Sensing Properties of NiO Loaded SnO2 Nanoparticles—Specific Selectivity to H2S

NiO-loaded SnO2 powders were prepared involving two chemical procedures. The mesoporous SnO2 support was synthesized by a hydrothermal route using Brij 35 non-ionic surfactant as a template. The nickel loadings of 1 and 10 wt.%. NiO were deposited by the wet impregnation method. The H2S sensing properties of xNiO-(1-x)SnO2 (x = 0, 1, 10%) thick layers deposited onto commercial substrates have been investigated with respect to different potential interfering gases (NO2, CO, CO2, CH4, NH3 and SO2) over a wide range of operating temperatures and relative humidity specific for in-field conditions. Following the correlation of the sensing results with the morphological ones, 1wt.% NiO/SnO2 was selected for simultaneous electrical resistance and work function investigations. The purpose was to depict the sensing mechanism by splitting between specific changes over the electron affinity induced by the surface coverage with hydroxyl dipoles and over the band bending induced by the variable surface charge under H2S exposure. Thus, it was found that different gas-interaction partners are dependent upon the amount of H2S, mirrored through the threshold value of 5 ppm H2S, which from an applicative point of view, represents the lower limit of health effects, an eight-hour TWA.


Introduction
Semiconducting metal oxide (SMOX)-based gas sensors are one of the widest spread devices for the detection of different explosive and toxic gases, mainly due to their high sensitivity and low manufacturing costs [1]. When targeting applicative demands, properties such as sensitivity and selectivity should be enhanced to boost the potential development. Thus, one of the currently used strategies consists of developing heterojunctions aimed to improve the gas sensing performances of p-type SMOX sensors [2].
One of the most important aspects in preparing a heterojunction resides in the inner electrical properties of the materials involved, ready to take part in an equilibrium process after the junction. Thus, each part comes with a well-defined Fermi level which upon contact will be subject to a subsequent charge transfer through the interface in order to establish an energetic equilibrium. Consequently, a potential barrier will be formed at the interface between the involved SMOX materials. The appearance of either oxidizing or reducing gases in the surrounding atmosphere will lead to the modulation of the p−n junction directly reflected in the sensor signal of the heterojunction materials [3].
Under the real operating conditions (i.e., in-field conditions) moisture and interfering gases are always present and relevant for developing realistic gas sensing applications. Such aspects should be considered with respect to today's challenges [4]. The role of water corrosive and extremely toxic gas. The toxicity of H 2 S can be comparable with that of hydrogen cyanide, as a broad-spectrum poison and most of the SMOX materials suffer from the slow recovery transients in spite of higher sensitivity [16]. Because of their low fabrication costs and low level of H 2 S detection, SMOX based gas sensors have gained the attention of the scientific community as promising future monitoring systems. It was demonstrated that p−n heterojunctions act better towards H 2 S detection than simple single metal oxide components. Usually the base matrix material consists of a well-known n-type SnO 2 in combination with different p-type materials, such as CuO. One of the reasons resides in the complexity of the roles played by these two oxides towards sensing and transduction phenomena. However, one of their major drawbacks resides in the strong interference by moisture, downgrading the overall sensor signal. Therefore, NiO is considered a promotor in taking the role of RH upon itself, thus leaving the H 2 S reaction to its n-type SMOX partner. This is the main reason why the NiO-SnO 2 heterojunction was chosen for H 2 S detection under real operating conditions.
In this paper, an enhancement of the H 2 S gas response in the presence of moisture using xNiO-(1-x) SnO 2 (x = 0, 1, 10%) based gas sensors is reported and the associated gas sensing mechanism is described throughout via phenomenological investigations involving simultaneous electrical resistance and work function measurements.

Materials Preparation and Sensors Fabrication
SnO 2 sensing powders, 1 and 10 wt.% NiO loaded, were prepared with mesoporous SnO 2 support involving two preparation procedures. The SnO 2 support was obtained by a hydrothermal method using Brij 35 non-ionic surfactant as template. In the first step the surfactant was dissolved in water under vigorous stirring for 2h. An aqueous solution of tin (IV) chloride pentahydrate (SnCl 4 5H 2 O) 98% was added dropwise. The pH was adjusted to 2, using nitric acid (HNO 3 ). The mixture was loaded into an autoclave equipped with Teflon liner and heat treated at 180 • C for 24 h. The autogenous pressure was~25 atm. The precipitate was washed with water, dried at 80 • C and finally calcined to 550 • C in air. The desired nickel loadings of 1 and 10 wt.%. NiO were deposited by wetness impregnation method. The corresponding amount of Ni(NO 3 ) 2 hydrate was dissolved in deionized water and added dropwise over support, followed by ultrasonication for 5 min, drying at room temperature for 24 h and thermal treatment at 400 • C, in air, for the nitrate decomposition leading to NiO formation.
In order to deposit the active material on top of the commercial alumina sensors, screen printing technique has been used. Thus, the powders were mixed with 1,2 propanediol and the obtained paste was deposited onto commercial alumina (Al 2 O 3 ) substrates provided with Pt electrodes and backside heater. The substrates consisted of nine interdigitated Pt fingers (with 200 µm in between) displayed in a 7 × 3.5 mm aspect ratio on the front side, being necessary for the measurement of the sensor electrical resistance changes. On the other side, a Pt heater was designed in order to keep the sensor at the desired operating temperature. The obtained sensors were labeled: SnO 2 , SnO 2 -1%NiO and SnO 2 -10%NiO.

Structural, Morphological and Surface Chemistry Investigations
X-ray diffraction (XRD) patterns of the powdery samples were recorded with the Bruker D8 Advance X-ray diffractometer (λ = 0.154184 nm) in Bragg−Brentano configuration. All XRD measurements were performed with the same experimental parameters in a wide 2θ range (10-140 • ) for a high accuracy of the structural data (lattice parameters, average crystallite size) obtained by Rietveld refinement (Topas v.3 software).
A JEOL2100 instrument equipped with JEOL EDS (energy dispersive X-ray) detector was used for CTEM (conventional transmission electron microscopy), SAED (selected area diffraction), HRTEM (high resolution transmission electron microscopy) and STEM (scanning transmission electron microscopy) investigations, regarding morphological, structural and chemical properties of NiO loaded SnO 2 systems. X-ray photoelectron spectroscopy (XPS) was carried out on PHI Quantera equipment with a base pressure in the analysis chamber of 10 −9 Torr and the monochromatized Al Kα radiation (1486.6 eV). For calibration, we used C1s line (BE = 284.8 eV) characteristic for the adsorbed hydrocarbon on the surface sample (C−C or (CH)n bonding). A dual beam (electrons and Ar ions) was used as neutralizer in order to overcome the charging effect.

Gas Sensing Investigations
In order to characterize the electrical resistance changes of the gas sensors, a special computer-controlled gas mixing system (GMS) for delivering the target gas concentrations in a reproducible way was used. Gas sensors have been mounted into a four-socket Teflon (PTFE) gas cell provided with stainless-steel electrical connections and Viton sealing in order to prevent possible outgassing. The computer-controlled mass-flow meters and electrovalves allowed the desired concentrations of the H 2 S and other potential interfering gases to be obtained under constant gas flow, miming the in-field working conditions. The general scheme for the GMS is presented in Figure 1.
a wide 2θ range (10-140°) for a high accuracy of the structural data (lattice parameters, average crystallite size) obtained by Rietveld refinement (Topas v.3 software).
A JEOL2100 instrument equipped with JEOL EDS (energy dispersive X-ray) detector was used for CTEM (conventional transmission electron microscopy), SAED (selected area diffraction), HRTEM (high resolution transmission electron microscopy) and STEM (scanning transmission electron microscopy) investigations, regarding morphological, structural and chemical properties of NiO loaded SnO2 systems.
X-ray photoelectron spectroscopy (XPS) was carried out on PHI Quantera equipment with a base pressure in the analysis chamber of 10 −9 Torr and the monochromatized Al Kα radiation (1486.6 eV). For calibration, we used C1s line (BE = 284.8 eV) characteristic for the adsorbed hydrocarbon on the surface sample (C−C or (CH)n bonding). A dual beam (electrons and Ar ions) was used as neutralizer in order to overcome the charging effect.

Gas Sensing Investigations
In order to characterize the electrical resistance changes of the gas sensors, a special computer-controlled gas mixing system (GMS) for delivering the target gas concentrations in a reproducible way was used. Gas sensors have been mounted into a four-socket Teflon (PTFE) gas cell provided with stainless-steel electrical connections and Viton sealing in order to prevent possible outgassing. The computer-controlled mass-flow meters and electrovalves allowed the desired concentrations of the H2S and other potential interfering gases to be obtained under constant gas flow, miming the in-field working conditions. The general scheme for the GMS is presented in Figure 1. The gas flow throughout the system was kept constant for all measurements at 200 mL/min. The GMS operated with high purity gases (5.0) from cylinders while the relative humidity (RH) was generated by passing the carrier gas (dry synthetic air) through a vaporizer. By adjusting the air flow, the desired RH level was attained. In order to gain insights about the possible gas sensing mechanism towards H2S detection, electrical resistance and work function changes were recorded simultaneously with a Keithley 6517A electrometer (DC) and McAllister KP 6500 Kelvin Probe (ΔCPD). The Kelvin Probe (KP) measured the contact potential difference (CPD) established between the sensitive layer and KP metallic tip. While the latter is gas inert, the changes in CPD induced by the variations in the test gas atmosphere represents the layer's relative work function modifications according to the relation (1).
where: ΔΦ represents the work function changes; ΔCPD represents the contact potential difference and q is the elementary charge. The gas flow throughout the system was kept constant for all measurements at 200 mL/min. The GMS operated with high purity gases (5.0) from cylinders while the relative humidity (RH) was generated by passing the carrier gas (dry synthetic air) through a vaporizer. By adjusting the air flow, the desired RH level was attained. In order to gain insights about the possible gas sensing mechanism towards H 2 S detection, electrical resistance and work function changes were recorded simultaneously with a Keithley 6517A electrometer (DC) and McAllister KP 6500 Kelvin Probe (∆CPD). The Kelvin Probe (KP) measured the contact potential difference (CPD) established between the sensitive layer and KP metallic tip. While the latter is gas inert, the changes in CPD induced by the variations in the test gas atmosphere represents the layer's relative work function modifications according to the relation (1).
where: ∆Φ represents the work function changes; ∆CPD represents the contact potential difference and q is the elementary charge. Figure 2 shows the XRD patterns of the samples, highlighting the absence of any secondary phase up to the highest doping level, in the detection limit of 1-2%. All three sam-ples presented similar XRD patterns, indexed as tetragonal SnO 2 , space group: P42/mnm (136), ICDD-04-014-0193. The XRD pattern for undoped SnO 2 was identical to the pattern of the 1% NiO doped SnO 2 sample.

XRD Results
Chemosensors 2021, 9, x FOR PEER REVIEW 5 of 16 Figure 2 shows the XRD patterns of the samples, highlighting the absence of any secondary phase up to the highest doping level, in the detection limit of 1-2%. All three samples presented similar XRD patterns, indexed as tetragonal SnO2, space group: P42/mnm (136), ICDD-04-014-0193. The XRD pattern for undoped SnO2 was identical to the pattern of the 1% NiO doped SnO2 sample. The calculated lattice parameters were identical for all the samples: a = 0.4739 ± 0.0001 nm and c = 0.3186 ± 0.0001 nm. This result is expected when Ni 2+ ions substitute Sn 4+ ions, as the effective ionic radii of Sn 4+ and Ni 2+ are equal (0.069 nm) in octahedral coordination [17]. The average crystallite size was similar in the unloaded and 1% NiO loaded SnO2 samples: d = 21.5 ± 0.5 nm, slightly decreasing to d = 20.5 ± 0.5 nm in the 10% NiO loaded SnO2 sample. Such a variation is practically within the errors limit, being smaller than expected for the 10% loading level.

Morphological Results
The images in Figure 3 show, for all the samples, mono-crystalline grains in the 20 nm range, some of them with extended defects. The SAED pattern revealed the tetragonal SnO2 structure, with no observable reflections from other crystal structures. The calculated lattice parameters were identical for all the samples: a = 0.4739 ± 0.0001 nm and c = 0.3186 ± 0.0001 nm. This result is expected when Ni 2+ ions substitute Sn 4+ ions, as the effective ionic radii of Sn 4+ and Ni 2+ are equal (0.069 nm) in octahedral coordination [17]. The average crystallite size was similar in the unloaded and 1% NiO loaded SnO 2 samples: d = 21.5 ± 0.5 nm, slightly decreasing to d = 20.5 ± 0.5 nm in the 10% NiO loaded SnO 2 sample. Such a variation is practically within the errors limit, being smaller than expected for the 10% loading level.

Morphological Results
The images in Figure 3 show, for all the samples, mono-crystalline grains in the 20 nm range, some of them with extended defects. The SAED pattern revealed the tetragonal SnO 2 structure, with no observable reflections from other crystal structures.
Size distributions ( Figure 4) provided more quantitative information regarding the morphology, i.e., the SnO 2 +1%NiO sample had a mean crystal size of 20.5 nm with a standard deviation of 6.7 nm while the other had a mean crystal size of 18 nm with a standard deviation of 7.1 nm. A statistic for more than 150 NPs performed on the unloaded SnO 2 sample (histogram not shown) indicated a mean crystal size of~20.7 with a standard deviation of 7.28.
The chemical composition as obtained by EDS spectra ( Figure 5) provided a %at. Sn:O:Ni stoichiometry of~29:61:0.13 for SnO 2 +1%NiO and of~34:63.5:2.3 for SnO 2 +10%NiO. In the reference sample, the EDS provided a %at. Sn:O ratio of~33.3:66.6, very close to the SnO 2 stoichiometry, with no Ni signal present in the spectra. Chemosensors 2021, 9, x FOR PEER REVIEW 6 of 16 Size distributions ( Figure 4) provided more quantitative information regarding the morphology, i.e., the SnO2+1%NiO sample had a mean crystal size of 20.5 nm with a standard deviation of 6.7 nm while the other had a mean crystal size of 18 nm with a standard deviation of 7.1 nm. A statistic for more than 150 NPs performed on the unloaded SnO2 sample (histogram not shown) indicated a mean crystal size of ~20.7 with a standard deviation of 7.28.   Size distributions ( Figure 4) provided more quantitative information regarding the morphology, i.e., the SnO2+1%NiO sample had a mean crystal size of 20.5 nm with a standard deviation of 6.7 nm while the other had a mean crystal size of 18 nm with a standard deviation of 7.1 nm. A statistic for more than 150 NPs performed on the unloaded SnO2 sample (histogram not shown) indicated a mean crystal size of ~20.7 with a standard deviation of 7.28.  While the inherent EDS error can be significant and most of the time increased by imperfect acquisition conditions, the NiO loading was significantly lower than its nominal value. It is worth mentioning that the experimental conditions were optimized, as observed from the spectroscopic analysis on pure SnO 2 .
The spatial distribution of elements as provided by EDS coupled STEM analysis, showed that in the case of highest doped sample, a certain segregation of Ni-rich entities (whether it was Ni or NiO was difficult to tell) took place. Correlated with the size distributions, a natural decrease of the mean crystal size with the increase of the loading could be observed. While the decrease was still uncertain for the case of SnO 2 +1%NiO, for the higher doping where segregation appeared, it became noticeable. Consequently, even though not observed in the diffraction patterns, the potential formation of N-rich secondary crystalline phases (which may contribute to the crystal size evaluations) needs to be considered for further investigations.
value. It is worth mentioning that the experimental conditions were optimized, as observed from the spectroscopic analysis on pure SnO2.
The spatial distribution of elements as provided by EDS coupled STEM analysis, showed that in the case of highest doped sample, a certain segregation of Ni-rich entities (whether it was Ni or NiO was difficult to tell) took place. Correlated with the size distributions, a natural decrease of the mean crystal size with the increase of the loading could be observed. While the decrease was still uncertain for the case of SnO2+1%NiO, for the higher doping where segregation appeared, it became noticeable. Consequently, even though not observed in the diffraction patterns, the potential formation of N-rich secondary crystalline phases (which may contribute to the crystal size evaluations) needs to be considered for further investigations.

Surface Chemistry Results
The surface chemistry of the sensors was investigated by X-ray photolectron spectroscopy (XPS). Thus, in the Figures 6a−c and 7a,b are shown the high-resolution spectra recorded for the most prominent transitions of the elements detected on the surface, as follows: Ni2p3, Sn3d, C1s and O1s, respectively. Ni was detected on the surface as Ni 2+ at BE ~855.6 eV. It is worth mentioning that the BE shifted toward higher BE compared with the standard NiO (BE in the range 854-855 eV according to NIST XPS databases) [18] is attributed to the interaction of the NiO cu SnO2 lattice. Sn was detected on the surface as Sn4+ characteristic to SnO2, for the BE of Sn3d5/2 at ~486.4 eV. We assessed the surface contamination by recording the C1s HR spectra for all sensors by superimposed C1s spectra as shown in Figure 6c. We found out that the surface of the sensitive layers was not contaminated except for an unavoidable, low amount of carbon which was attributed mainly to the adventitious carbon adsorbed from the atmosphere (C-C and CHn). This finding clearly proves that the organic part from both synthesis routes and the deposition procedure of the sensitive layers was completely removed by the thermal treatment.

Surface Chemistry Results
The surface chemistry of the sensors was investigated by X-ray photolectron spectroscopy (XPS). Thus, in the Figures 6a−c and 7a,b are shown the high-resolution spectra recorded for the most prominent transitions of the elements detected on the surface, as follows: Ni2p3, Sn3d, C1s and O1s, respectively. Ni was detected on the surface as Ni 2+ at BE~855.6 eV. It is worth mentioning that the BE shifted toward higher BE compared with the standard NiO (BE in the range 854-855 eV according to NIST XPS databases) [18] is attributed to the interaction of the NiO cu SnO 2 lattice. Sn was detected on the surface as Sn4+ characteristic to SnO 2 , for the BE of Sn3d5/2 at~486.4 eV. We assessed the surface contamination by recording the C1s HR spectra for all sensors by superimposed C1s spectra as shown in Figure 6c. We found out that the surface of the sensitive layers was not contaminated except for an unavoidable, low amount of carbon which was attributed mainly to the adventitious carbon adsorbed from the atmosphere (C-C and CHn). This finding clearly proves that the organic part from both synthesis routes and the deposition procedure of the sensitive layers was completely removed by the thermal treatment.
In Figure 7a are depicted the normalized O1s spectra for SnO 2 , SnO 2 -1%NiO and SnO 2 -10%NiO. The oxygen chemistry highlighted by O1s analysis revealed no significant changes to the surface after NiO deposition. By the deconvolution process, we found the largest amount of oxygen bound in the SnO 2 lattice (~75%), OH adsorbed groups (~21%), as well as a tiny amount of water (~4%) (Figure 7b). It can be seen that SnO 2 exhibited a slightly more hydroxylated surface. Table 1 shows the XPS data which comprise the surface composition and the corresponding BEs (eV) assessed by X-ray photoelectron spectroscopy, from the HR spectra. The Ni content was found to be close to the nominal value percentage as follows:~0.95 atom %, equivalent to~1. In Figure 7a are depicted the normalized O1s spectra for SnO2, SnO2-1%NiO and SnO2-10%NiO. The oxygen chemistry highlighted by O1s analysis revealed no significant changes to the surface after NiO deposition. By the deconvolution process, we found the largest amount of oxygen bound in the SnO2 lattice (~75%), OH adsorbed groups (~21%), as well as a tiny amount of water (~4%) (Figure 7b). It can be seen that SnO2 exhibited a slightly more hydroxylated surface.  Table 1 shows the XPS data which comprise the surface composition and the corresponding BEs (eV) assessed by X-ray photoelectron spectroscopy, from the HR spectra. The Ni content was found to be close to the nominal value percentage as follows:~0.95 atom %, equivalent to ~1.2 wt.% for SnO2-1%NiO and ~5.2 atom %, equivalent to ~6.5 wt.%. The experimental errors in the XPS data quantification are in the range of ±10%, while for the BEs values of ±0.2 eV.

Gas Sensing Properties
It is known [19,20] that the gas sensing properties of a sensitive material are strongly related to the operating temperature through the nature of the surface adsorbed species further involved in the gas detection processes. The changes in the surrounding relative humidity (%RH) must be considered when an applicative potential is demanded [21]. Therefore, we pursued the evaluation of the relative humidity influence upon the electrical resistance of xNiO-(1-x)SnO 2 (x = 0, 1, 10%) through the whole range of operating temperatures, directly linked to the specific surface interactions involving free charge carrier exchange. As such, the corresponding difference between dry air atmosphere and air with 50% RH (as the accepted average value for in-field conditions) has been evaluated. On the other hand, the differences induced by the various RH levels were considered to be references for the following changes induced by the gas−surface interactions with different gas noxes, potentially present in the surrounding atmosphere (see Figure 8).

Gas Sensing Properties
It is known [19,20] that the gas sensing properties of a sensitive material are strongly related to the operating temperature through the nature of the surface adsorbed species further involved in the gas detection processes. The changes in the surrounding relative humidity (%RH) must be considered when an applicative potential is demanded [21]. Therefore, we pursued the evaluation of the relative humidity influence upon the electrical resistance of xNiO-(1-x)SnO2 (x = 0, 1, 10%) through the whole range of operating temperatures, directly linked to the specific surface interactions involving free charge carrier exchange. As such, the corresponding difference between dry air atmosphere and air with 50% RH (as the accepted average value for in-field conditions) has been evaluated. On the other hand, the differences induced by the various RH levels were considered to be references for the following changes induced by the gas−surface interactions with different gas noxes, potentially present in the surrounding atmosphere (see Figure 8). The doping level with a p-type MOX material (NiO) of the base n-type SnO2 is reflected through its electrical resistance dependence with respect to the operating temperature. Accordingly, one can see that the electrical resistance increases with the increase in the amount (%) of NiO loading, independent of the RH level. Moreover, the RH influence decreases with the increase in the amount of NiO loading, while the influence of RH level is negligible over the whole range of operating temperatures. A possible explanation is related to the fact that the presence of moisture in the surrounding atmosphere decreases the electrical resistance of the base SnO2 material as a consequence of hemolytic dissociation [22] in a hydroxyl group OHwhich can share its electronic pair with the lattice cation (usually Snlatt) and in a weakly bounded proton H + that may easily react with the lattice oxygen according to the following relation: Herein, the presence of Ni 2+ cations are responsible for OH group trapping, leaving Sn 2+ available for subsequent gas surface interaction.
In order to identify the gas-specific finger print of (1-x) SnO2 -xNiO (x = 0, 1, 10%) a common gas testing protocol involving different gases (NO2, CO, CO2, CH4, NH3, SO2 and H2S) was addressed over the whole range of operating temperatures (Figure 9). All the concentrations of the specific interfering gases were chosen according to the EU exposure limits. The sensor signal was defined as:  The doping level with a p-type MOX material (NiO) of the base n-type SnO 2 is reflected through its electrical resistance dependence with respect to the operating temperature. Accordingly, one can see that the electrical resistance increases with the increase in the amount (%) of NiO loading, independent of the RH level. Moreover, the RH influence decreases with the increase in the amount of NiO loading, while the influence of RH level is negligible over the whole range of operating temperatures. A possible explanation is related to the fact that the presence of moisture in the surrounding atmosphere decreases the electrical resistance of the base SnO 2 material as a consequence of hemolytic dissociation [22] in a hydroxyl group OHwhich can share its electronic pair with the lattice cation (usually Sn latt ) and in a weakly bounded proton H + that may easily react with the lattice oxygen according to the following relation: Herein, the presence of Ni 2+ cations are responsible for OH group trapping, leaving Sn 2+ available for subsequent gas surface interaction.
In order to identify the gas-specific finger print of (1-x) SnO 2 -xNiO (x = 0, 1, 10%) a common gas testing protocol involving different gases (NO 2 , CO, CO 2 , CH 4 , NH 3 , SO 2 and H 2 S) was addressed over the whole range of operating temperatures (Figure 9). All the concentrations of the specific interfering gases were chosen according to the EU exposure limits. The sensor signal was defined as: S red = R air /R gas for reducing gases (3) and S ox = R gas /R air for oxidizing gases (4) where R air is the electrical resistance under reference atmosphere and R gas is the electrical resistance under test gas conditions. mosensors 2021, 9, x FOR PEER REVIEW 11 of 16 and S ox = R gas /R air for oxidizing gases (4) where Rair is the electrical resistance under reference atmosphere and Rgas is the electrical resistance under test gas conditions. Depending on the nature of the target gas (reducing or oxidizing), the calculated sensor signals varied with respect to the operating temperature in the range of 1 to 30 for all investigated gas types, except for H2S. In this case, the maximum signal occurred at 200 °C for SnO2-1%NiO (S~1000) and SnO2-10%NiO (S~940) and highlighted the specific selectivity to H2S of both NiO loaded materials. Correlating the sensing results with the morphological ones that highlighted the appearance of segregation for the highest doped  Figure 9. Sensor signal of (1-x) SnO 2 -xNiO (x = 0, 1, 10%) for: 3 ppm NO 2 (a); 50 ppm CO (b); 3000 ppm CO 2 (c); 2500 ppm CH 4 (d); 50 ppm NH 3 (e); 7 ppm SO 2 (f) and 10 ppm H 2 S (g), over a wide range of operating temperatures and fixed 50%RH.
Depending on the nature of the target gas (reducing or oxidizing), the calculated sensor signals varied with respect to the operating temperature in the range of 1 to 30 for all investigated gas types, except for H 2 S. In this case, the maximum signal occurred at 200 • C for SnO 2 -1%NiO (S~1000) and SnO 2 -10%NiO (S~940) and highlighted the specific selectivity to H 2 S of both NiO loaded materials. Correlating the sensing results with the morphological ones that highlighted the appearance of segregation for the highest doped NiO based SnO 2 material, SnO 2 -1%NiO was selected for supplementary investigations involving simultaneous electrical resistance and work function measurements under H 2 S exposure [23]. Thus, the surface contact potential differences (∆CPD) of SnO 2 -1%NiO exposed to different H 2 S concentrations under 50% RH and operated at 200 • C, bring insights about the surface adsorbed species, which may or may not exchange free charge carriers with the investigated material but in turn contribute to the coverage of the surface with specific dipolar species. Accordingly, the change in the work function (∆Φ) has two components, which are directly influenced by the gas−surface interactions, i.e., electron affinity (∆χ) and band bending (q∆Vs) changes while the electrochemical potential (µ) remains constant (Figure 10a). The former is influenced by the net coverage in surface with dipoles while the latter is influenced by the changes in the net surface charge [24]. specific dipolar species. Accordingly, the change in the work function (∆Φ) has two components, which are directly influenced by the gas−surface interactions, i.e., electron affinity (Δχ) and band bending (qΔVs) changes while the electrochemical potential (µ) remains constant (Figure 10a). The former is influenced by the net coverage in surface with dipoles while the latter is influenced by the changes in the net surface charge [24].
With the above-mentioned discussion one can write the following relation: where: Experimentally, the contact potential difference (ΔCPD) and subsequently the changes between Rair and RH2S were measured, thus allowing quantitative information to be obtained about the changes in electron affinity (Δχ) (see Figure 10b). The work function behavior with respect to the H2S concentration resembles the one of electron affinity, i.e., with an abrupt increase for the first two gas concentrations, reaching maximum, then followed by a monotonous decrease (crossing the abscissa) until a steady state is accomplished for the last concentrations. Considering the operating temperature of 200 °C, a possible explanation is given by the interplay between the hydroxyl groups and surface oxygen species, the role of Ni 2+ cations being to take over the additional surface hydroxylation. As depicted from the XPS investigations ( Figure 7b) the largest amount of oxygen is bound in the SnO2 lattice being available for H2S interaction with respect to the following equation: With the above-mentioned discussion one can write the following relation: where: Experimentally, the contact potential difference (∆CPD) and subsequently the changes between R air and R H2S were measured, thus allowing quantitative information to be obtained about the changes in electron affinity (∆χ) (see Figure 10b).
The work function behavior with respect to the H 2 S concentration resembles the one of electron affinity, i.e., with an abrupt increase for the first two gas concentrations, reaching maximum, then followed by a monotonous decrease (crossing the abscissa) until a steady state is accomplished for the last concentrations. Considering the operating temperature of 200 • C, a possible explanation is given by the interplay between the hydroxyl groups and surface oxygen species, the role of Ni 2+ cations being to take over the additional surface hydroxylation. As depicted from the XPS investigations ( Figure 7b) the largest amount of oxygen is bound in the SnO 2 lattice being available for H 2 S interaction with respect to the following equation: Accordingly, for H 2 S concentrations below 5 ppm, the surface hydroxylation dominates, reflected by increasing electronic affinity. As H 2 S concentration increases, the (∆χ) slightly decreases indicating a change in the reaction partners until a steady state is ob-tained, strengthening the idea of having different interaction mechanisms. The changes occurring in the potential barrier (q∆Vs) are described by a monotonous decrease up to 5 ppm of H 2 S concentration, followed by a turning point of the slope for the highest concentrations. Such behavior can be attributed to the strong reducing potential of H 2 S over the preadsorbed oxygen species, with the release of free charge carriers in the conduction band.
As mentioned before (5), there is a direct relationship between the surface band bending (q∆Vs) and sensor signal via: q∆Vs = k B Tln(S H2S ) → S H2S = exp q∆Vs k B T allowing the calculation of the sensor signal dependencies with respect to the H 2 S concentrations.
As can be observed in Figure 11a, the sensor signal of the investigated material shows a good linear relationship for the lowest H 2 S concentrations, indicating promising involvement towards applicative potential with H 2 S gas sensors. It should be noted that the threshold value of 5 ppm H 2 S represents, from the application point of view, the lower limit of health effects, an eight-hour TWA.
Accordingly, for H2S concentrations below 5 ppm, the surface hydroxylation dominates, reflected by increasing electronic affinity. As H2S concentration increases, the (Δχ) slightly decreases indicating a change in the reaction partners until a steady state is obtained, strengthening the idea of having different interaction mechanisms. The changes occurring in the potential barrier (qΔVs) are described by a monotonous decrease up to 5 ppm of H2S concentration, followed by a turning point of the slope for the highest concentrations. Such behavior can be attributed to the strong reducing potential of H2S over the preadsorbed oxygen species, with the release of free charge carriers in the conduction band.
As mentioned before (5), there is a direct relationship between the surface band bending (qΔVs) and sensor signal via: qΔVs = k B Tln(S H2S ) → S H2S = exp ( qΔVs k B T ) allowing the calculation of the sensor signal dependencies with respect to the H2S concentrations.
As can be observed in Figure 11a, the sensor signal of the investigated material shows a good linear relationship for the lowest H2S concentrations, indicating promising involvement towards applicative potential with H2S gas sensors. It should be noted that the threshold value of 5 ppm H2S represents, from the application point of view, the lower limit of health effects, an eight-hour TWA. On the other hand, we call the first derivative of the sensor signal in order to enhance the resolution of possible inflection points. As illustrated in Figure 11b, the peak center of the derivative matched with the 5 ppm of H2S concentration, indicating a turning point in the detection mechanism as seen through the simultaneous electrical resistance and work function measurements.
Additionally, in order to compare the results for H2S detection with different types of SMOX gas sensors, we have proceeded to a literature review ( Table 2). On the other hand, we call the first derivative of the sensor signal in order to enhance the resolution of possible inflection points. As illustrated in Figure 11b, the peak center of the derivative matched with the 5 ppm of H 2 S concentration, indicating a turning point in the detection mechanism as seen through the simultaneous electrical resistance and work function measurements.
Additionally, in order to compare the results for H 2 S detection with different types of SMOX gas sensors, we have proceeded to a literature review ( Table 2).
As can be observed, the interest has been focused on low operating temperature and high sensitivity to H 2 S, involving different types of heterostructures, but all the investigations were performed under ideal laboratory conditions. From an applicative perspective, the right solution is to address different gas sensing approaches by attempting to exploit the sensitive materials' performances close to the in-field conditions. Therefore, the take-off herein depicted, was to highlight the novelty of nickel impregnation of base SnO 2 as main mediator within the oxygen−water interplay. In addition to the moisture influence over the sensing performances, the present work also considered the possible cross-sensitivity issues with other potential interfering gases present in the surrounding atmosphere.
Transients and stability have been also evaluated. The response time is defined as the time necessary for the electrical resistance of the sensor to attain a threshold level of 90% from the steady state conditions in a test gas situation and the recovery time is defined as the time necessary to return to 90% from the steady state conditions after the gas stimulus has ended. By considering the fact that the gas flow conditions and test chamber volume are constant during the measurements, the aforementioned characteristics strongly depend on the surface equilibria where the gas interaction takes place [30]. As for the investigated SnO 2 -1%NiO sensitive material, the response times were found to vary between 3 and 31 min with minima for the lowest and the highest concentrations, as well as for the turning point at 5 ppm of H 2 S. The recovery time was found to be around 30 min (Figure 12a). As can be observed, the interest has been focused on low operating temperature and high sensitivity to H2S, involving different types of heterostructures, but all the investigations were performed under ideal laboratory conditions. From an applicative perspective, the right solution is to address different gas sensing approaches by attempting to exploit the sensitive materials' performances close to the in-field conditions. Therefore, the takeoff herein depicted, was to highlight the novelty of nickel impregnation of base SnO2 as main mediator within the oxygen−water interplay. In addition to the moisture influence over the sensing performances, the present work also considered the possible cross-sensitivity issues with other potential interfering gases present in the surrounding atmosphere.
Transients and stability have been also evaluated. The response time is defined as the time necessary for the electrical resistance of the sensor to attain a threshold level of 90% from the steady state conditions in a test gas situation and the recovery time is defined as the time necessary to return to 90% from the steady state conditions after the gas stimulus has ended. By considering the fact that the gas flow conditions and test chamber volume are constant during the measurements, the aforementioned characteristics strongly depend on the surface equilibria where the gas interaction takes place [30]. As for the investigated SnO2-1%NiO sensitive material, the response times were found to vary between 3 and 31 min with minima for the lowest and the highest concentrations, as well as for the turning point at 5 ppm of H2S. The recovery time was found to be around 30 min ( Figure  12a). Aiming to find an explanation for such behavior, in the case of thick, porous films is not an easy task. Therefore, most probably the interaction mechanism involves either different percolation paths or various reaction partners for H2S detection [31]. Aiming to find an explanation for such behavior, in the case of thick, porous films is not an easy task. Therefore, most probably the interaction mechanism involves either different percolation paths or various reaction partners for H 2 S detection [31].
It is known that the stability of a sensitive material is an important parameter when targeting a potential application. Herein, the base line resistance was monitored under dynamic air flow conditions with 50% RH at the optimum operating temperature. Figure  12b shows the electrical resistance changes during several periods of time when the sensor was operated under normal in-field conditions. The sensor's resistance slightly decreased over eight days of investigation. Such an aspect is quite normal for a SMOX-based gas sensor when operated against a humid background and can be attributed both to the surface bonded hydroxyl species and to the common aging effect.

Conclusions
Sensitive materials based on xNiO-(1-x)SnO 2 (x = 0, 1, 10%) have been prepared by NiO wet impregnation of SnO 2 obtained by a hydrothermal chemical route assisted by Brij 35 as non-ionic surfactant. An operating temperature screening was performed in order to establish the electrical resistance behavior of said materials, relative to the amount (%) of NiO and to the RH level of the simulated in-field atmosphere. Thus, it was possible to identify the role of NiO loading over the RH influence for the whole range of operating temperatures. The specific selectivity to H 2 S was demonstrated by using a general gas sensing protocol involving other potential interfering gases such as: NO 2 , CO, CO 2 , CH 4 , NH 3 and SO 2 over the whole range of operating temperatures with specific concentrations in accordance with the European exposure limits. The selected 1 wt.% NiO/SnO 2 was subjected to simultaneous electrical resistance and work function investigations. Thus, we could understand the way in which different concentrations of H 2 S induced different reaction mechanisms with a turning point at the threshold limit of 5 ppm H 2 S. The good linear relationship for lower H 2 S concentrations under in-field conditions indicates a promising applicative potential.