Fabrication of Nanosized Island-Like CdO Crystallites-Decorated TiO2 Rod Nanocomposites via a Combinational Methodology and Their Low-Concentration NO2 Gas-Sensing Behavior

TiO2–CdO composite rods were synthesized through a hydrothermal method and sputtering thin-film deposition. The hydrothermally derived TiO2 rods exhibited a rectangular cross-sectional crystal feature with a smooth surface, and the as-synthesized CdO thin film exhibited a rounded granular surface feature. Structural analyses revealed that the CdO thin film sputtered onto the surfaces of the TiO2 rods formed a discontinuous shell layer comprising many island-like CdO crystallites. The TiO2–CdO composite rods were highly crystalline, and their surfaces were rugged. A comparison of the NO2 gas-sensing properties of the CdO thin film, TiO2 rods, and TiO2–CdO composite rods revealed that the composite rods exhibited superior gas-sensing responses to NO2 gas than did the CdO thin film and TiO2 rods, which can be attributed to the microstructural differences and the formation of heterojunctions between the TiO2 core and CdO crystallites.


Introduction
Gas sensors based on oxide semiconductors, such as TiO 2 , have the potential to detect harmful and toxic gases with high sensitivity [1][2][3][4]. TiO 2 nanostructures have a large surface-to-volume ratio because of their one-dimensional structure, which is advantageous in providing high surface effects between the oxide surface and the detected gases. Several studies have investigated methods to fabricate TiO 2 -based one-dimensional structures with various morphologies for application as gas-sensing devices. For example, the high acetone gas-sensing response of electrospinning-synthesized TiO 2 nanorods was investigated at 500 • C [5]. TiO 2 nanotubes synthesized through anodization of Ti foil at room temperature were used to detect H 2 gas [6]. TiO 2 nanowires were synthesized by subjecting sol-gel derived solid thin films to a photolithographic process for application in visible ethanol vapor sensing [7]. TiO 2 nanofibers are formed via the reaction of dense polycrystalline TiO 2 in a H 2 /N 2 environment at an elevated temperature and are investigated as gas sensing materials [8]. Moreover, gas sensors made from nano-heterostructures are shown to be a promising approach to enhance the gas-sensing response of the oxide semiconductors [9]. To further improve the gas-sensing performance of sensors composed of pure TiO 2 nanostructures, TiO 2 nanostructures on the surfaces of these sensors can be modified by using binary oxides such as ZnO, SnO 2 , and In 2 O 3 . Many core-shell structures of TiO 2 -based nanomaterials have been synthesized, and their applications in gas sensing have been investigated. ZnO sheet-decorated TiO 2 fibers, synthesized through a combination of the electrospinning technique and the hydrothermal growth method, exhibited a response of 15.7 to 100 ppm ethanol at 280 • C, which is much higher than that of pure TiO 2 and ZnO [10]. Heterostructures comprising SnO 2 nanoparticles deposited on TiO 2 nanobelts through a hydrothermal method exhibited high gas-sensing sensitivity for acetone vapor because of their large surface area [11]. Porous In 2 O 3 /TiO 2 composite nanofibers synthesized through a facile electrospinning-based synthesis method, followed by appropriate thermal treatment under ambient conditions, exhibited higher gas-sensing responses to NO x gas than those of pure TiO 2 nanofibers [12].
Few studies have used CdO-another promising gas-sensing binary oxide semiconductor-to decorate the surfaces of TiO 2 for improving its gas-sensing performance toward various toxic gases. CdO has a bandgap in the visible-light region, and its native oxygen vacancies make it an n-type semiconductor with low resistance [13]. CdO thin films and nanostructures have been synthesized and examined through various physical and chemical techniques [14][15][16]. For example, the gas-sensing properties of chemical bath-deposited CdO thin films were studied for sensing liquefied petroleum gas at an operating temperature of 300 • C [17]. Highly crystalline CdO nanostructures prepared through microwave-assisted growth were investigated as the sensing layer in resistive sensors and tested for NO 2 sensing [18]. Recently, gas sensors fabricated using composite materials incorporating CdO have been shown to exhibit superior gas-sensing performance than do the constituent individual oxides. For example, two-phase mixed CdO-MnO 2 thin films enhanced the performance of the constituent individual oxide films when sensing reduced gases [19]. Moreover, the formation of CdO-ZnO heterostructures improved the applicability of CdO in gas sensors [20]. Incorporating CdO into one-dimensional TiO 2 structures is promising for improving the gas-sensing response of one-dimensional TiO 2 -based sensors. However, limited studies have investigated the gas-sensing properties of one-dimensional TiO 2 -CdO composites. Therefore, in this study, island-like CdO crystallites were decorated onto TiO 2 rods to form a core-shell structure, and the correlation between the microstructure and the gas-sensing performance of TiO 2 -CdO composite rods for sensing low-concentration NO 2 gas was investigated.

Materials and Methods
In this study, TiO 2 rods sputtering deposited a 100-nm-thick CdO thin film and were used as samples for investigating the structure-dependent NO 2 gas-sensing response of the novel TiO 2 -CdO composite rods. The TiO 2 rods were grown on glass substrates using a hydrothermal method in this study. An amount of 7.4 mL of DI water was mixed with 12.6 mL of concentrated HCl (35%) in a Teflon-lined digestion autoclave. The mixture was stirred at ambient conditions for 5 min and after that, 0.25 mL of TiCl 4 was added into the mixed solution for preparation of the TiO 2 rods. The hydrothermal synthesis was conducted at 180 • C for 3 h. After the synthesis reaction, the autoclave was cooled down to room temperature. Then the samples were removed, washed with DI water, and air dried. TiO 2 -CdO composite rods were fabricated by sputtering CdO thin films onto the surfaces of the hydrothermally derived TiO 2 rod templates. The CdO thin films were fabricated through RF magnetron sputtering using a CdO target in mixed Ar/O 2 ambient with a ratio of 20/5 at 350 • C. The working pressure during thin-film deposition was fixed at 2.67 Pa, and the RF sputtering power was fixed at 60 W.
Crystal structures of the as-synthesized samples were investigated by X-ray diffraction (XRD; D2 PHASER, Bruker, Karlsruhe, Germany) using Cu Kα radiation. The surface morphologies of the samples were characterized by scanning electron microscopy (SEM; S-4800, Hitachi, Tokyo, Japan). High-resolution transmission electron microscopy (HRTEM; Tecnai F20 G2, Philips, Amsterdam, The Netherlands) was used to investigate the detailed microstructures of the TiO 2 -CdO composite rods. The attached energy dispersive X-ray spectroscopy (EDS) was used to investigate the composition of the TiO 2 and TiO 2 -CdO rod samples. The analyses of the transmittance and reflectance spectra of various samples were conducted by using a UV-Vis spectrophotometer (V750, Jasco, Tokyo, Japan). An X-ray photoelectron spectroscopy (XPS; ULVAC-PHI XPS, ULVAC, Chigasaki, Japan) analysis was used to determine the chemical binding status of the constituent elements of the samples. The gas sensing system used in this study consists of a gas flow system consisting of two lines. One is for a cylinder containing the background gas N 2 ; the other one was connected to a cylinder containing diluted NO 2 gas. The NO 2 concentration for the gas-sensing tests was varied from 1.0 ppm to 5.0 ppm by controlling the relative flow rates of background gas and target gas through mass flow controllers. Before conducting the gas-sensing tests, the test chamber was pumped to a vacuum environment. Subsequently, highly pure nitrogen gas at a consistent flow rate of 1000 sccm was introduced into the test chamber for 10 min; thereafter, the sample holder was heated to the desired temperatures using a direct heating approach. After the electric resistance of the sensors was stabilized, NO 2 gas at concentrations of 1.0, 2.5, and 5.0 ppm were introduced into the test chamber to investigate the NO 2 gas-sensing performance of the samples. The gas-sensing response of the sensors made from the CdO film, TiO 2 rods and TiO 2 -CdO rods to NO 2 gas is defined as the Rg/Ra. Ra is the sensor electrical resistance in the absence of a target gas, and Rg is that in the target gas. The measurements of a gas-sensing response were conducted with various concentrations of NO 2 gas. The 100 ppm CO, H 2 , and NH 3 gases were used for a gas-sensing selectivity test. The gas-sensing responses for these reducing gases are defined as Ra/Rg. Figure 1a shows the surface topography of a CdO film. The surface was rough and comprised many rounded, surface granular crystallites with a diameter of 20-45 nm. Figure 1b,c depict the morphologies of the TiO 2 and TiO 2 -CdO composite rods. The TiO 2 rods were dense and homogeneously distributed over the substrate. The pure TiO 2 rods exhibited a smooth surface (inset, Figure 1b). The diameter of the TiO 2 rods ranged from 100 to 140 nm, and the tip of the TiO 2 rods exhibited a rectangular feature. By contrast, the TiO 2 rods coated with CdO shell layers exhibited a rugged surface feature and high surface roughness (Figure 1c). High-magnification SEM revealed that the surfaces of the TiO 2 -CdO composite rods were comprised of many small, irregular-shaped, and clustered crystallites (inset, Figure 1c). is for a cylinder containing the background gas N2; the other one was connected to a cylinder containing diluted NO2 gas. The NO2 concentration for the gas-sensing tests was varied from 1.0 ppm to 5.0 ppm by controlling the relative flow rates of background gas and target gas through mass flow controllers. Before conducting the gas-sensing tests, the test chamber was pumped to a vacuum environment. Subsequently, highly pure nitrogen gas at a consistent flow rate of 1000 sccm was introduced into the test chamber for 10 min; thereafter, the sample holder was heated to the desired temperatures using a direct heating approach. After the electric resistance of the sensors was stabilized, NO2 gas at concentrations of 1.0, 2.5, and 5.0 ppm were introduced into the test chamber to investigate the NO2 gas-sensing performance of the samples. The gas-sensing response of the sensors made from the CdO film, TiO2 rods and TiO2-CdO rods to NO2 gas is defined as the Rg/Ra. Ra is the sensor electrical resistance in the absence of a target gas, and Rg is that in the target gas. The measurements of a gas-sensing response were conducted with various concentrations of NO2 gas. The 100 ppm CO, H2, and NH3 gases were used for a gas-sensing selectivity test. The gas-sensing responses for these reducing gases are defined as Ra/Rg. Figure 1a shows the surface topography of a CdO film. The surface was rough and comprised many rounded, surface granular crystallites with a diameter of 20-45 nm. Figures 1b-c depict the morphologies of the TiO2 and TiO2-CdO composite rods. The TiO2 rods were dense and homogeneously distributed over the substrate. The pure TiO2 rods exhibited a smooth surface (inset, Figure 1b). The diameter of the TiO2 rods ranged from 100 to 140 nm, and the tip of the TiO2 rods exhibited a rectangular feature. By contrast, the TiO2 rods coated with CdO shell layers exhibited a rugged surface feature and high surface roughness (Figure 1c). High-magnification SEM revealed that the surfaces of the TiO2-CdO composite rods were comprised of many small, irregular-shaped, and clustered crystallites (inset, Figure 1c).   Figure 2c presents the XRD pattern of the TiO2 rods decorated with CdO crystallites through sputtering deposition, in which the Bragg reflection centered at approximately 33.02° can be ascribed to cubic CdO(111); the appearance of this reflection was consistent with the XRD findings for the CdO thin film, which indicated that the sputtering-deposited CdO thin film had a highly (111)-oriented crystallographic feature.   Figure 2c presents the XRD pattern of the TiO 2 rods decorated with CdO crystallites through sputtering deposition, in which the Bragg reflection centered at approximately 33.02 • can be ascribed to cubic CdO(111); the appearance of this reflection was consistent with the XRD findings for the CdO thin film, which indicated that the sputtering-deposited CdO thin film had a highly (111)-oriented crystallographic feature. Figure 3a is a low-magnification TEM image of a single TiO2-CdO composite rod. As shown in the figure, the surface morphology of the shell layer indicated that this layer was comprised of many discontinuous island-like crystallites, and TiO2-CdO exhibited a rugged surface feature. The thickness distribution of the island-like CdO shell layer over the CdO crystallites was inhomogeneous and ranged from 15 to 30 nm. Notably, the thickness of the CdO shell layer on the TiO2 rod surface was less than the original two-dimensional 100-nm-thick CdO film; this was because the deposition of the two-dimensional continuous CdO film onto the three-dimensional free-standing TiO2 rods engendered a large surface dispersion effect and decreased the coverage thickness of the CdO on the TiO2 surface [21].   Figure 4a shows the optical transmittance spectrum of the CdO film. The transmittance of the film substantially decreased in the wavelength range of 500-600 nm, indicating that the CdO film has a bandgap in the visible-light region. The optical bandgap of the CdO film was evaluated using the Tauc model [22]. The inset in Figure 4a illustrates the plot of (ahv) 2 vs. hv, where a is the absorption coefficient and hv is the photon energy. The bandgap of the CdO film, calculated by extrapolating the linear portion of the curve until it intercepted the horizontal axis (photon energy), was approximately Figure 3a is a low-magnification TEM image of a single TiO 2 -CdO composite rod. As shown in the figure, the surface morphology of the shell layer indicated that this layer was comprised of many discontinuous island-like crystallites, and TiO 2 -CdO exhibited a rugged surface feature. The thickness distribution of the island-like CdO shell layer over the CdO crystallites was inhomogeneous and ranged from 15 to 30 nm. Notably, the thickness of the CdO shell layer on the TiO 2 rod surface was less than the original two-dimensional 100-nm-thick CdO film; this was because the deposition of the two-dimensional continuous CdO film onto the three-dimensional free-standing TiO 2 rods engendered a large surface dispersion effect and decreased the coverage thickness of the CdO on the TiO 2 surface [21]. Figure 3a is a low-magnification TEM image of a single TiO2-CdO composite rod. As shown in the figure, the surface morphology of the shell layer indicated that this layer was comprised of many discontinuous island-like crystallites, and TiO2-CdO exhibited a rugged surface feature. The thickness distribution of the island-like CdO shell layer over the CdO crystallites was inhomogeneous and ranged from 15 to 30 nm. Notably, the thickness of the CdO shell layer on the TiO2 rod surface was less than the original two-dimensional 100-nm-thick CdO film; this was because the deposition of the two-dimensional continuous CdO film onto the three-dimensional free-standing TiO2 rods engendered a large surface dispersion effect and decreased the coverage thickness of the CdO on the TiO2 surface [21].   Figure 4a shows the optical transmittance spectrum of the CdO film. The transmittance of the film substantially decreased in the wavelength range of 500-600 nm, indicating that the CdO film has a bandgap in the visible-light region. The optical bandgap of the CdO film was evaluated using the Tauc model [22]. The inset in Figure 4a illustrates the plot of (ahv) 2 vs. hv, where a is the absorption coefficient and hv is the photon energy. The bandgap of the CdO film, calculated by extrapolating the linear portion of the curve until it intercepted the horizontal axis (photon energy), was approximately  Figure 4a shows the optical transmittance spectrum of the CdO film. The transmittance of the film substantially decreased in the wavelength range of 500-600 nm, indicating that the CdO film has a bandgap in the visible-light region. The optical bandgap of the CdO film was evaluated using the Tauc model [22]. The inset in Figure 4a illustrates the plot of (ahv) 2 vs. hv, where a is the absorption coefficient and hv is the photon energy. The bandgap of the CdO film, calculated by extrapolating the linear portion of the curve until it intercepted the horizontal axis (photon energy), was approximately 2.36 eV. Furthermore, the diffuse reflectance spectra of the TiO 2 and TiO 2 -CdO rods were recorded and converted into absorption coefficient spectra by applying the Kubelka-Munk function [23] (Figure 4b). The optical absorbance edge of the TiO 2 rods was located in the UV region at a wavelength of approximately 400 nm. The optical bandgap of the TiO 2 rods was 3.02 eV (the inset in Figure 4b), which is similar to that reported for rutile TiO 2 [24]. Notably, the optical absorbance edge of the TiO 2 -CdO rods broadened and extended to the visible-light region. This clear red-shift of the optical absorbance edge of the TiO 2 -CdO rods demonstrated the successful decoration of the CdO crystallites, which has a narrower bandgap, onto the surfaces of the TiO 2 rods through sputtering. 2.36 eV. Furthermore, the diffuse reflectance spectra of the TiO2 and TiO2-CdO rods were recorded and converted into absorption coefficient spectra by applying the Kubelka-Munk function [23] ( Figure 4b). The optical absorbance edge of the TiO2 rods was located in the UV region at a wavelength of approximately 400 nm. The optical bandgap of the TiO2 rods was 3.02 eV (the inset in Figure 4b), which is similar to that reported for rutile TiO2 [24]. Notably, the optical absorbance edge of the TiO2-CdO rods broadened and extended to the visible-light region. This clear red-shift of the optical absorbance edge of the TiO2-CdO rods demonstrated the successful decoration of the CdO crystallites, which has a narrower bandgap, onto the surfaces of the TiO2 rods through sputtering.  Figure 5a presents the narrow-scan Cd 3d XPS spectrum for the CdO film. The Cd 3d XPS spectrum contained the main 3d5/2 and 3d3/2 spin-orbit components with the binding energies of approximately 404.56 and 411.36 eV, respectively. These binding energies are ascribed to the Cd 2+ bonding state in the CdO phase [25]. Figure 5b shows the O1s spectrum of the CdO film. The asymmetricity of the spectrum revealed that the as-synthesized CdO film contained oxygen in multiple valence states. The low-binding-energy peak can be attributed to the Cd-O bonds, the medium-binding-energy peak results from the oxygen vacancy in the oxide lattice, and the relatively high-binding-energy peak is associated with the chemisorbed oxygen species on the film surface [26]. Figure 5c depicts the Ti 2p core spectra of the TiO2 rods. The 2p3/2 and 2p1/2 peaks were deconvoluted into four subpeaks. The subpeaks located in the higher-binding-energy regions of the 2p3/2 and 2p1/2 peaks can be assigned to the Ti 4+ valence state, whereas those located in the lower-binding-energy regions can be assigned to the Ti 3+ valance state. The presence of the mixed Ti 4+ /Ti 3+ valance state indicated the presence of oxygen vacancies in the surfaces of the as-synthesized TiO2 rods [27]. The O1s spectrum of the TiO2 rods exhibited a marked asymmetric curve feature (Figure 5d). The O1s peak was deconvoluted into three subpeaks. The low-binding-energy component can be attributed to the Ti-O bonds, the medium-binding-energy component results from the oxygen vacancy in the oxide lattice, and the high-binding-energy component can be ascribed to the oxygen species chemisorbed from the ambient air [28]. The integrated area ratio of chemisorbed oxygen and the oxygen vacancy was 4.3% and approximately 43.8%, respectively. Figure 5e shows the XPS spectral signal of Cd 3d associated with the island-like CdO crystallites decorated onto the surfaces of the TiO2 rods. Figure 5f shows the Ti 2p core spectra of the surfaces of the TiO2 rods in the TiO2-CdO core-shell composites, which clearly demonstrates the mixed Ti 4+ /Ti 3+ valance state in the TiO2 rods. No substantial differences were observed in the Ti 4+ /Ti 3+ valance state of TiO2 rods with and without decoration with the island-like CdO crystallites. By contrast, the O1s spectral intensity substantially increased in the relatively high-binding-energy region of the TiO2 rods after decoration, revealing an increase in the number of crystal defects and chemisorbed oxygen species on the surfaces of the composite rods [2]. In addition, when the TiO2 rods were decorated with the island-like CdO crystallites, the integrated area ratio of chemisorbed oxygen and oxygen vacancy increased to 8.4% and 45.5%, respectively (Figure 5g). The increased oxygen vacancy in the composite rods can be  Figure 5a presents the narrow-scan Cd 3d XPS spectrum for the CdO film. The Cd 3d XPS spectrum contained the main 3d 5/2 and 3d 3/2 spin-orbit components with the binding energies of approximately 404.56 and 411.36 eV, respectively. These binding energies are ascribed to the Cd 2+ bonding state in the CdO phase [25]. Figure 5b shows the O1s spectrum of the CdO film. The asymmetricity of the spectrum revealed that the as-synthesized CdO film contained oxygen in multiple valence states. The low-binding-energy peak can be attributed to the Cd-O bonds, the medium-binding-energy peak results from the oxygen vacancy in the oxide lattice, and the relatively high-binding-energy peak is associated with the chemisorbed oxygen species on the film surface [26]. Figure 5c depicts the Ti 2p core spectra of the TiO 2 rods. The 2p 3/2 and 2p 1/2 peaks were deconvoluted into four subpeaks. The subpeaks located in the higher-binding-energy regions of the 2p 3/2 and 2p 1/2 peaks can be assigned to the Ti 4+ valence state, whereas those located in the lower-binding-energy regions can be assigned to the Ti 3+ valance state. The presence of the mixed Ti 4+ /Ti 3+ valance state indicated the presence of oxygen vacancies in the surfaces of the as-synthesized TiO 2 rods [27]. The O1s spectrum of the TiO 2 rods exhibited a marked asymmetric curve feature (Figure 5d). The O1s peak was deconvoluted into three subpeaks. The low-binding-energy component can be attributed to the Ti-O bonds, the medium-binding-energy component results from the oxygen vacancy in the oxide lattice, and the high-binding-energy component can be ascribed to the oxygen species chemisorbed from the ambient air [28]. The integrated area ratio of chemisorbed oxygen and the oxygen vacancy was 4.3% and approximately 43.8%, respectively. Figure 5e shows the XPS spectral signal of Cd 3d associated with the island-like CdO crystallites decorated onto the surfaces of the TiO 2 rods. Figure 5f shows the Ti 2p core spectra of the surfaces of the TiO 2 rods in the TiO 2 -CdO core-shell composites, which clearly demonstrates the mixed Ti 4+ /Ti 3+ valance state in the TiO 2 rods. No substantial differences were observed in the Ti 4+ /Ti 3+ valance state of TiO 2 rods with and without decoration with the island-like CdO crystallites. By contrast, the O1s spectral intensity substantially increased in the relatively high-binding-energy region of the TiO 2 rods after decoration, revealing an increase in the number of crystal defects and chemisorbed oxygen species on the surfaces of the composite rods [2]. In addition, when the TiO 2 rods were decorated with the island-like CdO crystallites, the integrated area ratio of chemisorbed oxygen and oxygen vacancy increased to 8.4% and 45.5%, respectively (Figure 5g). The increased oxygen vacancy in the composite rods can be attributed to the intrinsic crystal defects in the CdO crystallites, and the increase in the number of chemisorbed oxygen species can be attributed to the increase in the surface area of the composite rods due to its rugged surface morphology. attributed to the intrinsic crystal defects in the CdO crystallites, and the increase in the number of chemisorbed oxygen species can be attributed to the increase in the surface area of the composite rods due to its rugged surface morphology.  Figure 6 presents the gas-sensing responses of the CdO thin film, TiO2 rods, and TiO2-CdO composite rods as a function of the operating temperature on exposure to 1.0 ppm NO2 gas. To prevent microstructural changes in the deposited CdO thin film, the operating temperature was not increased beyond the CdO thin-film growth temperature of 350 °C. The CdO thin film exhibited its  Figure 6 presents the gas-sensing responses of the CdO thin film, TiO 2 rods, and TiO 2 -CdO composite rods as a function of the operating temperature on exposure to 1.0 ppm NO 2 gas. To prevent microstructural changes in the deposited CdO thin film, the operating temperature was not increased beyond the CdO thin-film growth temperature of 350 • C. The CdO thin film exhibited its maximum gas-sensing response at a high operating temperature of 325 • C. By contrast, the gas-sensing response of the TiO 2 rods and the TiO 2 -CdO rods peaked at 275 • C. The optimal operating temperature of the oxide sensors is highly associated with the balance between the chemical reactions and the gas diffusion rate [1,29]. Therefore, the differences in the specific surface area of the samples might crucially affect the optimal operating temperatures of the investigated samples. For example, because of its relatively low specific surface area, the CdO thin film required a relatively higher operating temperature to achieve effective chemisorption between the film surface and the reaction gas. By contrast, in one-dimensional TiO 2 rods-which have a higher specific surface area than do the two-dimensional CdO thin films-equilibrium between the surface reaction with NO 2 gas and the diffusion of NO 2 gas to the surface of the TiO 2 rods was achieved at a relatively low sensor operating temperature of 275 • C. The optimal operating temperature of the TiO 2 -CdO rods was similar to that of the TiO 2 rods, revealing that the decoration of the TiO 2 rods with island-like CdO crystallites did not markedly alter the temperature at which the aforementioned equilibrium was achieved.

Result and Discussion
Materials 2017, 10, 778 7 of 13 maximum gas-sensing response at a high operating temperature of 325 °C. By contrast, the gassensing response of the TiO2 rods and the TiO2-CdO rods peaked at 275 °C. The optimal operating temperature of the oxide sensors is highly associated with the balance between the chemical reactions and the gas diffusion rate [1,29]. Therefore, the differences in the specific surface area of the samples might crucially affect the optimal operating temperatures of the investigated samples. For example, because of its relatively low specific surface area, the CdO thin film required a relatively higher operating temperature to achieve effective chemisorption between the film surface and the reaction gas. By contrast, in one-dimensional TiO2 rods-which have a higher specific surface area than do the two-dimensional CdO thin films-equilibrium between the surface reaction with NO2 gas and the diffusion of NO2 gas to the surface of the TiO2 rods was achieved at a relatively low sensor operating temperature of 275 °C. The optimal operating temperature of the TiO2-CdO rods was similar to that of the TiO2 rods, revealing that the decoration of the TiO2 rods with island-like CdO crystallites did not markedly alter the temperature at which the aforementioned equilibrium was achieved. Figures 7a-c show the gas-sensing response curves of the CdO film, TiO2 rods, and TiO2-CdO rods on exposure to various NO2 gas concentrations, respectively. The sensor operating temperature for the CdO film is 325 °C and those for the TiO2 and TiO2-CdO rods are 275 °C. TiO2 and CdO are n-type semiconductors; therefore, the resistance of the fabricated sensors increased upon their exposure to NO2 gas. This is because the adsorption of NO2 gas molecules onto the surfaces of the CdO film, TiO2 rods, and TiO2-CdO rods engendered the extraction of surface electrons, resulting in an increase in sensor resistance [30]. The reactions between the NO2 gas molecules and the surface electrons of the CdO film, TiO2 rods, and TiO2-CdO rods are as follows [30]: The CdO film, TiO2 rods, and TiO2-CdO rods exhibited evident NO2 gas-sensing behavior on exposure to various NO2 gas concentrations (Figure 7a-c); their gas-sensing responses are summarized in Figure 7d. The responses of the CdO film to 1.0, 2.5, and 5.0 ppm NO2 were approximately 1.78, 2.01, and 2.21, respectively. Those of the TiO2 rods were approximately 3.21, 4.26, and 4.73, respectively, and those of the TiO2 rods decorated with CdO crystallites were higher, at 4.44, 6.21, and 7.41, respectively. The gas-sensing responses of the TiO2−CdO rods were higher than those of the TiO2 rods and CdO film at all investigated NO2 gas concentrations, whereas the CdO film Figure 7a-c show the gas-sensing response curves of the CdO film, TiO 2 rods, and TiO 2 -CdO rods on exposure to various NO 2 gas concentrations, respectively. The sensor operating temperature for the CdO film is 325 • C and those for the TiO 2 and TiO 2 -CdO rods are 275 • C. TiO 2 and CdO are n-type semiconductors; therefore, the resistance of the fabricated sensors increased upon their exposure to NO 2 gas. This is because the adsorption of NO 2 gas molecules onto the surfaces of the CdO film, TiO 2 rods, and TiO 2 -CdO rods engendered the extraction of surface electrons, resulting in an increase in sensor resistance [30]. The reactions between the NO 2 gas molecules and the surface electrons of the CdO film, TiO 2 rods, and TiO 2 -CdO rods are as follows [30]: The CdO film, TiO 2 rods, and TiO 2 -CdO rods exhibited evident NO 2 gas-sensing behavior on exposure to various NO 2 gas concentrations (Figure 7a-c); their gas-sensing responses are summarized in Figure 7d. The responses of the CdO film to 1.0, 2.5, and 5.0 ppm NO 2 were approximately 1.78, 2.01, and 2.21, respectively. Those of the TiO 2 rods were approximately 3.21, 4.26, and 4.73, respectively, and those of the TiO 2 rods decorated with CdO crystallites were higher, at 4.44, 6.21, and 7.41, respectively. The gas-sensing responses of the TiO 2 −CdO rods were higher than those of the TiO 2 rods and CdO film at all investigated NO 2 gas concentrations, whereas the CdO film exhibited the lowest responses. In general, the gas-sensing responses of the CdO film, TiO 2 rods, and TiO 2 -CdO rods increased with the NO 2 concentration because of the increase in the number of NO 2 gas molecules interacting with the surfaces of the samples in the NO 2 -rich test environment. However, the gas-sensing responses of the CdO film increased by only approximately 24% as the NO 2 gas concentration increased from 1.0 to 5.0 ppm, whereas the corresponding change for the TiO 2 rods and TiO 2 -CdO rods was 48% and 67%, respectively. Clearly, the one-dimensional TiO 2 rods with a relative large specific surface area are more sensitive to NO 2 gas at various gas concentrations than is the CdO film, whereas the sensitivity of TiO 2 -CdO rods to NO 2 gas is even higher than that of the TiO 2 rods. This result can be attributed to the differences in the rod surface morphology and the formation of heterojunctions between the TiO 2 core and the CdO shell. The rugged and discontinuous island-like CdO crystallites on the surfaces of the TiO 2 rods provided more sites for NO 2 gas adsorption, as has been reported for one-dimensional semiconductors [31,32]. In addition, crystalline defects, such as surface oxygen vacancies and numerous boundaries in the island-like CdO crystallites, are favorable sites for NO 2 absorption, and the presence of more such defects on the surfaces of the TiO 2 -CdO rods than on the surfaces of the TiO 2 rods improved the adsorption efficiency of the TiO 2 rods decorated with island-like CdO crystallites. In particular, an increase in oxygen vacancy defects in oxides is highly correlated with an increase in their gas-sensing performance [33]. A similar enhancement in gas-adsorption efficiency has been achieved for TiO 2 nanobelts and ZnO nanorods through surface decoration of SnO 2 and TiO 2 nanoparticles, respectively [11,34]. The formation of TiO 2 -CdO core-shell heterostructures is expected to bend the energy bands at the TiO 2 /CdO interface. The interface potential barrier might be another factor to dominate the superior NO 2 gas-sensing sensitivity of the TiO 2 -CdO rods than that of the pure TiO 2 rods. Notably, the gas-sensing sensitivity of the TiO 2 -CdO rods on exposure to 0.25 ppm NO 2 gas was still visible and reached to 3.91 in Figure 7e; however, no reliable sensing response curve was obtained for the pure TiO 2 rods in this work. Furthermore, the gas-sensing selectivity of the TiO 2 -CdO rods was investigated on exposure to various target gases in Figure 7f. Comparatively, the TiO 2 -CdO rods are more sensitive to detect low-concentration NO 2 gas in this work. The response times of the TiO 2 -CdO rods are approximately 41-53 s on exposure to 1.0-5.0 ppm NO 2 gas, respectively. The recovery times are in the range of 305-341 s when the TiO 2 -CdO rods were exposed to 1.0-5.0 ppm NO 2 gas, respectively. Table 1 summarized NO 2 gas-sensing performance of other one-dimensional TiO 2 -based heterostructures prepared by various methods [35][36][37][38]. Comparatively, the TiO 2 -CdO rods have potential in gas sensor applications in detecting low-concentration NO 2 gas. exhibited the lowest responses. In general, the gas-sensing responses of the CdO film, TiO2 rods, and TiO2-CdO rods increased with the NO2 concentration because of the increase in the number of NO2 gas molecules interacting with the surfaces of the samples in the NO2-rich test environment. However, the gas-sensing responses of the CdO film increased by only approximately 24% as the NO2 gas concentration increased from 1.0 to 5.0 ppm, whereas the corresponding change for the TiO2 rods and TiO2-CdO rods was 48% and 67%, respectively. Clearly, the one-dimensional TiO2 rods with a relative large specific surface area are more sensitive to NO2 gas at various gas concentrations than is the CdO film, whereas the sensitivity of TiO2-CdO rods to NO2 gas is even higher than that of the TiO2 rods. This result can be attributed to the differences in the rod surface morphology and the formation of heterojunctions between the TiO2 core and the CdO shell. The rugged and discontinuous island-like CdO crystallites on the surfaces of the TiO2 rods provided more sites for NO2 gas adsorption, as has been reported for one-dimensional semiconductors [31,32]. In addition, crystalline defects, such as surface oxygen vacancies and numerous boundaries in the island-like CdO crystallites, are favorable sites for NO2 absorption, and the presence of more such defects on the surfaces of the TiO2-CdO rods than on the surfaces of the TiO2 rods improved the adsorption efficiency of the TiO2 rods decorated with island-like CdO crystallites. In particular, an increase in oxygen vacancy defects in oxides is highly correlated with an increase in their gas-sensing performance [33]. A similar enhancement in gas-adsorption efficiency has been achieved for TiO2 nanobelts and ZnO nanorods through surface decoration of SnO2 and TiO2 nanoparticles, respectively [11,34]. The formation of TiO2-CdO core-shell heterostructures is expected to bend the energy bands at the TiO2/CdO interface. The interface potential barrier might be another factor to dominate the superior NO2 gas-sensing sensitivity of the TiO2-CdO rods than that of the pure TiO2 rods. Notably, the gas-sensing sensitivity of the TiO2-CdO rods on exposure to 0.25 ppm NO2 gas was still visible and reached to 3.91 in Figure 7e; however, no reliable sensing response curve was obtained for the pure TiO2 rods in this work. Furthermore, the gas-sensing selectivity of the TiO2-CdO rods was investigated on exposure to various target gases in Figure 7f. Comparatively, the TiO2-CdO rods are more sensitive to detect low-concentration NO2 gas in this work. The response times of the TiO2-CdO rods are approximately 41-53 s on exposure to 1.0-5.0 ppm NO2 gas, respectively. The recovery times are in the range of 305-341 s when the TiO2-CdO rods were exposed to 1.0-5.0 ppm NO2 gas, respectively. Table 1 summarized NO2 gas-sensing performance of other one-dimensional TiO2-based heterostructures prepared by various methods [35][36][37][38]. Comparatively, the TiO2-CdO rods have potential in gas sensor applications in detecting low-concentration NO2 gas.   By contrast, a formation of the heterojunction between the TiO2 and CdO is also an important factor that affects the NO2 gas-sensing response of the composite rods. Figure 8 depicts the energy diagram for the TiO2 and CdO phases [39,40]. As evident from the figure, many discontinuous electron-depleted regions were formed at the TiO2 core side when the island-like CdO crystallites discontinuously covered the surfaces of the TiO2 rods. Figure 9 illustrates the effect of interfacial potential barriers on the NO2 gas-sensing response of the TiO2 rods decorated with CdO crystallites. Many discontinuous depletion regions were initially formed over the surface of the TiO2 rods after they were coated with the CdO crystallites. Upon exposure to NO2 gas, the adsorption of the NO2 gas molecules onto the surface of the TiO2 rods resulted in the extraction of the surface electrons. This further induced the formation of a layer-like surface-depletion region on the TiO2 rod surface. The presence of this surface-depletion region markedly decreased the channel size for carrier  By contrast, a formation of the heterojunction between the TiO 2 and CdO is also an important factor that affects the NO 2 gas-sensing response of the composite rods. Figure 8 depicts the energy diagram for the TiO 2 and CdO phases [39,40]. As evident from the figure, many discontinuous electron-depleted regions were formed at the TiO 2 core side when the island-like CdO crystallites discontinuously covered the surfaces of the TiO 2 rods. Figure 9 illustrates the effect of interfacial potential barriers on the NO 2 gas-sensing response of the TiO 2 rods decorated with CdO crystallites. Many discontinuous depletion regions were initially formed over the surface of the TiO 2 rods after they were coated with the CdO crystallites. Upon exposure to NO 2 gas, the adsorption of the NO 2 gas molecules onto the surface of the TiO 2 rods resulted in the extraction of the surface electrons. This further induced the formation of a layer-like surface-depletion region on the TiO 2 rod surface. The presence of this surface-depletion region markedly decreased the channel size for carrier transportation, increasing the resistance of the sensor made from the TiO 2 rods. By contrast, when the TiO 2 -CdO rod was exposed to the NO 2 gas, the inhomogeneous coverage of the island-like CdO crystallites absorbed more NO 2 gas molecules onto the surface of the TiO 2 -CdO rod than did the pure TiO 2 rod; this is because of the defective crystalline feature of the CdO and the increased surface area of the composite rod. The relatively large fluctuation in the charge density, caused by the adsorption of NO 2 gas molecules onto the surfaces of the TiO 2 -CdO rods, further affected the size variation of the depletion region [31]. Consequently, the thickness distribution of the depletion regions over the TiO 2 rod surface became more inhomogeneous, substantially decreasing the carrier transportation efficiency in the TiO 2 rods. Therefore, the variation in sensor resistance on exposure to NO 2 gas would be substantially larger for the TiO 2 -CdO rods than for the pure TiO 2 rods. On the basis of the microstructural differences between the TiO 2 and TiO 2 -CdO rods and the formation of numerous interfacial depletion regions in the TiO 2 -CdO rod, it can be concluded that the TiO 2 -CdO rods exhibit the highest NO 2 gas detection ability among the samples investigated in this study. transportation, increasing the resistance of the sensor made from the TiO2 rods. By contrast, when the TiO2-CdO rod was exposed to the NO2 gas, the inhomogeneous coverage of the island-like CdO crystallites absorbed more NO2 gas molecules onto the surface of the TiO2-CdO rod than did the pure TiO2 rod; this is because of the defective crystalline feature of the CdO and the increased surface area of the composite rod. The relatively large fluctuation in the charge density, caused by the adsorption of NO2 gas molecules onto the surfaces of the TiO2-CdO rods, further affected the size variation of the depletion region [31]. Consequently, the thickness distribution of the depletion regions over the TiO2 rod surface became more inhomogeneous, substantially decreasing the carrier transportation efficiency in the TiO2 rods. Therefore, the variation in sensor resistance on exposure to NO2 gas would be substantially larger for the TiO2-CdO rods than for the pure TiO2 rods. On the basis of the microstructural differences between the TiO2 and TiO2-CdO rods and the formation of numerous interfacial depletion regions in the TiO2-CdO rod, it can be concluded that the TiO2-CdO rods exhibit the highest NO2 gas detection ability among the samples investigated in this study.   transportation, increasing the resistance of the sensor made from the TiO2 rods. By contrast, when the TiO2-CdO rod was exposed to the NO2 gas, the inhomogeneous coverage of the island-like CdO crystallites absorbed more NO2 gas molecules onto the surface of the TiO2-CdO rod than did the pure TiO2 rod; this is because of the defective crystalline feature of the CdO and the increased surface area of the composite rod. The relatively large fluctuation in the charge density, caused by the adsorption of NO2 gas molecules onto the surfaces of the TiO2-CdO rods, further affected the size variation of the depletion region [31]. Consequently, the thickness distribution of the depletion regions over the TiO2 rod surface became more inhomogeneous, substantially decreasing the carrier transportation efficiency in the TiO2 rods. Therefore, the variation in sensor resistance on exposure to NO2 gas would be substantially larger for the TiO2-CdO rods than for the pure TiO2 rods. On the basis of the microstructural differences between the TiO2 and TiO2-CdO rods and the formation of numerous interfacial depletion regions in the TiO2-CdO rod, it can be concluded that the TiO2-CdO rods exhibit the highest NO2 gas detection ability among the samples investigated in this study.

Conclusions
In summary, TiO 2 -CdO core-shell composite rods were initially synthesized through a combination of hydrothermal and sputtering methods. Electron microscopy analyses revealed that the surfaces of the as-synthesized TiO 2 rods were smooth. Moreover, the decoration of CdO thin film onto the surfaces of the TiO 2 rods engendered the formation of a discontinuous shell layer, which comprised many island-like CdO crystallites of varying size. Structural analyses revealed that the as-synthesized TiO 2 -CdO composite rods were highly crystalline, and the XPS spectra provided evidence that oxygen vacancies exist in the surfaces of the TiO 2 rods and CdO crystallites. The superior NO 2 gas-sensing response of the TiO 2 -CdO composite rods relative to that of pure TiO 2 rods and the CdO thin film in this study can be attributed to the increased surface area and surface defect density and the formation of the TiO 2 /CdO heterojunctions.