Novel Christmas Branched Like NiO/NiWO4/WO3 (p–p–n) Nanowire Heterostructures for Chemical Sensing

Establishing a platform comprising different nanostructured oxides is an emerging idea to develop highly sensitive and selective sensing devices. Herein, novel 3D‐heterostructures (p–p–n) consisting of 1D nanowires of NiO and WO3 along with their intermediate reactive product, i.e., NiWO4 seed, are produced by a two‐steps vapor phase growth method. In‐depth morphological and structural investigations describing the growth mechanism of these heterostructures are presented. Finally, the p–p–n heterostructures are integrated into conductometric sensing devices and their performances are investigated toward different gases. It is observed that by modulating the charge‐carrier transport with temperature, the heterostructure sensors exhibit selective behavior toward different gas analytes. Indeed, at 300 °C, the heterostructure sensors show relatively selective behavior toward NO2, while at 400 °C, high selectivity toward VOCs is observed. The improvement in sensing performances is mainly based on charge carrier transport through the two interfaces (one at WO3/NiWO4 (n–p) and the other at NiWO4/NiO (p–p)) and the modulation of charge carriers in the electron depletion layer of WO3 and hole accumulation layer of NiO and NiWO4. The remarkable performance of these complex heterostructures with low ppb‐level detection limits makes them excellent candidates for chemical/ gas sensing applications in e‐noses.

dangerous limits, it can easily volatilize and become extremely harmful for human health (Threshold Limit Value = 250 ppm). [3][4][5][6] Besides its harmful nature, the amount of acetone exhaled in human breath is highly important as it is used for the detection of diseases such as diabetes. [7] Indeed, the presence of 1.8 ppm of acetone in exhaled breath indicates the Diabetic Ketoacidosis (DKA). [8] Another highly toxic and harmful volatile inorganic compound, which is also extremely dangerous for plants, aquatic and land animals and for human health, is nitrogen dioxide (NO 2 ). The main source of NO 2 is the combustion of fossil fuels and the exhaust of motor engines. The Occupational Safety and Health Administration (US) [4] set that the maximum permissible exposure for general industries is 5 ppm, reduced to 1 ppm for short-term exposure. However, the conventional threshold limit value (TLV 2018) is much lower, i.e., around 0.2 ppm. Thus, it is necessary to develop highly sensitive sensors for the detection of NO 2 and acetone.
Nowadays, the research in the chemical/ gas sensor technology is majorly focused on the improvement of selectivity, sensitivity and fast reactivity. [9][10][11] Among all, the selectivity of metal oxide-based sensors toward specific compounds is one of the dominating concerns, as it is the most difficult to achieve. The selectivity of metal oxide-based gas sensors can be improved by functionalization, doping/addition or building a junction between two similar or dissimilar materials. [12][13][14][15][16][17] On the other hand, building a sensor array consisting of various metal-oxide sensing elements combined with pattern recognition algorithms is another successful method to overcome this limitation. [18][19][20] In particular, nano-heterostructure devices based on metaloxides received significant attention due to their superior sensing performances as compared to individual nano-material. [21] The interface between two nanostructured materials modifies the charge carrier transport and increases the surface area for reaction with gas analyte by bringing both nanostructured materials on single sensing platform.
In this work, we are presenting the design of metal oxide based 3D nanostructure system composed of three different materials and their integration into a single sensing platform. From the class of different metal oxides, nickel oxide (NiO) is a rarely investigated p-type semiconducting material that has been gathering great attention due to its unique

Introduction
It is unquestionable that volatile organic compounds (VOCs) are among the most common air pollutants and are highly harmful for human life beyond their exposure limits. [1][2][3] Indoor environment pollution caused by VOCs is a major issue. For example, acetone is one of the highly used solvents in industries and also laboratories. [4] If its concentration exceeds the physical/chemical properties. [22][23][24] NiO is a wide band gap (3.6-4.0 eV) semiconducting material with high thermal stability and excellent optical and electrical properties, which lead this material to be used in various applications. [25][26][27] Even though this material is rarely investigated, our previous reports on NiO nanowires prove its tremendous potential in the gas sensing field. [28,29] On the other hand, WO 3 is an n-type semiconducting material that has been extensively studied in the gas sensing field thanks to its extremely good sensing capabilities. [30][31][32][33] Specifically, we report for the first time the synthesis of NiO and WO 3 nano-heterostructures using vapor-phase growth method for the selective sensing of acetone and NO 2 . During the synthesis process of NiO/WO 3 heterostructures, we have observed the intermediate growth of a ternary material, i.e., nickel tungstate (NiWO 4 ) constructing a branch-like structure. This branch-like heterostructure consists of a big branch of NiO followed by NiWO 4 seeds on top, which lead to small branches of WO 3 . The growth optimization with morphological and structural characterization has been discussed in detail. This work is mainly focused on the development of high-performance selective gas sensing device based on NiO/NiWO 4 /WO 3 heterostructure directly synthetized on the active transducers and to understand their gas sensing mechanism.

Results and Discussion
The morphology of the nanostructures was investigated by FE-SEM. In line with our previous report, [28] bare NiO (Figure 1a) samples on alumina substrates exhibit a scattered and homogenous nanowire morphology with a diameter in range of 20 to 60 nm and a length at micrometer scale. In Figure 1b-f the heterostructure morphology shows the growth of small WO 3 nanowires on top of each NiO nanowire, effectively designing branch-like nanostructures.
For structural characterization, GI-XRD measurements have been performed on both pristine NiO nanowires and NiO/ WO 3 heterostructures. The detailed discussion of the crystalline structure is included in the Supporting Information. Figure S1 in the Supporting Information reports the diffraction pattern of NiO nanowire and NiO/WO 3 heterostructures fabricated on alumina substrate. The NiO cubic phase has been clearly www.afm-journal.de www.advancedsciencenews.com observed in the GI XRD spectrum of heterostructure along with the major contribution from alumina substrate. Spectra comparison of NiO nanowire with NiO/WO 3 heterostructures reveals the presence of a ternary phase of nickel and tungsten, i.e., NiWO 4 . The peak observed at 2θ equal to 31° is attributed to the nickel tungstate (NiWO 4 ) crystallographic phase according to the (JCPDS #15−0755), indicating the monoclinic phase. [34] Therefore, we can conclude that GI-XRD spectrum of the heterostructure system consists of three different phases: NiO, WO 3 , and NiWO 4 . However, due the low intensity of the peaks related to this phase, which can be attributed to the low amount of material, it is difficult to confirm its presence only by GI-XRD.
For structural analysis of intermediate NiWO 4 , the heterostructure samples were investigated by Transmission Electron Microscopy (TEM). Firstly, the diameter of the branches of WO 3 formed on the NiO core ranges between 6 and 11 nm, as presented in the Conventional TEM (CTEM) image (Figure 2a; Figure S2, Supporting Information ). The average diameter of the small nanowires was calculated around 8.6 nm. Figure 2b,c demonstrates the formation of a nucleation site on the NiO nanowire, from which a small branch initiates. This kind of nucleation can be explained using vapor-solid (VS) mechanism related to the vapor phase growth method. In VS growth, the source materials (WO 3 ) vaporized at high temperature and directly condensed on the target surface (NiO NWs). Once the condensation process starts, the initially condensed molecules create a seed crystal that serve as nucleation sites for the further growth of nanostructures.
In Figure 2b reports the Selected Area Electron Diffraction (SAED) pattern of the heterostructures taken from the nanostructures presented in Figure 2a. The superposition of different lattices due to the materials composing the heterostructure is clearly evident from the coexistence of multiple diffraction spots.  (Figure 2c) clearly shows the interface between two phases related to WO 3 and NiWO 4 . Indeed, from Figure 2a,c, the formation of a NiWO 4 seed is clearly visible. The NiWO 4 intermediate phase between core of NiO and WO 3 was found highly crystalline with orientation of (110) corresponding to a monoclinic crystalline phase of nickel tungstate. Furthermore, EDX maps of the selected chemical elements (Figure 2e-g) show a well identified thick core of NiO with scattered branches of WO 3 .
To further discriminate the NiWO 4 phase, Raman spectroscopy was performed on the heterostructures, NiO nanowires and alumina substrate. Figure 3 shows the comparison of the Raman spectra obtained for bare NiO nanowires, heterostructure and Al 2 O 3 substrate. Spectrum obtained for the NiO/WO 3 www.afm-journal.de www.advancedsciencenews.com heterostructures clearly exhibits the presence of peaks related to all three materials.
The Raman peaks of NiWO 4 are mainly related to tungsten sub-lattice, as nickel has relatively weak Raman activity. The major peaks related to NiWO 4 monoclinic structure exhibit at 890, 698, 1030, and 552 cm -1 . [35,36] The monoclinic NiWO 4 has wolframite structure and usually consists of six internal stretching modes caused by each six WO bond. Raman spectrum ( Figure 3) shows an intense peak at 890 cm -1 that is associated to A g mode. The higher intensity of this peak is related to the regular octahedron symmetric stretch of shortest bond of WO. Furthermore, other related peaks (at 550 and 698 cm -1 ) are associated to the asymmetric stretching of the WO bonds. [32,37] Indeed the peaks corresponding to the WO 3 monoclinic phase are also clearly observed at 805, 328, and 268 cm -1 . The intense peak at 805 cm -1 corresponds to WOW stretch vibrations of the bridging oxygen. [38] Moreover, the two peaks observed at higher Raman shift (at 1090 and 1490 cm -1 ) are related to the NiO cubic phase. [28] The first peak observed at 1090 cm -1 is associated to 2LO modes while the second one at 1490cm -1 to two magnon (2M) scattering. [39] These peaks are often not visible in the case of pure NiWO 4 . However, in the case of fabricated NWO heterostructure system, the peaks related to all three materials (NiO, WO 3 and intermediate NiWO 4 ) are fully distinguishable.
Furthermore, XPS analysis has been performed on the fabricated heterostructures (see Figure S3 in the Supporting Information ). The obtained survey spectrum show the presence of Ni, W, O and C according to their respected binding energy. The analysis shows that the Ni element is mainly present in +2 oxidation state, while W is in +6 oxidation state. Moreover, the O1 s spectrum indicates O-bond of Ni and W in NiWO4 lattice [40,41]

Discussion of NWO Heterostructures Growth Mechanism
The morphological and structural characterizations performed on the fabricated NWO heterostructure systems indicate the presence of three different materials. Specifically, in-between the junction NiWO 4 was formed due to the reaction between the NiO and WO 3 , resulting into the formation of NiO/NiWO 4 / WO 3 (p/p/n) heterostructure. The possible explanation for this demeanor can be understood by the schematic shown in Figure 4. After the growth of NiO nanowires, WO 3 vapors start to condensate on NiO surface. The reaction occurring between NiO (s) and WO 3(v) creates a new NiWO 4 phase. Indeed, these transition metal tungstate of type M 2+ WO 4 are ternary oxide semiconductors, highly studied in literature for various applications such as supercapacitors, biosensors and gas sensors. [35,[42][43][44] In an interesting report given by Jacob et al. [45] the phase transition diagram of N-W-O system at temperature of 1200 K was studied. The authors described that at a temperature close to 950 K  (pressure = 1 atm.), the tungsten-oxygen system consists of different W-O phases such as WO 2 , WO 2.72 , WO 2.9 and WO 3 . Moreover, in presence of Ni and oxygen the only ternary oxide phase that has a stable existence is nickel tungstate. At the given experimental conditions, the existing different phases of W-O transit to formation of NiWO 4 by combining the standard free energy. By simply taking out the free energies at 1200 K, the combination of the binary components of NiO and WO 3 leads to the formation of NiWO 4 , which is the only stable ternary phase In our work, during the vapor phase growth, WO 3 powder was heated at temperature of 1100 °C to induce its evaporation. The induced vapors consisted of different phases of tungstenoxygen, as it is well known that the materials stoichiometry may change during the vapor phase growth due to some oxygen losses in the system due to high temperature and pressure.
When tungsten-oxygen vapor reaches the NiO NWs surface, it reacts with NiO creating the most stable ternary NiWO 4 phase.
Thus, according to our proposed scheme, during the heterostructure growth tungsten-oxygen vapors condensate and initiate the seed formation on the surface of NiO nanowires. These seeds continue to grow as NiWO 4 as long as the incoming vapors of WO 3 are able to reach and react with the NiO. As the growth process continues, NiWO 4 seeds fully covered the NiO NWs surface. Due to this, the further incoming WO 3 vapors are unable to reach NiO NWs and the formed seeds of NiWO 4 act as nucleation sites for WO 3 nanowires growth.

Electrical and Gas Sensing Characteristics
The electrical properties of a batch of six NWO sensors and two of NiO sensors has been investigated toward different reducing and oxidizing chemical compounds. Figure 5a reports  (50 ppm) and acetone (50 ppm). c) Dynamic response of NiO and NWO sensing device toward reducing gases ethanol (10, 10, 20, 50 ppm) and acetone (30,30,50, 100 ppm). d) Calibration plots for NiO and NWO toward ethanol and acetone at their optimal working temperature of 400 °C. All the measurements were performed at relativity humidity of 50% at 20 °C. The legends for each panel are shown at the bottom. www.advancedsciencenews.com the variation of baseline conductance of the sensors (NiO NWs and NWO heterostructure) under different operating temperatures. The sensors conductance increased with the temperature increase, exhibiting the typical intrinsic semiconducting behavior of metal oxide based sensors. These curves were taken in air without other chemical compounds, to understand the nature of the sensing devices. It is clearly observed that, with the formation of the NWO heterostructures, the baseline resistance of the sensors increased. In particular, the conductance drop in heterostructure devices is almost one order of magnitude at temperatures higher than 200 °C as compared to NiO NWs. However, the trends of conductance changes with temperature for both types of devices are quite similar. This higher electrical resistance of NWO heterostructures is attributed to the formation of two interfaces: one at NiO/NiWO 4 and other at NiWO 4 /WO 3 . In order to reach the equilibrium and Fermi levels alignment, the charge carrier diffusion through these interfaces occurs. This diffusion of charge carriers creates the potential barriers at these interfaces, hence increasing the devices resistance.
All sensing devices were tested toward different reducing and oxidizing gases in a range of temperature from 200 to 500 °C. Firstly, these novel heterostructures were tested toward VOCs such as ethanol and acetone at different concentrations and temperatures. Figure 5b compares the response of NWO heterostructures with NiO NWs sensors at different temperature toward 50 ppm of ethanol and acetone. The heterostructure sensors exhibit highest response at the optimal working temperature of 400 °C for ethanol and acetone. Indeed, the heterostructure sensor response is higher toward acetone compared to ethanol. This demonstrates that with the heterostructures formation the sensors sensitivity increases comparing to bare NiO nanowire sensors. Furthermore, Figure 5c shows isothermal dynamic-transient responses toward different concentrations of ethanol (10, 10, 20, and 50 ppm) and acetone (30,30,50, and 100 ppm) at 400 °C (optimal working temperature) with relative humidity of 50%. Both NiO and NWO sensors show an electrical conductance decrease when exposed to a reducing gas. This is the typical behavior of a p-type metal oxide sensor under reducing compound exposure. This confirms that the holes charge carriers majorly dominate the conductance of the heterostructures system. The dominance of holes in the heterostructure is expected as both NiO [24,46] and NiWO 4 [47,48] are p-type semiconductors. Figure 5d shows the calibration curves in log-log scale for the NWO and NiO sensors for ethanol and acetone at the optimal working temperature of 400 °C. The curves clearly indicate that NiO NWs show more selective response toward ethanol in comparison to acetone. While, for the heterostructure sensors, the higher selective response was observed toward acetone. Furthermore, all the calibration curves follow a typical power law relation (Equation (7)) for semiconductor metal oxide-sensing devices, confirming the absence of any saturation process. All of the curves were fitted with the power trend law and the values of sensing parameters (coefficients A and B; detection limits) obtained after fitting are presented in Table S1 in the Supporting Information. The detection limits were calculated for all of the sensors by considering the minimum response value of 1 in Equation (7). The lowest detection limits calculated for NiO NWs sensors are higher than 1 ppm for both the gases, while the detection limits of the heterostructured sensors is lower to 0.5 and 0.7 ppm for ethanol and acetone, respectively.
In order to have a fair comparison between heterostructure and NiO NWs sensors, we have considered our previous report on VLS grown pristine WO 3 [32] nanowires sensors. Figure S4 in the Supporting Information reports the comparison of sensing responses of all three type of sensors at 50 ppm of ethanol and acetone at their optimal working temperatures. It should be noted that for the WO 3 nanowires sensors the optimal working temperature was 500 °C instead of 400 °C as for NiO and NWO. Clearly, among all different kind of sensors, pristine WO 3 nanowires are the least sensitive toward these VOCs. On the contrary, NWO heterostructures devices show the highest responses as compared to both pristine materials. Indeed, heterostructure sensors exhibit 5-fold higher response as compared to NiO NWs sensors for acetone.
To accomplish a complete sensing profile of the NWO heterostructure sensors, their performances were tested toward various gases such as hydrogen, toluene, carbon monoxide and nitrogen dioxide. Figure 6a reports NWO sensors response with respect to different operating temperatures at a fixed relative humidity of 50%. The humidity value was kept constant for these preliminary investigations to achieve a test environment close to real conditions. Different concentrations values of the gases have been chosen according to their exposure limits. For instance, testing a very high concentration such as 50 ppm of NO 2 has no significance in real testing conditions. The average exposure to NO 2 , in a period of one hour, should be lower than 0.2 ppm according to European Union (EU) Air Quality Standards. [49] We already observed that NWO heterostructure sensors show a high response toward acetone at higher operating temperatures, i.e., at 400 and 500 °C. At 300 °C, instead, sensors show a higher response toward NO 2 . This demonstrates that by modulating the operating temperature it is possible to change sensor's sensitivity toward different analytes. For an exhaustive study of the NWO heterostructures and NiO NWs sensors for NO 2 detection at their optimal working temperature of 300 °C, the sensors were tested toward different concentration of NO 2 at relative humidity of 50% (see Figure S5, Supporting Information). The sensors conductance increases when exposed to NO 2 (oxidizing gas), indicating a typical p-type semiconducting behavior for both the materials.
Furthermore, Figure 6b reports the calibration curve, i.e., response versus concentration, for NO 2 (0.5, 1, 2, 3, 5 10 ppm) at the optimal working temperature. Calculated coefficients with power fitting (eqn. 3) of the curve have been reported in Table S1 in the Supporting Information. Considering the minimum response of 1, the detection limit of NWO sensors was found in the ppb level, showing that the sensors are highly sensitive toward NO 2 at these operating conditions. As discussed earlier, all sensors were tested at a fixed relative humidity of 50%. However, the performances of heterostructure sensors were also investigated in dry air and in much higher humid environment (75%, see Figure 6c) at fixed concentration 5 ppm of NO 2 . The response of the sensors resulted higher in dry atmosphere than in the presence of humidity. More specifically, the response dropped more than half in humidity from the value in dry air. www.advancedsciencenews.com This effect is really common in metal oxide based sensors, ordinarily known as poisoning effect of water on metal oxide surface in the presence of humidity. [50,51] In a wet environment, the moisture adsorbs on the metal oxide surface sites, reducing oxygen adsorption. Due to this, the interaction of the gas molecule with the metal oxide surface decreases and a reduction of sensors response is observed. However, NWO heterostructured sensors showed remarkable sensing performance in presence of humidity, highlighting the potential of heterostructure sensors to work in real environmental conditions. Furthermore, Figure 6d compares the response of NWO heterostructured sensors with the host NiO NWs and previously reported VLS grown WO 3 [32] nanowires. WO 3 sensors exhibit the worst response in comparison to NiO NWs and NWO heterostructures. The NWO heterostructures show superior NO 2 sensing performances comparing to both pristine materials. Finally, these results emphasize that building a single sensing platform, combining the properties of two different materials, may enhance the device performances but also increase the system complexity. However, collective sensing data make these materials potential candidates to use in an e-nose or in sensor arrays. To evaluate the heterostructures dynamic response, the response and recovery times of the NWO heterostructure sensor toward different gas analytes have been reported in Figure S6 in the Supporting Information.

Gas Sensing Mechanism
The gas sensing mechanism of this complex NWO heterostructures and pristine NiO and WO 3 nanowires can be described by oxygen chemisorption model for metal oxides. [52] The basic terminology of this model is the occurrence of chemisorption of oxygen ions on metal oxides surface when placed in Figure 6. a) Temperature dependence response of NWO heterostructure sensors toward different interfering gases, CO (50 ppm), NO 2 (2 ppm), ethanol (50 ppm), acetone (50 ppm), H 2 (50 ppm) and toluene (50 ppm). b) Calibration plot for NWO sensors toward different concentration of NO 2 (0.5, 1, 2, 3, 5, 10 ppm) at optimal working temperature of 300 °C. (50% RH at 20 °C) c) Response of NWO sensors at fixed concentration of NO 2 (5 ppm) and temperature of 300 °C in dry air and 75% of relative humidity. d) Comparison of NO 2 (1 ppm) sensing performance of fabricated NWO and NiO sensors together with already reported VLS grown WO 3 [32] nanowires. (At 300 °C, 50% RH at 20 °C). www.advancedsciencenews.com air. These chemisorbed oxygen ions extract electrons from the metal oxides surface leading to the formation of an electronic core-shell structure. [52] Prior to discuss the sensing mechanism of this complex heterostructures, the understanding of the sensing behavior of single material which is NiO is important. Figure 7 depicts the three stages in which NiO NWs are in vacuum, air and finally exposed to acetone gas. It should be pointed out that NiO is a p-type semiconducting materials and holes are the majority charge carriers. [28] On the contrary, WO 3 is a n-type material in which electrons are the majority charge carriers. [32] In the first stage (Figure 7), NiO nanowires are in vacuum and the electronic bands are flat. When p-type NiO NWs are placed in air, the chemisorption of oxygen occurs on their surface and the types of these chemisorbed ions depends on the operating temperature. Generally, in the temperature range between 150 °C ≤ T ≤400 °C these chemisorbed oxygen ions are in the form of O¯. [53] The overall effect of this chemisorption is the extraction of electrons from the NiO surface. This extraction of electrons forms narrow semiconducting hole accumulation layer (HAL) near the surface with an insulation core. When this oxygen chemisorbed NiO surface is exposed to a reducing gas, such as acetone, the electrons donation from acetone to NiO nanowires surface occurs, resulting into an electron-hole recombination in HAL. Thus, the density of holes reduces and the resistance of p-type NiO nanowires sensors decreases. On the other hand, the chemisorption of oxygen ions on n-type WO 3 surface causes the formation of electron depletion layer (EDL) near the surface along with the semiconducting core. Hence, when exposed to reducing gases, the resistance of the WO 3 nanowires sensor decreases due to the narrowing of EDL width ( Figure S7, Supporting Information). The effect is vice-versa when these semiconductors are exposed to oxidizing gas such as NO 2 .
Furthermore, as described earlier, NWO heterostructure sensor showed superior performance as compared to the pristine NiO and WO 3 nanowires. One of the reasons behind their superior performance is the enhancement in effective surface area of the sensor due to the formation of heterostructure involving three different materials. For further deep analysis, it is important firstly to understand the electrical conduction within the heterostructure itself. This type of heterostructure can be consider as n/p/p junction in which the conduction is controlled by the dominating charge carriers, i.e., hole as heterostructure carries more p-type character. In NWO heterostructures two interfaces formed, i.e., one between WO 3 and NiWO 4 (n-p) while other at NiWO 4 /NiO (p-p) side, resulting in a double heterojunction (DH) device. Figure 8 represents the different stages of NWO heterostructure, i.e., in vacuum, in air and exposed to gas analytes at different operating temperature. As shown in Figure 8a, when the heterostructure is formed (in vacuum), the charge carriers diffusion occurs due to the difference in their densities on the two interfaces, causing the formation of depletion region (interface potential barriers). Furthermore, when this heterostructure is exposed to air, the chemisorption of O¯ ions results into the formation of electron core-shell structure within the semiconductors. In particular, EDL is formed on WO 3 with semiconducting core, while in NiWO 4 and NiO, HAL is formed along with resistive core. It should be noted that due to the existence of resistive cores in NiWO 4 and NiO, these two p-type semiconductors will be highly resistive as compared to WO 3 in which semiconducting core exists (low resistance) and their response will be determined by narrow HAL. Moreover, under the influence of temperature, the charge carriers gain enough energy to overcome both depletion regions and the conductance of the heterostructure increases (see Figure 5a). Specifically, with the increase in temperature more and more holes are injected from both p-type semiconductors and start dominating the overall heterostructure conductance. The temperature will also cause the electron-hole recombination largely at WO 3 /NiWO 4 interface in which both electrons of WO 3 and holes of NiWO 4 will consumed. However, in whole NWO heterostructure, the electron-hole recombination will not be significantly affecting the holes density due to dominating p-type character. While the density of electrons in NWO heterostructure will be greatly reduced. Indeed, this electron-hole recombination will get more and more effective as the temperature increases. Thus, the density of free electrons in the conduction band of WO 3 will be more at 300 °C in comparison to higher temperatures. However, NWO sensors exhibit p-type behavior for both oxidizing and reducing gases between temperature 200-500 °C, which implies that the dominating charge carriers in this heterostructure will always be holes. We believe that the crux, that plays an important role in this distinctive performance of NWO heterostructure sensors, is the hole mobility which increase with operating temperature. Indeed, as it seen from Figure 5a, at 400 °C the heterostructure conductance is www.afm-journal.de www.advancedsciencenews.com 3-times higher than at 300 °C. This abrupt increase in conductance suggests that the holes obtain enough energy to overcome the interfaces potential barriers (WO 3 /NiWO 4 and NiWO 4 /NiO) and start dominating throughout the heterostructure.
At 300 °C, when NWO heterostructure sensor are exposed to both oxidizing (NO 2 ) and reducing gas (acetone), the sensors show highest selective response toward NO 2 (see Figure 8b). NO 2 is an electron loving gas and has much higher electron affinity (≈ 2.28 eV) in comparison to ionosorbed oxygen (0.43 eV) at the surface. [54] When the heterostructure system is exposed to NO 2 , physisorbed NO 2 molecules create new surface acceptor levels deeper than surface oxygen ions. Hence, the bound electrons from the adsorbed oxygen are transferred to NO 2 molecule forming NO 2 species. In this heterostructured system, the incoming gas can react with all three different surfaces. However, as previously discussed, at this particular temperature (300 °C) the outer WO 3 has higher electron density. Therefore, NO 2 can oxidize the surface deeper and can capture electrons from the WO 3 surface, which can further enhance the response toward NO 2 gas. Only limited response has been achieved for reducing gases due to the deficiency of mobile holes at this particular temperature. The NO 2 gas will go through the following reactions [55] NO gas NO ads This capture of electrons by the NO 2 gas generates additional holes in the NWO heterostructure (WO 3 surface), which increases the overall conductance of NWO heterostructure sensor system. However, at 400 °C a large number of free mobile holes exists in the NWO heterostructure and the sensors respond more toward reducing gas as compared to oxidizing one (see Figure 8c). In particular, when heterostructure sensors are exposed to acetone at 400 °C, after the reaction with the adsorbed oxygen ions it donates electrons to the heterostructure system. Because of the presence of a higher hole density at this temperature, the injected electrons can recombine reducing the overall conductance of the heterostructure sensor system. The reaction between the acetone molecules can be expressed by following equation [56] C H O 8O 3 CO 3H O 8e 3 6 (ads) 2 2 On the other hand, only limited response has been achieved for oxidizing gases due to the deficiency of electrons at 400 °C. The enhancement in the response of this heterostructured system can be explained due the formation of interfaces between the three different materials (Figure 8c), in comparison to the host NiO NWs.
Furthermore, the increase of surface to volume ratio could be another reason for the further improvement in the sensing performance. However, the selective behavior toward different gases can only be comprehensible due to the presence of these interfaces that bring the distinctive properties of three semiconductors into a single sensing platform. Moreover in literature many reports have been published on the enhancement of the sensing performances in metal oxide based gas sensors due to the presence of p/n (NiO/ZnO), [56] (MoO 3 / ZnO), [57] (SnO 2 /NiO) [50] and n/n (WO 3 /ZnO), [58] (ZnO/SnO 2 ) [59] junctions. Most interestingly, it has been discussed in literature that adding W or WO 3 to NiO or vice versa can improve the sensing properties of the material. [60][61][62][63] On the other hand, Bao et al. [61] report the enhancement of room temperature sensing toward NO 2 sensing by synthesizing the NiO/WO 3 plate like heterogeneous nanocomposites. Authors explain the enhanced sensing behavior due to the formation of a depletion region between the two materials. These NiO/WO 3 nanocomposites showed response values around 5 and 3 toward 30 ppm of NO 2 at room temperature and 250 °C, respectively. Compared to the results presented in this work (R = 27 for 1 ppm NO 2 ), Bao et al. nanocomposites have the advantage of RT sensing capability. However, testing such high concentrations of NO 2 has very low impact for the development of real time NO 2 sensing device, as the dangerous limit for NO 2 is lower than 2 ppm. [29] Furthermore, comparison with the literature (see Table S2, Supporting Information) confirms the superior performance of NWO heterostructures.
Hence, according to all our observations the fundamental reason behind the superior performance of our novel heterostructure is the formation of interface between three semiconductor materials with tunable distinctive electrical properties on a single sensing platform. By controlling the charge carrier's density and their flow within the heterostructure with temperature, the sensors respond selectively toward both oxidizing and reducing gas analytes.

Conclusion
A novel and complex heterostructure of NiO/NiWO 4 /WO 3 have been successfully synthesized by single vapor-phase growth technique on alumina substrates. As synthesized host NiO nanowires show dense and long nanowire morphology with the diameter in the range of 20-60 nm. The SEM and HRTEM studies reveal that small WO 3 nanowires (diameter 6-11 nm) grew effectively on NiO nanowires surface first by creating a seed of a ternary material NiWO 4 , which acts as a nucleation site for the WO 3 nanowires. Raman spectroscopy provides a clear evidence of the presence of three different metal oxides in the NWO heterostructure system. Furthermore, the NWO heterostructured sensor shows the superior sensing performances compared to the ones of NiO and WO 3 nanowires. In particular, the heterostructured sensors show higher response toward VOCs at temperature of 400 °C and toward NO 2 at 300 °C. Based on the observations and in-depth analysis, the fundamental reason behind their superior performances is the formation of an interface between three nanostructured oxides with tunable and distinctive electrical properties. By controlling the charge carrier's density and their flow within the heterostructure with temperature, the sensors respond selectively toward both oxidizing and reducing gas analytes. Specifically, at 300 °C free electrons in the conduction band of WO 3 exist in the heterostructure, which makes them more selectively www.afm-journal.de www.advancedsciencenews.com reactive toward NO 2 (oxidizing gas) with a response of 195 toward 5 ppm of NO 2 . At the same time, at 400 °C, a significant electron-hole recombination consumes the free electrons and large number of holes exists in heterostructures, which makes them more selective toward the electron donating gases. Due to this, at 400 °C sensor showed selectivity toward the VOCs (reducing gases) with a response of 35 and 48 toward 50 ppm of ethanol and acetone respectively. Overall, we are convinced that the proposed novel heterostructure sensors would pave a way to build highly sensitive devices in the application of 1D heterostructure gas sensor and to develop a sensor array for interesting future gas sensing technology.

Experimental Methods
Synthesis: For the fabrication of nanostructures, polycrystalline alumina substrates (Al 2 O 3 , Kyocera, Japan, 99% purity, dimensions = 2 mm × 2 mm) were used. Prior to the deposition, all the substrates were ultrasonically cleaned with acetone and dried with synthetic air. An ultrathin layer of Au catalyst on Al 2 O 3 substrates was deposited using RF magnetron sputtering (Kenotec Sputtering system, Italy). The deposition of Au was performed at the power of 70 W with 7 SCCM of Ar flow at a pressure of 5 × 10 -3 mbar for 5 seconds.
The synthesis of NiO nanowires was carried out in a tubular furnace using vapor-liquid-solid (VLS) growth mechanism. Based on the previously obtained results, [28] the growth process was carried out at evaporation temperature for NiO powder (Sigma-Aldrich 99% purity CAS number: 1313-99-1) at 1400 °C, Ar flow of 100 sccm at pressure of 1 mbar. The Au catalyzed Al 2 O 3 substrates were kept at the temperature of 930 °C and the deposition time was set for 15 min. Furthermore, for the fabrication of heterostructures, WO 3 nanowires were grown directly onto the NiO nanowires via Vapor-Solid (VS) growth mechanism. The source material, WO 3 powder (Sigma-Aldrich 99% purity CAS number: 1314-35-8), was heated up to 1100 °C while the NiO nanowires were placed at lower substrate temperature of 580 °C. During the deposition process, the argon flow was set to 100 SCCM and pressure was kept at 1 mbar with a deposition time of 15 min.
Characterization: A field-emission scanning electron microscope (FE-SEM, model LEO 1525, ZEISS), was used to investigate the morphology of the fabricated nanostructures. Morphological and structural investigations were carried out using a JEOL JEM ARM200F analytical microscope operated at 200 kV and equipped with a JEOL JED-2300T unit for Energy Dispersive X-Ray (EDX) analysis. Samples were observed at magnification from 60 kx to 2 Mx. Samples were obtained by grazing the nanowires from the substrate with a diamond pen. A drop of ethanol was placed at the place of the grazing. A copper grid with lacey carbon film was put on top of the drop and then it was left to dry.
Raman spectra were measured by using a XploRA Nano system (Horiba Jobin Yvon Srl, Italy) formed by a confocal microscope (Olympus BX) and a 1800 gr/mm reticule. A Peltier-cooled Open Electrode CCD was used to record the Raman spectra excited by a 638 solid-state laser. Spectra were recorded in the wavelength range 200-1800 cm -1 . Nearambient pressure XRay photoelectron spectroscopy (SPECS GmbH, Germany, Al Kα monochromatized source) was used to perform the chemical analysis of the surface of the heterostructures in the hundreds pascals range atmosphere at room. All the experiments were performed at room temperature. The BE axis for each spectrum has been aligned by taking as reference the C1s core level of the adventitious carbon at 284.8 eV. Further, KolXPD software has been used for the fitting and analysis of XPS data, Gas Sensing Characteristics: A batch of heterostructured NiO/WO 3 and pristine NiO nanowire sensing devices were fabricated by DC magnetron sputtering. For the complete device fabrication, Pt interdigitated electrodes (IDEs) and a Pt heating element were deposited by a two-step procedure, on top of the NiO/WO 3 and pristine NiO sensing elements and on the backside of the alumina substrates, respectively. The twostep procedure consists in: a) TiW adhesion layer by DC magnetron sputtering (70 W Ar plasma, 100 nm, 300 °C, ≈ 5.3 × 10 −3 mbar); (b) Pt electrodes, using the same parameters used for the adhesion layer (thickness≈1 µm). Devices were finally mounted on transistor outline (TO) packages using electro-soldered gold wires.
Gas sensing tests were performed in a sealed climatic chamber (20 °C; relative humidity level = 50%, in order to simulate real ambient conditions). A constant synthetic airflow [rate = 200 standard cubic centimeters per minute (SCCM)] at atmospheric pressure was used as gas carrier. The atmosphere composition was controlled using mass flow controllers (MKS, Germany), mixing flows coming from certified gas bottles (SOL, Italy) containing a precise concentration of target analytes diluted in synthetic air. The output signal was measured by applying a constant bias of 1 V to the sensing materials, recording the resulting current by means of a picoammeter (Keithley, USA). Prior to gas sensing measurements, all sensors were thermally stabilized at the desired working temperature for 8 hours. The devices reproducibility and repeatability were investigated by performing measurements on up to 6 identical sensors under the same experimental conditions, yielding an estimated maximum uncertainty. The response was calculated by the variation of the conductance using the following formulas for reducing and oxidizing gases, respectively where R gas and G gas are respectively the sensor resistance and conductance in presence of gas, and R air and G air in synthetic air. Sensing performances were evaluated toward various concentrations of nitrogen dioxide (NO 2 ), ethanol, acetone, hydrogen (H 2 ), toluene and carbon monoxide (CO) were supplied by SOL (Italy) in a certified bottle. Firstly, a temperature screening was performed, in the range of 200-500 °C, to identify the optimal working temperature for each chemical compound. Secondly, calibration curves were measured, at the optimal working temperature for acetone, ethanol and NO 2 . Power-law trend line was calculated also, as it is very common for metal oxide conductometric devices: [64] A C where C is the concentration of the target compound, A and B are constants related to the material composition and the involved surface chemical reactions. Moreover, detection limits, response and recovery times, and sensing responses under influence of different interfering gases are presented.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.