The XRD patterns of as-synthesized CTO fibers were calcined at various temperatures between 700°C and 1000°C to investigate the evolution of the crystalline phases as shown in Fig. 1. Sharp peaks in Fig. 1 reflect the formation of a well-ordered crystal at 1000°C.The patterns of the CTO were identical to that of Cr2O3 (JCPDS File no.38–1479), and no peaks corresponding to TiO2 or CTO are observed, which may due to the small mass percent of TiO2 in CTO fibers. Formation and solid solution behavior of TiO2 in Cr2O3 are expected when both Cr3+ and Ti4+ ions have the same ionic radius (0.061 nm). With the help of Scherrer equation, as calculated by the relatively large diffraction peaks assigning to (012), (104), (110), and (116) planes, the average crystallite size of the CTO samples are about 27, 32, 38, and 42 nm, respectively.
Features of the reticular structure of fibers were also examined by SEM. Figure 2 shows SEM micrographs of a typical structure of the CTO fibers: (i) as-synthesized composite fibers and (ii) the calcined fibers, the inset is the CTO fibers with high amplification. It can be observed that the distributions of fibers are fairly random with no distinct alignment. The surface of the hybrid fibers is smooth and the diameters are around 500 nm. After the calcination process, the diameters of fibers shrink significantly to ~200 nm with the removal of organic phase, and the surface of the fibers becomes coarse, which consisted of nanocrystallines.
Figure 2. SEM micrographs of (a) as-synthesized composite fibers and (b) the calcined fibers of CTO, the inset is the CTO fibers with high amplification.
Download figure to PowerPoint
The EPMA study of CTO showed that the content of the as-prepared CTO is Cr1.93Ti0.07O2.7. Table 1 lists the compositions of the as-prepared fibers. We can see from Table 1 that the as-prepared fibers generate an abundance of oxygen vacancies in these samples, which is responsible for the active sites to reversibly adsorb oxygen. As the gas-sensing catalytic activity is related to the surface adsorbed oxygen, it can be inferred that the catalytic activity is controlled by the nonstoichiometric character of the gas-sensing material. Hence, the nonstoichiometric “CTO” could enhance and improve the activity of catalytic reaction on its surface.
Table 1. Composition of As-synthesized CTO Fibers
Here, the investigation focuses on the ethanol-sensing properties of the as-synthesized CTO fibers. The behavior of the gas-sensing properties are determined by the interactions of oxygen ions and gas molecules around the surface oxygen vacant sites of the sensing material. These reactions' processes can be expressed as follows:
where R- represents the carbonic long chain. The gas response is believed to be due to surface catalytic reaction of ethanol with the O− forming a surface hydroxyl species and an unstable alkoxyl radical as an intermediate product, which releases the surface-trapped electrons, then oxidizes to water and carbon dioxide at the working temperature of 400°C.
The responses of the CTO fibers tested in 1–1000 ppm of ethanol are presented in Fig. 3. The operating temperature of the sensor was 400°C. In our measurement, the response time and recovery time are about 110 and 200 s to 50 ppm ethanol, respectively.
Figure 3. Response of the CTO fibers versus ethanol concentration, the inset is the response and recovery characteristics curves in the range of 1–25 ppm.
Download figure to PowerPoint
The experimental data could be fitted by the sensor response law developed by Gurlo et al. The correlation between the sensor response (S) and gas concentration (C) is approximated by an equation in a form of S = 1 + aCbin this report, where a and b are variables, that can be fitted to a = (0.2144 ± 0.00307) and b = (0.6186 ± 0.0022), respectively.
The sensor noise can be calculated using the residual between experimental data and the corresponding curve-fitting value using the root-mean-square deviation (rmsd).
Where yi is the measured data point, y is the corresponding value calculated from the curve-fitting equation, and N is the number of data points used in the curve fitting. The sensor noise is 0.0315 for the ethanol sensor in Fig. 3. According to the IUPAC definition, when the signal-to-noise ratio equals 3, the signal is considered to be a true signal.
Using the above equations, the minimum response is calculated to be 1.095 and the ethanol detection limit is about 0.268 ppm in Fig. 3.
The cross-response (selectivity) of CTO sensors to formaldehyde, ethanol, ammonia, and isopropanol were examined as shown in Fig. 4. The results indicate that the CTO sensors are more sensitive to ethanol and isopropanol than to formaldehyde and ammonia. The sensors show almost the same response to ethanol and isopropanol, and the main reason relating to this phenomenon may be that ethanol and isopropanol have the same functional group (-OH) involved in the redox reaction. The different activity energy for the interaction between oxygen ions on the surface of CTO fibers with diverse functional groups of gas molecules is predictable. As proposed in reference, to obtain high selectivity of gas-sensing materials among a variety of gas vapors, a combination of the sensor process and catalyst design approach should be developed. Further investigations are needed in this respect to achieve the selective sensing properties of the CTO fibers.
In addition, it was known that the sensing performance of gas-sensing materials was considerably influenced by the humidity of the environment. Figure 5 shows the change in the base resistance and the response of gas sensor-based CTO fibers to ethanol at 15% RH and 50% RH, respectively. In general, surface resistance of metal oxide-sensing materials is influenced by the presence of moisture as the preadsorbed oxygen on the surface can be displaced by competitive water physisorption. For the p-type CTO fibers, an increase in RH led to an increase in resistance. The relative variation in base resistance was found to be 4.75% and the sensor relative response decreased by a factor of 13.38% and 7.45% to 50 and 100 ppm ethanol when the humidity varied from 15% RH to 50% RH, respectively. However, even after reduction, the relative response of the sensor was high enough for a good sensor categorization, and the response and recovery characteristics remained unchanged.
Figure 5. Resistance of the CTO fibers to different concentrations of ethanol in humid air of 15% RH and 50% RH at 400°C.
Download figure to PowerPoint
Figure 6 gives the response transient curves of two CTO sensors to 100 ppm ethanol of three cycles in 20% RH. We can see that two sensors have good consistency and reproducibility in three response–recovery cycles.
The sensor stability test was performed every 5 days for 30 days as shown in Fig. 7. There are almost no changes in the response values, which suggests that these sensors could be operated to continuously monitor ethanol. The better stability of the fiber mats sensors is explained in reticular framework of reduced propensity of fibers to sinter at elevated temperature.