Jiawen JianLi contributed experiments; Fu-an Li, Jinxia Wang and Jiawen JianJin analyzed data; Jinxia Wang and Jiawen Jian contributed reagents/materials/analysis tools; Jie Zou and Han wrote thethe paper. reagents/materials/analysis tools; Jie Zou and Han Jin wrote the paper. Conflicts of Interest: The authors declare no conflict of interest.
Conflicts of Interest: The authors declare no conflict of interest.
Abbreviations
Abbreviations
The following abbreviations are used in this manuscript:
The following abbreviations are used in this manuscript: YSZ
Yttria-Stabilized Zirconia
Sensing Electrode YSZ SE Yttria-Stabilized Zirconia Reference Electrode SE RE Sensing Electrode principleElectrode component analysis RE PCAReference PCA principle component analysis
Sensors 2017, 17, 573
8 of 9
References 1. 2. 3.
4. 5. 6. 7. 8. 9. 10.
11. 12. 13. 14. 15. 16.
17. 18. 19. 20. 21. 22. 23.
Karthikeyan, S.; Prathima, A.; Sabariswaran, K. An environmental effect of nano additive on performance and emission in a biofuel fuelled marine engine. Indian J. Geo. Mar. Sci. 2015, 44, 896–901. Tzamkiozis, T.; Ntziachristos, L.; Samaras, Z. Diesel passenger car pm emissions: From euro 1 to euro 4 with particle filter. Atmos. Environ. 2010, 44, 909–916. [CrossRef] O’Driscoll, R.; ApSimon, H.M.; Oxley, T.; Molden, N.; Stettler, M.E.J.; Thiyagarajah, A. A portable emissions measurement system (pems) study of nox and primary NO2 emissions from euro 6 diesel passenger cars and comparison with copert emission factors. Atmos. Environ. 2016, 145, 81–91. [CrossRef] Zhang, H.; Wang, J.M.; Wang, Y.Y. Sensor reduction in diesel engine two-cell selective catalytic reduction (scr) systems for automotive applications. IEEE-ASME Trans. Mechatron. 2015, 20, 2222–2233. [CrossRef] Simons, H.; Simon, U. Zeolites as nanoporous, gas-sensitive materials for in situ monitoring of denox-scr. Beilstein J. Nanotechnol. 2012, 3, 667–673. [CrossRef] [PubMed] Hannon, A.; Lu, Y.; Li, J.; Meyyappan, M. A sensor array for the detection and discrimination of methane and other environmental pollutant gases. Sensors 2016, 16, 1163. [CrossRef] [PubMed] Chiu, S.W.; Tang, K.T. Towards a chemiresistive sensor-integrated electronic nose: A review. Sensors 2013, 13, 14214–14247. [CrossRef] [PubMed] Miura, N.; Sato, T.; Anggraini, S.A.; Ikeda, H.; Zhuiykov, S. A review of mixed-potential type zirconia-based gas sensors. Ionics 2014, 20, 901–925. [CrossRef] Rheaume, J.M.; Pisano, A.P. A review of recent progress in sensing of gas concentration by impedance change. Ionics 2011, 17, 99–108. [CrossRef] Zou, J.; Zheng, Y.; Li, J.; Zhan, Z.; Jian, J. Potentiometric NO2 sensors based on thin stabilized zirconia electrolytes and asymmetric (La0.8 Sr0.2 )0.95 MnO3 electrodes. Sensors 2015, 15, 17558–17571. [CrossRef] [PubMed] Lu, G.; Miura, N.; Yamazoe, N. High-temperature hydrogen sensor based on stabilized zirconia and a metal oxide electrode. Sens. Actuators B Chem. 1996, 35–36, 130–135. [CrossRef] Yamaura, H.; Jinkawa, T.; Tamaki, J.; Moriya, K.; Miura, N.; Yamazoe, N. Indium oxide-based gas sensor for selective detection of CO. Sens. Actuators B Chem. 1996, 36, 325–332. [CrossRef] Ueda, T.; Plashnitsa, V.V.; Nakatou, M.; Miura, N. Zirconia-based amperometric sensor using zno sensing-electrode for selective detection of propene. Electrochem. Commun. 2007, 9, 197–200. [CrossRef] Jin, H.; Breedon, M.; Miura, N. Sensing behavior of YSZ-based amperometric NO2 sensors consisting of Mn-based reference-electrode and In2 O3 sensing-electrode. Talanta 2012, 88, 318–323. [CrossRef] [PubMed] Lee, I.; Jung, B.; Park, J.; Lee, C.; Hwang, J.; Park, C.O. Mixed potential NH3 sensor with LaCoO3 reference electrode. Sens. Actuators B Chem. 2013, 176, 966–970. [CrossRef] Dai, L.; Liu, Y.; Meng, W.; Yang, G.; Zhou, H.; He, Z.; Li, Y.; Wang, L. Ammonia sensing characteristics of La10 Si5 MgO26 -based sensors using In2 O3 sensing electrode with different morphologies and cuo reference electrode. Sens. Actuators B Chem. 2016, 228, 716–724. [CrossRef] Liu, F.; Guan, Y.; Sun, H.; Xu, X.; Sun, R.; Liang, X.; Sun, P.; Gao, Y.; Lu, G. Ysz-based NO2 sensor utilizing hierarchical In2 O3 electrode. Sens. Actuators B Chem. 2016, 222, 698–706. [CrossRef] Miura, N.; Shiraishi, T.; Shimanoe, K.; Yamazoe, N. Mixed-potential-type propylene sensor based on stabilized zirconia and oxide electrode. Electrochem. Commun. 2000, 2, 77–80. [CrossRef] Miura, N.; Raisen, T.; Lu, G.; Yamazoe, N. Highly selective CO sensor using stabilized zirconia and a couple of oxide electrodes. Sens. Actuators B Chem. 1998, 47, 84–91. [CrossRef] Sun, F.; Li, X.; Liu, L.; Wang, J. Novel Zn–M–O (M = Sn, Co) sensing electrodes for selective mixed potential CO/C3 H8 sensors. Sens. Actuators B Chem. 2013, 184, 220–227. [CrossRef] Li, X.; Kale, G.M. Influence of thickness of ito sensing electrode film on sensing performance of planar mixed potential CO sensor. Sens. Actuators B Chem. 2006, 120, 150–155. [CrossRef] Chen, Y.; Xiao, J. Effects of ysz addition on the response of La2 CuO4 sensing electrode for a potentiometric NOx sensor. Sens. Actuators B Chem. 2014, 192, 730–736. Li, X.; Wang, C.; Wang, B.; Yuan, Y.; Huang, J.; Zhang, H.; Xia, F.; Xiao, J. Effects of sintering temperature on the NH3 sensing properties of Mg2 Cu0.25 Fe1 O3.75 electrode for ysz-based potentiometric NH3 sensor. Ceram. Int. 2016, 42, 2214–2220. [CrossRef]
Sensors 2017, 17, 573
24. 25. 26. 27.
28.
29. 30.
31.
9 of 9
Hu, X.; Tian, L.; Sun, H.; Wang, B.; Gao, Y.; Sun, P.; Liu, F.; Lu, G. Highly enhanced NO2 sensing performances of Cu-doped In2 O3 hierarchical flowers. Sens. Actuators B Chem. 2015, 221, 297–304. [CrossRef] Zhang, D.; Liu, Z.; Li, C.; Tang, T.; Liu, X.; Han, S.; Lei, B.; Zhou, C. Detection of NO2 down to ppb levels using individual and multiple In2 O3 nanowire devices. Nano Lett. 2004, 4, 1919–1924. [CrossRef] Tomchenko, A.A.; Harmer, G.P.; Marquis, B.T.; Allen, J.W. Semiconducting metal oxide sensor array for the selective detection of combustion gases. Sens. Actuators B Chem. 2003, 93, 126–134. [CrossRef] Natale, C.D.; Davide, F.A.M.; Amico, A.D.; Sberveglieri, G.; Nelli, P.; Faglia, G.; Perego, C. Complex chemical pattern recognition with sensor array the discrimination of vintage years of wine. Sens. Actuators B Chem. 1995, 25, 801–804. [CrossRef] Sheini, A.; Khajehsharifi, H.; Shahbazy, M.; Kompany-Zareh, M. A chemosensor array for the colorimetric identification of some carboxylic acids in human urine samples. Sens. Actuators B Chem. 2017, 242, 288–298. [CrossRef] Li, D.; Lei, T.; Zhang, S.; Shao, X.; Xie, C. A novel headspace integrated E-nose and its application in discrimination of chinese medical herbs. Sens. Actuators B Chem. 2015, 221, 556–563. [CrossRef] Zhang, L.; Tian, F.; Peng, X.; Dang, L.; Li, G.; Liu, S.; Kadri, C. Standardization of metal oxide sensor array using artificial neural networks through experimental design. Sens. Actuators B Chem. 2013, 177, 947–955. [CrossRef] Zampolli, S.; Elmi, I.; Ahmed, F.; Passini, M.; Cardinali, G.C.; Nicoletti, S.; Dori, L. An electronic nose based on solid state sensor arrays for low-cost indoor air quality monitoring applications. Sens. Actuators B Chem. 2004, 101, 39–46. [CrossRef] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
