Bridging Nanowires for Enhanced Gas Sensing Properties

: It is crucial to develop new bottom-up fabrication methods with control over the physical properties of the active materials to produce high-performance devices. This article reports well-controlled, without seed layer and site-selective hydrothermal method to produce ZnO bridging nanowires sensors. By controlling the growth environment, the performance of the sensor became more efﬁcient. The presented on-chip bridging nanowire sensor enhanced sensitivity toward acetone gas (200 ppm) around 63 and fast response time (420 ms) and recovery time (900 ms). The enhancement in the speed of response and recovery is ascribed to the exceptional NW-NW junction barrier that governs the sensor’s conductivity, and the excellent contact between ZnO nanowires and Au electrodes.


Introduction
The controlled synthesis of metal-oxide nanostructures (MONs)has advanced swiftly in recent years. Many researchers are still trying to understand how morphology, property, and application are related to achieving high functional devices. Among the most important devices are gas sensors, due to their crucial role in a wide range of important fields, including industrial process control, safety systems, disease diagnoses, and environmental monitoring [1][2][3][4][5][6][7][8][9].
As gas sensors, MONs are promising because they show high sensitivity and fast response to low concentrations of gases. As a consequence of their very small size, MONs resistance can easily be influenced by changes at their surfaces because the effect of surface interactions can extend to a distance from the surface that is comparable to the dimensions of the MONs [10][11][12][13][14][15][16][17][18]. The sensing mechanism of MONs depends on activating the surface atmospheric oxygen at relatively high operating temperatures. Accordingly, the reactions on the surface will free more electrons and transfer them from the surface to the bulk of MONs, and this process is responsible for changing their resistance [4].
Gas sensors fabricated using ZnO thin film are usually not sensitive enough and take a long time to respond. Alternatively, ZnO nanostructures are more sensitive and can detect a much lower concentration of gases in a relatively shorter time because of their higher surface-to-volume ratio and smaller size [10][11][12][13][14][15][16][17][18].
In the fabrication of BNWs sensors, substrates with prepatterned Au electrodes were put face-down in a vial containing Zn(NO 3 ) 2 ·6H 2 O (25 mM), HMTA (12.5 mM), and NH 3 ·H 2 O (0.35 M) as an initial growth stage where the vial was heated in an oven at 85 • C for 45 min. The secondary growth stage was started by replacing the growth solution in the vial with a mixture solution of Zn(NO 3 ) 2 ·6H 2 O (10 mM), HMTA (5 mM), NH 3 ·H 2 O (0.35 M), and PEI (5 mM). The process was conducted within 3 h at the same temperature. After the growth is terminated, the substrate was washed with DI water and annealed at 150 • C for 10 min.
ZnO NWA and SNWs gas sensors were fabricated for sensor performance comparison. For ZnO NWA devices, the prepatterned substrate was spin-coated with 5 mM zinc acetate dehydrate Zn(CH 3 COO) 2 ·2H 2 O solution in acetone at 1000 rpm for 30 s. The spin-cast layer on the substrate was cured on a hot plate at 150 • C for 5 min to stabilize the film structure. The spin-coating and curing processes were repeated five times to obtain a uniform film, which served as the seeding layer. Afterward, the film was thermally annealed at 350 • C for 30 min, and then allowed to cool down. The thermal decomposition (of the zinc acetate) created ZnO nanocrystals on the substrate that act as a seed layer for subsequent ZnO array growth. The precursor solution for the hydrothermal reaction consists of Zn(NO 3 ) 2 ·6H 2 O(25 mM), HMTA (12.5 mM), PEI, (5 mM), and NH 3 ·H 2 O (0.35 M). The seeded substrate was then placed in a vial that contains 15 mL of the growth solution. The vial was covered and then placed in an oven and heated at 85 • C for 7 h. after the growth is terminated, the substrate was then rinsed with DI water and dried in air at 150 • C for 30 min.
SNWs devices were fabricated by drop-casting a solution containing ZnO NWs on substrates with Au prepatterned electrodes. The substrate was annealed at 200 • C for 2 h for better contact between the NWs and the electrodes.
The morphology and crystal structure of as-prepared products were observed using Philips XL-20 scanning electron microscope at 10 kV. The electrical characteristics and gas sensing properties of the fabricated devices were recorded using a probe station in a homemade chamber attached to a Keithley 4200 semiconductor analyzer. The sensor response, S, is defined as S = (I g − I a )/I a , where I g is the current sensor value in the tested gas environment and I a is the current value in air. The measurements were performed under a fixed bias of 2 V. The response time, t r , is defined as the time required for the current to reach 90% of the equilibrium value after injecting the gas, and the recovery time, t d , is defined as the time necessary for the sensor to return to 10% above the original current value in the air after releasing the gas from the test chamber.

Results and Discussion
Three ZnO gas sensors with different device structures, NWA, SNW, and BNW, were fabricated using hydrothermal synthesis, and their gas sensing properties were tested. Due to the critical impact of the operating temperature on the ZnO gas sensors, all fabricated sensors were tested at different temperatures to find the optimum condition for acetone gas sensing. Figure 1 shows the responses of all tested sensors to 200 ppm acetone as a function of the operating temperature. All sensors behave similarly in general. For all sensors, the sensitivity increases as the temperature increases, up to a certain point, and after that, the sensitivity starts to decrease with further increases in the temperature. ZnO NWA, SNW, and BNW sensors were most sensitive at 375, 325, and 275 • C, respectively. ZnO needs enough thermal energy to be activated and be able to sense. When reducing gases, such as acetone, are presented, the adsorbed oxygen on the ZnO surface takes part in the oxidation of acetone. The oxygen ions on the surface of ZnO react with the acetone molecules and give up electrons to the conduction band. At low operating temperature, the sensitivity Crystals 2021, 11, 743 4 of 11 is low, due to insufficient chemical activation [22]. Ionized oxygen is chemisorbed onto the surface of ZnO in its molecular form, O 2− which is less reactive [23]. At higher temperatures around the optimum, oxygen ion molecules are dissociated into oxygen ions with singly, O − , or doubly negative electric charges, O 2− , by attracting an electron from the conduction band of the ZnO. This substantial resistance increase, due to the depletion of ZnO by the adsorbed oxygen enhances the sensitivity of ZnO gas sensors [24,25]. These two processes, thermal excitation of electrons and chemisorption of oxygen species, carry on until the whole ZnO surface is covered with oxygen species, and this is where the optimum operating temperature is reached. After this stage, no more oxygen species can be absorbed, and electrons thermal excitation overcomes, leading to lower sensitivity, as seen in Figure 1. In fact, working at very high temperatures improves the activation significantly and forces the adsorbed gas molecules on the sensor surface to depart before getting the chance to exchange charges with the surface [4,7]. NWA, SNW, and BNW sensors were most sensitive at 375, 325, and 275 °C, respectively. ZnO needs enough thermal energy to be activated and be able to sense. When reducing gases, such as acetone, are presented, the adsorbed oxygen on the ZnO surface takes part in the oxidation of acetone. The oxygen ions on the surface of ZnO react with the acetone molecules and give up electrons to the conduction band. At low operating temperature, the sensitivity is low, due to insufficient chemical activation [22]. Ionized oxygen is chemisorbed onto the surface of ZnO in its molecular form, O 2− which is less reactive [23]. At higher temperatures around the optimum, oxygen ion molecules are dissociated into oxygen ions with singly, O − , or doubly negative electric charges, O 2− , by attracting an electron from the conduction band of the ZnO. This substantial resistance increase, due to the depletion of ZnO by the adsorbed oxygen enhances the sensitivity of ZnO gas sensors [24,25]. These two processes, thermal excitation of electrons and chemisorption of oxygen species, carry on until the whole ZnO surface is covered with oxygen species, and this is where the optimum operating temperature is reached. After this stage, no more oxygen species can be absorbed, and electrons thermal excitation overcomes, leading to lower sensitivity, as seen in Figure 1. In fact, working at very high temperatures improves the activation significantly and forces the adsorbed gas molecules on the sensor surface to depart before getting the chance to exchange charges with the surface [4,7].  Figure 2a,b depicts two SEM images of the NWA devices. One can see the wellaligned ZnO NWs and how uniform they are. A schematic of the NWA device is presented in Figure 2c. The sensitivity characteristics of the NWA sensor are shown in Figure 2d. The response of the sensor to 200 ppm of acetone gas at 375 °C is about 2.7. The sensor took 42 s to respond and 56 s to recover to its original state. The two electrodes in this device are connected by the seed layer under the NW array that consists of a polycrystalline thin film with nanoparticles ranging between (300-400 nm). The device response is stable, but it takes very long to reach its baseline again. Moreover, it showed poor sensitivity and extended the response and recovery times. Furthermore, in this device's fabrication procedure, the seed-layer needed to be annealed at a relatively high temperature (400 °C ), which does not allow the use of flexible substrates.
ZnO SNWs gas sensor was also tested at its optimum operating temperature under an acetone gas concentration of 200 ppm. SEM images, schematic, and gas sensing characteristics of the ZnO SNWs sensor are shown in Figure 3. The sensitivity of this device to acetone has been significantly improved relative to the previous ZnO NWA sensor. The response to 200 ppm of acetone is about 26 and the response time and recovery-time are 6 and 8 s, respectively. The enhancement in this device performance over the NWA sensor   Figure 2c. The sensitivity characteristics of the NWA sensor are shown in Figure 2d. The response of the sensor to 200 ppm of acetone gas at 375 • C is about 2.7. The sensor took 42 s to respond and 56 s to recover to its original state. The two electrodes in this device are connected by the seed layer under the NW array that consists of a polycrystalline thin film with nanoparticles ranging between (300-400 nm). The device response is stable, but it takes very long to reach its baseline again. Moreover, it showed poor sensitivity and extended the response and recovery times. Furthermore, in this device's fabrication procedure, the seed-layer needed to be annealed at a relatively high temperature (400 • C), which does not allow the use of flexible substrates.
ZnO SNWs gas sensor was also tested at its optimum operating temperature under an acetone gas concentration of 200 ppm. SEM images, schematic, and gas sensing characteristics of the ZnO SNWs sensor are shown in Figure 3. The sensitivity of this device to acetone has been significantly improved relative to the previous ZnO NWA sensor. The response to 200 ppm of acetone is about 26 and the response time and recovery-time are 6 and 8 s, respectively. The enhancement in this device performance over the NWA sensor was anticipated because of the reduced size of the active sensing part in the SNWs sensor and increased surface to volume ratio [26,27]. was anticipated because of the reduced size of the active sensing part in the SNWs sensor and increased surface to volume ratio [26,27].   was anticipated because of the reduced size of the active sensing part in the SNWs sensor and increased surface to volume ratio [26,27].  Our goal in this study is to develop a more efficient device structure that can improve sensor performance and be fabricated using flexible and transparent substrates. To reach this goal we should achieve the following: (1) Create better contact between the NWs and the electrodes by growing the NWs on the electrodes without any separating seed layer, (2) control the charge transport by including (NW-NW junction) where the effect of gas present at the on the potential barrier is greatest, and (3) reduce the size of the active sensing material.
To achieve these tasks, we had to modify our hydrothermal growth method. The growth solution and the substrate were altered to produce ZnO NWs directly and selectively on the electrodes, as shown in the SEM images in Our goal in this study is to develop a more efficient device structure that can improve sensor performance and be fabricated using flexible and transparent substrates. To reach this goal we should achieve the following: (1) Create better contact between the NWs and the electrodes by growing the NWs on the electrodes without any separating seed layer, (2) control the charge transport by including (NW-NW junction) where the effect of gas present at the on the potential barrier is greatest, and (3) reduce the size of the active sensing material.
To achieve these tasks, we had to modify our hydrothermal growth method. The growth solution and the substrate were altered to produce ZnO NWs directly and selectively on the electrodes, as shown in the SEM images in  From our experiments and the literature [21], we know that increasing the concentration of ammonium hydroxide in the growth environment can improve the heterogeneous nucleation of ZnO on Au surfaces, and consequently, the NWs' density the cost of slowing the growth process. Once again, we decided to use PEI to accelerate and enhance the growth process of ZnO NWs. However, we discovered that adding PEI from the start, before the nucleation occurs, interrupts the whole growth process. A similar observation was also reported by another study in the literature [28]. Thus, we decided to introduce PEI gradually after the nucleation occurs after 30 min from the beginning of the growth process.
The modified hydrothermal synthesis method was used to fabricate ZnO BNWs sensors on flexible and transparent substrates. SEM images at different magnifications and schematic of the BNWs sensor are presented in Figure 5a-d, respectively. In these SEM images, one can see the bridging NWs and the contact nano junctions between them. The BNWs sensor performance characteristics are shown in the plot in Figure 5e. The sensor response to 200 ppm acetone is around 63. The response-time and the recovery-time of the BNWs device are 0.42 and 0.90 s, respectively. The BNWs sensor is faster than the NWA 105 times and 15 times faster than the SNWs sensor. The enhancement in the BNWs sensor performance stems from several factors, including the more efficient device structure with junctions between the NWs and stronger contact between the NWs and the electrodes.  From our experiments and the literature [21], we know that increasing the concentration of ammonium hydroxide in the growth environment can improve the heterogeneous nucleation of ZnO on Au surfaces, and consequently, the NWs' density the cost of slowing the growth process. Once again, we decided to use PEI to accelerate and enhance the growth process of ZnO NWs. However, we discovered that adding PEI from the start, before the nucleation occurs, interrupts the whole growth process. A similar observation was also reported by another study in the literature [28]. Thus, we decided to introduce PEI gradually after the nucleation occurs after 30 min from the beginning of the growth process.
The modified hydrothermal synthesis method was used to fabricate ZnO BNWs sensors on flexible and transparent substrates. SEM images at different magnifications and schematic of the BNWs sensor are presented in Figure 5a-d, respectively. In these SEM images, one can see the bridging NWs and the contact nano junctions between them. The BNWs sensor performance characteristics are shown in the plot in Figure 5e. The sensor response to 200 ppm acetone is around 63. The response-time and the recovery-time of the BNWs device are 0.42 and 0.90 s, respectively. The BNWs sensor is faster than the NWA 105 times and 15 times faster than the SNWs sensor. The enhancement in the BNWs sensor performance stems from several factors, including the more efficient device structure with junctions between the NWs and stronger contact between the NWs and the electrodes.  Figure 6 shows the sensitivity of the different fabricated gas sensors as a function of gas concentration. Each sensor was tested at its optimum working temperature. The order of sensitivities was as follows: BNW > SNW > NWA along with the whole concentration range. The ZnO BNW sensor was the most sensitive, while the NWA sensor displayed the lowest. When the concentration of acetone exceeded 500 ppm, all sensors started to show a sign of saturation at different levels. Table 1 compares the performance of the different fabricated gas sensors in this work to ZnO gas sensors of different morphologies from the literature toward acetone gas.  Figure 6 shows the sensitivity of the different fabricated gas sensors as a function of gas concentration. Each sensor was tested at its optimum working temperature. The order of sensitivities was as follows: BNW > SNW > NWA along with the whole concentration range. The ZnO BNW sensor was the most sensitive, while the NWA sensor displayed the lowest. When the concentration of acetone exceeded 500 ppm, all sensors started to show a sign of saturation at different levels. Table 1 compares the performance of the different fabricated gas sensors in this work to ZnO gas sensors of different morphologies from the literature toward acetone gas.  Figure 6 shows the sensitivity of the different fabricated gas sensors as a function of gas concentration. Each sensor was tested at its optimum working temperature. The order of sensitivities was as follows: BNW > SNW > NWA along with the whole concentration range. The ZnO BNW sensor was the most sensitive, while the NWA sensor displayed the lowest. When the concentration of acetone exceeded 500 ppm, all sensors started to show a sign of saturation at different levels. Table 1 compares the performance of the different fabricated gas sensors in this work to ZnO gas sensors of different morphologies from the literature toward acetone gas. The generation and transportation of charges inside the SNWs sensor and the BNWs sensor are schematically represented in Figure 7. The scheme also compares the transport mechanisms in the SNWs and BNWs sensors. The sensing mechanism of ZnO nanostruc-tures depends significantly on the change in the nanostructure resistance, which is governed mainly by gas molecules adsorption and desorption processes on the surface [10][11][12][13][14][15][16][17][18][22][23][24][25][26]. This mechanism can be explained by the depletion region modulation model, as shown in Figure 7 [29,30]. In ambient air, the ZnO nanostructure adsorbs oxygen species on the surface, attracting free electrons from the conduction band and forming chemisorbed oxygen species, as shown in the first row of Figure 7. There are different oxygen species, such as molecular (O 2 − ) and atomic (O − , O 2− ) ions on the surface, depending on the operating temperature. For metal oxides in general, a relatively high temperature (150-500 • C) is required to reach a reasonable sensing performance. The process of attracting and seizing free electrons from the conduction band extends the width of the depletion region near the surface, and consequently, increases the resistance of the nanostructure, as represented in the second row of Figure 7. When acetone gas molecules are introduced in the gas chamber, as shown in the third row of Figure 7, they go through an oxidation process with the chemisorbed oxygen species on the surface. This oxidation process results in the release of all previously seized electrons from the conduction band as summarized in the following Equation (1) [31]: Crystals 2021, 11, x FOR PEER REVIEW 9 of 11

Conclusions
In conclusion, we reported a controlled seedless and site-selective hydrothermal synthesis method to produce ZnO NWs. Gas sensors with different device structures were This reaction reduces the width of the depletion region near the surface of the ZnO nanostructure and consequently decreases its resistance. On the ZnO nanostructure surface, oxygen vacancies behave like donors, while adsorbed oxygen ions act as surface acceptors. The more oxygen vacancies available on the surface of the ZnO nanostructure, the better its interaction with gas molecules and subsequently its gas sensing properties [29]. The conduction channel in the BNWs sensor includes NW-NW junctions, and charges need to cross them for the current to flow. The barrier consists of the surface depletion layers from both NWs. The expression for the ZnO NW conductance is as follow: where n 0 is the free charge carrier density, µ is the charge carriers' mobility, D is the NW diameter, t c is the thickness of the surface charge region, l is the length between the two electrodes. The thickness of the depletion region t c is expressed as: where L D is the Debye length and V S represents the adsorbate induced band bending. The Debye length is expressed as: where sults in the release of all previously seized electrons from the conduction band as summarized in the following Equation (1) [31]: This reaction reduces the width of the depletion region near the surface of the ZnO nanostructure and consequently decreases its resistance. On the ZnO nanostructure surface, oxygen vacancies behave like donors, while adsorbed oxygen ions act as surface acceptors. The more oxygen vacancies available on the surface of the ZnO nanostructure, the better its interaction with gas molecules and subsequently its gas sensing properties [29].
The conduction channel in the BNWs sensor includes NW-NW junctions, and charges need to cross them for the current to flow. The barrier consists of the surface depletion layers from both NWs. The expression for the ZnO NW conductance is as follow: where n0 is the free charge carrier density, µ is the charge carriers' mobility, D is the NW diameter, tc is the thickness of the surface charge region, l is the length between the two electrodes. The thickness of the depletion region tc is expressed as: where LD is the Debye length and VS represents the adsorbate induced band bending. The Debye length is expressed as: where ɛ is the relative dielectric permittivity [32]. The charges to flow between the NWs must first cross the depletion regions that act as barriers. Thus, for the NWs with diameters greater than the Debye length, the major modulator of charge flow in the BNWs sensors is the junction between the connected NWs. The improvement in response-time and recovery-time of the BNWs sensors can be ascribed mainly to the domination of the NW-NW junction barrier over the charge transport. Oxygen diffusion and adsorption again to deplete the NW channel takes a very long time. Therefore, if the sensor conductance is controlled by an NW conductance only, then it will take too long for the charge to recover. On the other hand, the conductance of the BNWs sensor is controlled more efficiently by the NW-NW junction barriers, where the flow of charges across the junction barrier is exponentially related to the barrier height. That is why the flow of charges is so sensitive to changes in the barrier. As soon as the gas gets out of the chamber, the electron-hole pairs recombine immediately, and the number of free charges in the NW will reduce accordingly, causing a swift increase in the effective barrier. Therefore, the junction barrier can lower the flowing current efficiently. The modulation of the barrier height is much faster than the oxygen diffusion process [2].
is the relative dielectric permittivity [32]. The charges to flow between the NWs must first cross the depletion regions that act as barriers. Thus, for the NWs with diameters greater than the Debye length, the major modulator of charge flow in the BNWs sensors is the junction between the connected NWs. The improvement in response-time and recovery-time of the BNWs sensors can be ascribed mainly to the domination of the NW-NW junction barrier over the charge transport. Oxygen diffusion and adsorption again to deplete the NW channel takes a very long time. Therefore, if the sensor conductance is controlled by an NW conductance only, then it will take too long for the charge to recover. On the other hand, the conductance of the BNWs sensor is controlled more efficiently by the NW-NW junction barriers, where the flow of charges across the junction barrier is exponentially related to the barrier height. That is why the flow of charges is so sensitive to changes in the barrier. As soon as the gas gets out of the chamber, the electron-hole pairs recombine immediately, and the number of free charges in the NW will reduce accordingly, causing a swift increase in the effective barrier. Therefore, the junction barrier can lower the flowing current efficiently. The modulation of the barrier height is much faster than the oxygen diffusion process [2].

Conclusions
In conclusion, we reported a controlled seedless and site-selective hydrothermal synthesis method to produce ZnO NWs. Gas sensors with different device structures were fabricated on-chip and tested. We demonstrated that by controlling the grown NWs, by tuning the growth parameters, the device performance can improve significantly. Using this method, high performance gas sensors were fabricated on flexible and transparent substrates and were compared with two other sensors having different device structures, the NWA and SNWs devices. The BNWs gas sensor demonstrated improved sensitivity toward acetone, excellent detectability, and short response-time (420 ms) and recovery-time 900 ms). The enhancement in response and recovery times of the BNWs device was ascribed to the NW-NW junction barrier dominated resistance, and the direct contact between the ZnO and the Au electrode was unavailable for the other devices.