Gas Sensing of Laser-Produced Hybrid TiO2-ZnO Nanomaterials under Room-Temperature Conditions

The preparation method can considerably affect the structural, morphological, and gas-sensing properties of mixed-oxide materials which often demonstrate superior photocatalytic and sensing performance in comparison with single-metal oxides. In this work, hybrids of semiconductor nanomaterials based on TiO2 and ZnO were prepared by laser ablation of Zn and Ti plates in water and then tested as chemiresistive gas sensors towards volatile organics (2-propanol, acetaldehyde, ethanol, methanol) and ammonia. An infrared millisecond pulsed laser with energy 2.0 J/pulse and a repetition rate of 5 Hz was applied to Zn and Ti metal targets in different ablation sequences to produce two nano-hybrids (TiO2/ZnO and ZnO/TiO2). The surface chemistry, morphology, crystallinity, and phase composition of the prepared hybrids were found to tune their gas-sensing properties. Among all tested gases, sample TiO2/ZnO showed selectivity to ethanol, while sample ZnO/TiO2 sensed 2-propanol at room temperature, both with a detection limit of ~50 ppm. The response and recovery times were found to be 24 and 607 s for the TiO2/ZnO sensor, and 54 and 50 s for its ZnO/TiO2 counterpart, respectively, towards 100 ppm of the target gas at room temperature.


Preparation of Hybrid ZnO-TiO 2 Nanomaterials
The experimental setup used to prepare materials is schematically shown in Figure 1. A millisecond pulsed Nd:YAG laser with a wavelength of 1064 nm, pulse peak power of 1 kW, pulse width of 2 ms, and repetition rate of 5 Hz was applied to ablate metal targets placed in a quartz cuvette. The beam was focused on the target surface by a lens with a focal length of 9.0 cm, with the diameter of ablated area being~150 µm. More detailed descriptions of the setup and procedures used can be found elsewhere [35]. A zinc plate (99.5% purity, 2 mm in thickness) was fixed in the cuvette filled with 15 mL of deionized water, after which the target was ablated for 30 min. Then, the Zn plate was replaced with a Ti plate (99.5% purity, 0.5 mm thick), and irradiation was continued for another 30 min. The prepared sample was denoted as ZnO/TiO 2 ( Figure 1). The second sample, TiO 2 /ZnO, was prepared by first ablating Ti plate in 15 mL of deionized water for 30 min, followed by ablation of Zn plate in the TiO 2 dispersion for another 30 min ( Figure 1). During the preparation, the temperature of dispersions was found to rise up tõ 67 • C, which is explained by heating caused by millisecond-long pulses. The as-prepared samples were centrifuged, the supernatant was removed, and NPs were concentrated in a volume of 1 mL. Then, the colloids were drop-cast on copper grids, on Si wafers, and on interdigitated electrodes for transmission electron microscopy (TEM), X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), and gas sensing test, respectively.
tailed descriptions of the setup and procedures used can be found elsewhere [35]. A zinc plate (99.5% purity, 2 mm in thickness) was fixed in the cuvette filled with 15 mL of deionized water, after which the target was ablated for 30 min. Then, the Zn plate was replaced with a Ti plate (99.5% purity, 0.5 mm thick), and irradiation was continued for another 30 min. The prepared sample was denoted as ZnO/TiO2 ( Figure 1). The second sample, TiO2/ZnO, was prepared by first ablating Ti plate in 15 mL of deionized water for 30 min, followed by ablation of Zn plate in the TiO2 dispersion for another 30 min ( Figure  1). During the preparation, the temperature of dispersions was found to rise up to ~67 °C, which is explained by heating caused by millisecond-long pulses. The as-prepared samples were centrifuged, the supernatant was removed, and NPs were concentrated in a volume of 1 mL. Then, the colloids were drop-cast on copper grids, on Si wafers, and on interdigitated electrodes for transmission electron microscopy (TEM), X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), and gas sensing test, respectively.

Characterization of Hybrids
TEM observations were carried out in a Hitachi HF-2200 microscope. Both samples were examined by X-ray photoelectron spectroscopy (XPS, Quantum 2000, ULVAC-PHI, Chigasaki, Japan) and X-ray powder diffraction (XRD, D8 Discover from Bruker, Leipzig, Germany), for which nanomaterials were dropcast on Si wafers.

Sensing Measurements
All the VOCs used in gas sensing tests, namely 2-propanol, acetaldehyde, ethanol, methanol, and ammonia solution, were purchased from Merck, India. To prepare sensors, the laser-prepared nanomaterials were dropcast on commercially available interdigitated electrodes (G-IDEAU5) supplied by Metrohm-DropSens (Oviedo, Spain). Gas sensing characteristics of the prepared samples were studied using a custom-made gas testing chamber integrated with high resistance electrometer (Keithley 6517B, Tektronix, Beaverton, OR, USA) [43,44]. The electrical contacts were established on interdigitated electrodes using gold wire bonder with a wire thickness of 20 μm [44]. The sensing elements were then placed inside the gas sensing chamber and connected to the high-resistance electrometer with a bias voltage of 5 V. The gas testing chamber was filled with ambient air, and resistance was measured to establish the baseline resistance (Ra) for the sensing studies.

Characterization of Hybrids
TEM observations were carried out in a Hitachi HF-2200 microscope. Both samples were examined by X-ray photoelectron spectroscopy (XPS, Quantum 2000, ULVAC-PHI, Chigasaki, Japan) and X-ray powder diffraction (XRD, D8 Discover from Bruker, Leipzig, Germany), for which nanomaterials were dropcast on Si wafers.

Sensing Measurements
All the VOCs used in gas sensing tests, namely 2-propanol, acetaldehyde, ethanol, methanol, and ammonia solution, were purchased from Merck, India. To prepare sensors, the laser-prepared nanomaterials were dropcast on commercially available interdigitated electrodes (G-IDEAU5) supplied by Metrohm-DropSens (Oviedo, Spain). Gas sensing characteristics of the prepared samples were studied using a custom-made gas testing chamber integrated with high resistance electrometer (Keithley 6517B, Tektronix, Beaverton, OR, USA) [43,44]. The electrical contacts were established on interdigitated electrodes using gold wire bonder with a wire thickness of 20 µm [44]. The sensing elements were then placed inside the gas sensing chamber and connected to the high-resistance electrometer with a bias voltage of 5 V. The gas testing chamber was filled with ambient air, and resistance was measured to establish the baseline resistance (R a ) for the sensing studies.
The concentration of target gases was determined following the protocol and formulas previously published elsewhere [43,44]. A chromatographic syringe was used to inject the desired volume of targeted volatile organic compounds (VOCs) into the test chamber. Subsequently, the sensing element's resistance was monitored continuously until it reached a steady state value in the presence of desired concentration of VOC. The static liquid-gas distribution was adapted to calculate the concentration of VOCs from the injected target analytes, as given in Equation (1) [43]: where Q is the liquid volume of VOC (L), V is the volume of the gas testing chamber (L), ϕ is the required gas volume fraction, d is the specific gravity of the VOC solution (g/cm 3 ), p is the purity of the specific VOC solutions, M is the molecular weight of the target molecule (g/mol), and T R and T B are the temperatures of ambient and testing chambers environment ( • C). Once the sensing element attained a steady state, the chamber was exposed to the ambient atmosphere to ensure the sensing element's recovery behavior, and the element's response characteristics were continuously recorded. This measurement was carried out for the desired concentrations of target analytes. The gas sensing response was determined by comparing the electrical resistance of the sensing element in dry air and in gas environment by means of Equation (2): where R a and R v are the resistance values of the sensing element in dry air and in presence of target gas, respectively.

Preparation and Characterization of Hybrids
The preparation of hybrid nanomaterials used in this study is schematically illustrated in Figure 1. Sample ZnO/TiO 2 was prepared as follows. At first, a Zn metal plate immersed in deionized water was ablated by a millisecond pulsed laser, which is known to produce ZnO nanorods [35]. Then, the Zn plate was replaced with a Ti one to be further ablated (in presence of already formed ZnO nanorods) for another 30 min (Figure 1). In a similar way, a hybrid sample TiO 2 /ZnO was fabricated: first by ablating a Ti target in deionized water to form TiO 2 NPs, after which a Zn target was ablated in their presence.
The phase composition of as-prepared materials was studied by XRD analysis. The XRD patterns for both samples are shown in Figure 2. One can clearly see eight sharp peaks at 31.7 • , 34.4 • , 36.2 • , 47.5 • , 56.5 • , 62.8 • , 67.7 • , and 69.8 • which correspond to the planes of the hexagonal wurtzite phase of ZnO (PDF 01-089-1397). In both hybrid nanomaterials, titania is seen to show broad low-intensity peaks assigned to three crystalline phases: anatase (PDF 00-001-0562), rutile (PDF 01-072-7374), and brookite (PDF 01-076-1934), all being indicated with circles and squares in Figure 2. TiO 2 patterns are more pronounced for sample ZnO/TiO 2 (red pattern) where titania was produced in presence of ZnO NPs. The peaks for TiO 2 in sample TiO 2 /ZnO are weaker, indicating lower crystallinity of titania, most probably due to secondary laser irradiation of TiO 2 NPs during the second step of this hybrid formation.
TEM images of samples ZnO/TiO 2 and TiO 2 /ZnO are shown in Figure 3a,b, respectively. It is seen that in both hybrids, ZnO NPs are presented as nanorods, while TiO 2 NPs are spherical. Both metal oxides are formed as two separate phases, which are well-mixed in homogeneous solid materials. Previously, Honda et al. reported on size-controlled ZnO nanorods prepared under different conditions by means of a millisecond laser (pulse width of 0.5, 1, and 2 ms and peak energy of 1 and 5 kW) [35]. According to Figure 3b, the formation and growth of ZnO NPs in titania-containing medium (sample TiO 2 /ZnO) tends to provide ZnO rods with a larger aspect ratio (longer and thinner nanorods), more specifically, with their length varying from 50 to 190 nm and their width ranging from 10 to 40 nm. Meanwhile, bigger and uniform ZnO nanorods with a length of 120-210 nm and width 25-55 nm were formed in sample ZnO/TiO 2 where ZnO NPs were then irradiated during the second stage for another 30 min. This can be explained by a longer heating period of the ZnO nanorods formed in sample ZnO/TiO 2 , in good agreement with previous reports [15,35]. TEM images of samples ZnO/TiO2 and TiO2/ZnO are shown in Figure 3a,b tively. It is seen that in both hybrids, ZnO NPs are presented as nanorods, while T are spherical. Both metal oxides are formed as two separate phases, which are we in homogeneous solid materials. Previously, Honda et al. reported on size-control nanorods prepared under different conditions by means of a millisecond lase width of 0.5, 1, and 2 ms and peak energy of 1 and 5 kW) [35]. According to Figur formation and growth of ZnO NPs in titania-containing medium (sample TiO2/Zn to provide ZnO rods with a larger aspect ratio (longer and thinner nanorods), m cifically, with their length varying from 50 to 190 nm and their width ranging fr 40 nm. Meanwhile, bigger and uniform ZnO nanorods with a length of 120-210 width 25-55 nm were formed in sample ZnO/TiO2 where ZnO NPs were then ir during the second stage for another 30 min. This can be explained by a longer period of the ZnO nanorods formed in sample ZnO/TiO2, in good agreement wi ous reports [15,35].
Ablation of metallic Ti in water was typically reported to produce spherical T [22,23,26]. The diameter of such TiO2 NPs observed in sample ZnO/TiO2 is in a n range (5-30 nm) than those formed in sample TiO2/ZnO (5-65 nm) [15]. This can p be attributed to the presence of ZnO nanorods in the liquid medium, which restr crystal growth of TiO2. The wider size distribution of TiO2 NPs in sample TiO2/Zn also be explained by secondary irradiation of TiO2 NPs during ablation of Zn plat dispersion (i.e., the second preparation stage) when they were subjected to ad fragmentation and melting-aggregation. Such fusion of TiO2 NPs is usually accom by a loss of crystallinity, as revealed by XRD patterns in Figure 2. A similar concl the effect of secondary irradiation of titania NPs also comes from XPS data (pres Figure 4) that confirm changes in the surface composition of analyzed NPs.  In Ti 2p XPS spectra, the Ti 2p3/2 and Ti 2p1/2 peaks are centered at 458.8 eV and 464.6 eV for sample ZnO/TiO2 (Figure 4a, top) and at 458.7 eV and 464.4 eV for sample TiO2/ZnO (Figure 4a, bottom). These values are typical for Ti 4+ ions in titania and confirm the formation of rutile and anatase as main phases in both hybrids. As depicted in Figure 4a, the Ti 2p3/2 peak of material TiO2/ZnO is much wider than that of its counterpart ZnO/TiO2. It can be fitted with three Gaussian peaks: (i) a main peak at 458.7 eV assigned to Ti 4+ ions [19,24], (ii) a peak located at lower binding energy (457.7 eV) and related to Ti 3+ ions [20,45,46], and (iii) a lowest-energy peak at 456.7 eV associated with Ti 2+ ions (Figure 4a, bottom) [19,23]. Ablation of metallic Ti in water was typically reported to produce spherical TiO 2 NPs [22,23,26]. The diameter of such TiO 2 NPs observed in sample ZnO/TiO 2 is in a narrower range (5-30 nm) than those formed in sample TiO 2 /ZnO (5-65 nm) [15]. This can probably be attributed to the presence of ZnO nanorods in the liquid medium, which restricted the crystal growth of TiO 2 . The wider size distribution of TiO 2 NPs in sample TiO 2 /ZnO might also be explained by secondary irradiation of TiO 2 NPs during ablation of Zn plate in TiO 2 dispersion (i.e., the second preparation stage) when they were subjected to additional fragmentation and melting-aggregation. Such fusion of TiO 2 NPs is usually accompanied by a loss of crystallinity, as revealed by XRD patterns in Figure 2. A similar conclusion on the effect of secondary irradiation of titania NPs also comes from XPS data (presented in Figure 4) that confirm changes in the surface composition of analyzed NPs. (Figure 4a, bottom). These values are typical for Ti 4+ ions in titania and confirm the formation of rutile and anatase as main phases in both hybrids. As depicted in Figure 4a, the Ti 2p3/2 peak of material TiO2/ZnO is much wider than that of its counterpart ZnO/TiO2. It can be fitted with three Gaussian peaks: (i) a main peak at 458.7 eV assigned to Ti 4+ ions [19,24], (ii) a peak located at lower binding energy (457.7 eV) and related to Ti 3+ ions [20,45,46], and (iii) a lowest-energy peak at 456.7 eV associated with Ti 2+ ions (Figure 4a, bottom) [19,23]. Based on the XPS results, we conclude that in sample TiO2/ZnO, its TiO2 NPs are more laser-modified because of longer irradiation, which leads to the formation of surface Ti 3+ and Ti 2+ species well seen in Figure 4a. Such species may form not only by gaining electrons from plasma, but also from Zn atoms through a redox process where surface Ti 4+ ions get reduced. Comparison with the Ti 2p3/2 peak of sample ZnO/TiO2 (Figure 4a) shows  (Figure 4a, bottom). These values are typical for Ti 4+ ions in titania and confirm the formation of rutile and anatase as main phases in both hybrids. As depicted in Figure 4a, the Ti 2p 3/2 peak of material TiO 2 /ZnO is much wider than that of its counterpart ZnO/TiO 2 . It can be fitted with three Gaussian peaks: (i) a main peak at 458.7 eV assigned to Ti 4+ ions [19,24], (ii) a peak located at lower binding energy (457.7 eV) and related to Ti 3+ ions [20,45,46], and (iii) a lowest-energy peak at 456.7 eV associated with Ti 2+ ions (Figure 4a, bottom) [19,23].
Based on the XPS results, we conclude that in sample TiO 2 /ZnO, its TiO 2 NPs are more laser-modified because of longer irradiation, which leads to the formation of surface Ti 3+ and Ti 2+ species well seen in Figure 4a. Such species may form not only by gaining electrons from plasma, but also from Zn atoms through a redox process where surface Ti 4+ ions get reduced. Comparison with the Ti 2p 3/2 peak of sample ZnO/TiO 2 (Figure 4a) shows that ablation of Ti plate in a ZnO colloid gives TiO 2 NPs with a very low density of surface Ti 3+ species. This conclusion is also confirmed by the O1s XPS spectra shown in Figure 4b. The O1s peak of sample ZnO/TiO 2 was deconvoluted into three components (Figure 4b, top), with the most intensive peak at 530.0 eV assigned to O 2ions in the crystal structure of both metal oxides, ZnO and TiO 2 [47], the peak at 531.1 eV associated with surface oxygen vacancies [15], and the peak at 532.2 eV related to surface hydroxyl groups [35]. The same components, fitted at 530.0, 531.2, and 532.3 eV, are also seen in the spectrum of sample TiO 2 /ZnO (Figure 4b, bottom), where two additional peaks at lower binding energy (529.0 and 528.0 eV) appear, being associated with oxygen ions bonded to titanium in lower oxidation states, Ti(III) and Ti(II) (i.e., O 2− -Ti 3+ and O 2− -Ti 2+ bonds), in accordance with the signals for Ti 3+ and Ti 2+ ions observed in Ti 2p XPS spectra (Figure 4a).
The doublets consisting of Zn 2p 3/2 and Zn 2p 1/2 peaks are observed at 1021.8 eV and 1044.8 eV (for sample ZnO/TiO 2 ) and at 1022.0 eV and 1045.0 eV (for sample TiO 2 /ZnO), (see Figure 4c). These peaks clearly indicate Zn 2+ species in ZnO NPs available in both hybrids [35,37,47]. In addition, a weak signal is observed in sample TiO 2 /ZnO at a lower binding energy (1020.6 eV) (Figure 4c, bottom), which is most likely related to a trace amount of metallic Zn inclusions that possibly remained due to incomplete oxidation of Zn atoms in a TiO 2 -containing dispersion [35]. Thus, XPS analysis of both hybrids revealed that: (i) Zn is always presented in the form of ZnO; (ii) during the second stage of laser processing, Ti(IV) is reduced to Ti(III) and Ti(II) on the surface of sample TiO 2 /ZnO; (iii) self-doping of TiO 2 with Ti 3+ and Ti 2+ is confirmed by both O1s and Ti 2p XPS spectra.

Selectivity and Response Ratio
To test the gas-sensing properties of the prepared samples ZnO/TiO 2 and TiO 2 /ZnO at room temperature, their responses towards 100 ppm of different target vapors (2-propanol, acetaldehyde, ethanol, methanol, and ammonia) were studied as shown in Figure 5a. The hybrid TiO 2 /ZnO showed a selective response towards ethanol (S = 34.1), whereas its ZnO/TiO 2 counterpart exhibited selective sensing towards 2-propanol (S = 6.6). Response ratio [20,25], or the selectivity factor, was calculated using Equation (3), and the obtained values are presented in Figure 5b: Selectivity f actor = S target gas S Inter f ering gas (3) (see Figure 4c). These peaks clearly indicate Zn 2+ species in ZnO NPs available in both hybrids [35,37,47]. In addition, a weak signal is observed in sample TiO2/ZnO at a lower binding energy (1020.6 eV) (Figure 4c, bottom), which is most likely related to a trace amount of metallic Zn inclusions that possibly remained due to incomplete oxidation of Zn atoms in a TiO2-containing dispersion [35]. Thus, XPS analysis of both hybrids revealed that: (i) Zn is always presented in the form of ZnO; (ii) during the second stage of laser processing, Ti(IV) is reduced to Ti(III) and Ti(II) on the surface of sample TiO2/ZnO; (iii) self-doping of TiO2 with Ti 3+ and Ti 2+ is confirmed by both O1s and Ti 2p XPS spectra.

Selectivity and Response Ratio
To test the gas-sensing properties of the prepared samples ZnO/TiO2 and TiO2/ZnO at room temperature, their responses towards 100 ppm of different target vapors (2-propanol, acetaldehyde, ethanol, methanol, and ammonia) were studied as shown in Figure  5a. The hybrid TiO2/ZnO showed a selective response towards ethanol (S = 34.1), whereas its ZnO/TiO2 counterpart exhibited selective sensing towards 2-propanol (S = 6.6). Response ratio [20,25], or the selectivity factor, was calculated using Equation (3)

Transient Response and Limit of Detection (LOD)
The transient resistance response of the samples in the presence of varied concentrations of 2-propanol and ethanol is shown in Figure 6a,b, respectively. The logarithmic response was observed to increase linearly with the logarithmic concentration of both 2-propanol and ethanol vapors. The observed trends were fitted to the following linear equations: y (2-propanol) = 1.423x − 2.56 and y (ethanol) = 1.799x − 2.31 with R 2 values being 0.95 and 0.94 for 2-propanol and ethanol, respectively (see Figure 6c). Thus, the observed R 2 values clearly indicate the linear relationships in both cases. Both samples ZnO/TiO 2 and TiO 2 /ZnO showed the lowest detection limit of 50 ppm towards 2-propanol and ethanol vapors, respectively.

Response and Recovery Times
Both the response and recovery times were defined as the time taken for the sensor to reach 90% of maximum response and recover back to surface resistance in absence of target gas, respectively [48]. Such times for sample TiO 2 /ZnO were found to be 24 and 607 s, respectively, towards 100 ppm of ethanol. Meanwhile, the response and recovery times of sample ZnO/TiO 2 were found to be 54 and 50 s, respectively, towards 100 ppm of 2-propanol (Figure 7). Though the sensor based on hybrid TiO 2 /ZnO showed a selective detection of ethanol at room temperature, its recovery time was relatively slow compared with that of its ZnO/TiO 2 counterpart, which is probably related to the slow desorption rate of ethanol molecules at room temperature. We assume that a possible reason can be the formation of stronger hydrogen bonds between the OH group in ethanol molecules and surface hydroxyl groups available on hybrid material, while the molecule of 2-propanol experiences sterical hindrance in hydrogen bonding. The transient resistance response of the samples in the presence of varied concentrations of 2-propanol and ethanol is shown in Figure 6a,b, respectively. The logarithmic response was observed to increase linearly with the logarithmic concentration of both 2propanol and ethanol vapors. The observed trends were fitted to the following linear equations: y(2-propanol) = 1.423x − 2.56 and y(ethanol) = 1.799x − 2.31 with R 2 values being 0.95 and 0.94 for 2-propanol and ethanol, respectively (see Figure 6c). Thus, the observed R 2 values clearly indicate the linear relationships in both cases. Both samples ZnO/TiO2 and TiO2/ZnO showed the lowest detection limit of 50 ppm towards 2-propanol and ethanol vapors, respectively.

Response and Recovery Times
Both the response and recovery times were defined as the time taken for the sensor to reach 90% of maximum response and recover back to surface resistance in absence of target gas, respectively [48]. Such times for sample TiO2/ZnO were found to be 24 and 607 s, respectively, towards 100 ppm of ethanol. Meanwhile, the response and recovery times of sample ZnO/TiO2 were found to be 54 and 50 s, respectively, towards 100 ppm of 2propanol ( Figure 7). Though the sensor based on hybrid TiO2/ZnO showed a selective detection of ethanol at room temperature, its recovery time was relatively slow compared with that of its ZnO/TiO2 counterpart, which is probably related to the slow desorption rate of ethanol molecules at room temperature. We assume that a possible reason can be the formation of stronger hydrogen bonds between the OH group in ethanol molecules and surface hydroxyl groups available on hybrid material, while the molecule of 2-propanol experiences sterical hindrance in hydrogen bonding.

Impact of Relative Humidity
Relative humidity is one of the parameters known to influence sensors that operate at room temperature. Hence, the sensing response of samples ZnO/TiO2 and TiO2/ZnO towards 100 ppm of 2-propanol (or ethanol) at different% RH levels was also investigated (see Figure 8a). For this, the values of % RH were varied to 32% and 72% from the actual relative humidity of 56%, which was achieved by maintaining corresponding saturated salt solutions of MgCl2 and NaCl inside the sensing chamber [49]. Humidity was evaluated by a digital humidity and temperature sensor (DHT 11) integrated into the sensing chamber used [43,48]. The sensing response was found to increase by 10 and 28% at lower humidity (32% RH) for 2-propanol and ethanol, respectively (see Figure 8a). The increased sensing response observed at lower humidity is ascribed to the reduced hindrance sorption of OH − ions on the surface of the sensing element, which follows the hopping charge transport behavior [50][51][52][53]. At the same time, the response was decreased by 44 and 20% as shown in Figure 8a at higher humidity (of 72%RH) for 2-propanol and ethanol, respec-

Impact of Relative Humidity
Relative humidity is one of the parameters known to influence sensors that operate at room temperature. Hence, the sensing response of samples ZnO/TiO 2 and TiO 2 /ZnO towards 100 ppm of 2-propanol (or ethanol) at different% RH levels was also investigated (see Figure 8a). For this, the values of % RH were varied to 32% and 72% from the actual relative humidity of 56%, which was achieved by maintaining corresponding saturated salt solutions of MgCl 2 and NaCl inside the sensing chamber [49]. Humidity was evaluated by a digital humidity and temperature sensor (DHT 11) integrated into the sensing chamber used [43,48]. The sensing response was found to increase by 10 and 28% at lower humidity (32% RH) for 2-propanol and ethanol, respectively (see Figure 8a). The increased sensing response observed at lower humidity is ascribed to the reduced hindrance sorption of OH − ions on the surface of the sensing element, which follows the hopping charge transport behavior [50][51][52][53]. At the same time, the response was decreased by 44 and 20% as shown in Figure 8a at higher humidity (of 72%RH) for 2-propanol and ethanol, respectively. This may be attributed to the reduced sorption process in the presence of excess OH − ions on the surface of the sensing element, which follows the Grotthuss physisorption charge transfer mechanism [50][51][52][53].
Relative humidity is one of the parameters known to influence sensors that operate at room temperature. Hence, the sensing response of samples ZnO/TiO2 and TiO2/ZnO towards 100 ppm of 2-propanol (or ethanol) at different% RH levels was also investigated (see Figure 8a). For this, the values of % RH were varied to 32% and 72% from the actual relative humidity of 56%, which was achieved by maintaining corresponding saturated salt solutions of MgCl2 and NaCl inside the sensing chamber [49]. Humidity was evaluated by a digital humidity and temperature sensor (DHT 11) integrated into the sensing chamber used [43,48]. The sensing response was found to increase by 10 and 28% at lower humidity (32% RH) for 2-propanol and ethanol, respectively (see Figure 8a). The increased sensing response observed at lower humidity is ascribed to the reduced hindrance sorption of OH − ions on the surface of the sensing element, which follows the hopping charge transport behavior [50][51][52][53]. At the same time, the response was decreased by 44 and 20% as shown in Figure 8a at higher humidity (of 72%RH) for 2-propanol and ethanol, respectively. This may be attributed to the reduced sorption process in the presence of excess OH − ions on the surface of the sensing element, which follows the Grotthuss physisorption charge transfer mechanism [50][51][52][53].

Stability of Performance over Time
The stability of sensors based on samples ZnO/TiO2 and TiO2/ZnO was evaluated as their sensing was tested towards 100 ppm of corresponding target gases over a period of

Stability of Performance over Time
The stability of sensors based on samples ZnO/TiO 2 and TiO 2 /ZnO was evaluated as their sensing was tested towards 100 ppm of corresponding target gases over a period of 30 days, during which the sensors were tested every 5 days. Both sensing elements are seen in Figure 8b,c to show good reproducible results with minimal changes in R a and R v after 30 days. This reveals the long-term stability of both hybrids ZnO/TiO 2 and TiO 2 /ZnO as they were subjected to multiple sorption-desorption cycles.

Sensing Mechanism
The band gap of TiO 2 (3.2 eV) is smaller than that of ZnO (3.3 eV) and the electron affinity of TiO 2 (4.2 eV) is larger than that of ZnO (4.1 eV), which predicts the formation of n-n homo-junction at their interface. Isolated bands of materials ZnO and TiO 2 should exist in each material and a heterojunction barrier is generated at their interface during contact. The surface resistance related to the heterojunction barrier can be expressed by Equation (4): where R is the resistance corresponding to heterojunction barrier, B is the constant related to ambient temperature, Φ is the heterojunction barrier, T is the absolute temperature and k is the Boltzmann constant. When a sensing element is maintained in the ambient atmosphere, it experiences chemisorption of atmospheric oxygen on its surface (Equations (5) and (6)). The adsorption of conduction band electrons of the sensing element by atmospheric oxygen molecules results in the increased space charge width around each grain. This process results in increased surface resistance and the latter resistance is considered the baseline for sensing measurements. Upon interaction with reducing type target analytes, the addition of conduction band electrons to the sensing element results in the reduction of space charge width. In turn, this reduces the surface resistance of the sensing element (R a ). The possible vapor solid interactions involved are expressed in Equations (7) and (8). During the desorption process, the surface resistance of the sensing element again reaches its baseline as a reversible process. Table 1 compares the sensing performance of hybrid materials prepared in the present work with that of similar materials previously reported in the literature. It is seen that, in comparison with their counterparts previously reported in the literature, samples ZnO/TiO 2 and TiO 2 /ZnO are seen to exhibit selective detection of propanol and ethanol at room temperature.

Conclusions
In this work, using the green preparation method of pulsed laser ablation in the liquid phase, we prepared two hybrid nanomaterials based on ZnO and TiO 2 . To produce such samples, Zn and Ti metal plates were ablated in water in different sequences: (i) Zn target followed by Ti target in presence of already formed ZnO nanoparticles and (ii) Ti target followed by Zn target in presence of already formed TiO 2 nanoparticles. Both hybrid materials were found to have spherical TiO 2 and rod-like ZnO nanoparticles as their components, demonstrating similar morphology irrespective of the preparation approaches. However, the surface chemistry of both materials was affected by laser irradiation time and liquid media, showing the formation of reduced oxidation states (+3) and (+2) of Ti in parallel with Ti(+4) in TiO 2 . When used as chemiresistor gas sensing elements at room temperature, these hybrids showed selectivity either to ethanol or 2-propanol and long-term stability performance. Thus, laser-prepared hybrid metal oxide nanomaterials are shown to demonstrate selective gas sensing depending on their preparation protocol.

Data Availability Statement:
The data presented in this study are available on request from the corresponding authors.