Journal of Materials Science & Technology  2019 , 35 (10): 2232-2237 https://doi.org/10.1016/j.jmst.2019.06.005

Orginal Article

Gas sensing selectivity of oxygen-regulated SnO2 films with different microstructure and texture

Ruiwu Lia, Yanwen Zhoua, Maolin Suna, Zhen Gonga, Yuanyuan Guoa, Xitao Yina*, Fayu Wua*, Wutong Dingb

a School of Material Science and Metallurgy, University of Science and Technology Liaoning, Anshan, 114051, China
b State Key Laboratory of Advanced Processing and Recycling of Nonferrous Metals, Lanzhou University of Technology, Lanzhou, 730050, China

Corresponding authors:   *Corresponding authors.E-mail addresses: yxtaj@163.com (X. Yin), fayuwu@ustl.edu.cn (F. Wu).*Corresponding authors.E-mail addresses: yxtaj@163.com (X. Yin), fayuwu@ustl.edu.cn (F. Wu).

Received: 2019-05-13

Revised:  2019-05-30

Accepted:  2019-06-1

Online:  2019-10-05

Copyright:  2019 Editorial board of Journal of Materials Science & Technology Copyright reserved, Editorial board of Journal of Materials Science & Technology

More

Abstract

The selectivity of gas sensing materials is increasingly important for their applications. The oxygen-regulated SnO2 films with (110) and (101) preferred orientation were obtained through magnetron sputtering, followed by annealing treatment. Their micro-structure, surface morphology and gas response were investigated by advanced structural characterization and property measurement. The results showed that the as-prepared (110)-oriented SnO2 film was oxygen-rich and had more adsorption sites while the as-prepared (101)-oriented SnO2 film was oxygen-poor and more sensitive to de-oxidation. H2 gas sensitivity, response speed, selectivity between H2 and CO of the (110)-orientated SnO2 film was superior to that of the (101)-orientated SnO2 film. After treated at high temperature and high vacuum, the reduction of gas-sensing properties of the annealed (110) SnO2 film was much more than that of the annealed (101) SnO2 film. The lattice oxygen was responsible for the difference in gas-sensing response between (110) and (101)-oriented SnO2 films under oxygen regulation. This work indicated the gas-sensing selectivity of the different crystal planes in SnO2 film, providing a significant reference for design and extension of the related materials.

Keywords: Gas sensor ; SnO2 ; Oxygen regulation ; Preferred orientation ; H2 selectivity

0

PDF (2135KB) Metadata Metrics Related articles

Cite this article Export EndNote Ris Bibtex

Ruiwu Li, Yanwen Zhou, Maolin Sun, Zhen Gong, Yuanyuan Guo, Xitao Yin, Fayu Wu, Wutong Ding. Gas sensing selectivity of oxygen-regulated SnO2 films with different microstructure and texture[J]. Journal of Materials Science & Technology, 2019, 35(10): 2232-2237 https://doi.org/10.1016/j.jmst.2019.06.005

1. Introduction

As one of the most common n-type semiconductor, Tin dioxide (SnO2) nano-materials with different structure and morphology exhibit extensive physical and chemical properties, and have been widely used in transparent conductive electrodes, Lithium-Ion batteries, solar cells, and gas sensors [[1], [2], [3], [4]]. Particularly, SnO2 nanomaterials have been deeply studied in the detection of various gases in the light of the fast response, high sensitivity, good chemical stability and strong selectivity [5].

A few studies indicate that the gas sensitivity of SnO2 nanomaterials depends on their size and morphology. For instance, Chao et al. reported that the gas-sensitive properties of SnO2 sensors can be improved by controlling SnO2 hollow nano-crystalline tubes to increase the specific surface area and the number of sensing active sites [6]. Wang et al. studied the SnO2 nanorod-based composite films with improved gas sensing properties of formaldehyde and methylbenzene at room temperature [7]. Yin et al. observed that doping the noble metals (Au, Pd, Pt) into SnO2 can enhance the sensitivity, and found that Au, Pd loading and In and Fe doping increased the selectivity to CO, while Pt loading increased the selectivity to H2 [8]. Xue et al. reported that the ordered mesoporous SnO2 also have improved sensitivity and selectivity [9]. All these studies focused on the modification of SnO2, which enhances the miscellaneous complexity of the SnO2 system and in turn limits the controllability and reliability of SnO2 system. Such related studies seem to ignore the exploration of gas response on the basic attributes of SnO2, and the gas response to the different crystal planes of SnO2 is not clear yet. Moreover, although density functional theory (DFT) simulation indicates that the surface atomic configuration may affect the surface adsorption property [10], which may lead to different catalytic [11] and gas-sensing properties [12]. By now, most of these studies concentrated on the low-energy (110) plane of SnO2 [13,14]. However, there is still lack of the investigations on the gas response of other crystal planes, which limited by the hard preparation of SnO2 gas sensor with specific dominant crystal planes.

Here, SnO2 films with (101) and (110) preferred orientations were alternatively prepared by oxygen addition during magnetron sputtering, and micro-structures in the orientated films were further regulated by oxygen subtraction during annealing. The characteristics of (101) and (110)-oriented SnO2 films under oxygen regulation were studied, aiming at explaining their gas sensitivity and selectivity between H2 and CO. In combination with atomic configuration and surface morphology, the response mechanism of target gas in N2 or air was further clarified. Our controllable preparation of SnO2 sensors with the specific dominant crystal planes would be of great significance for the development of sensors with high sensor sensitivity, selectivity and stability.

2. Experimental

2.1. Sample preparation

SnO2 films were deposited from powder targets in a radio frequency (RF, AE600X) magnetron sputtering system. SnO2 powders with the purity of 99.99% were spread onto the copper backing plate and gently tamped with a clean stainless steel disc to form a uniform and smooth surface, as described in previous work [15]. The glass substrate was ultrasonically washed with acetone, alcohol, deionized water for 20 min, respectively, and then dried with nitrogen. The substrate was kept at room temperature without subsequent heating. The background vacuum was evacuated to a base pressure of less than 3.0 × 10-3 Pa and backfilled with Argon gas to a pressure of 1.2 × 10-1 Pa, and then the films were prepared under a target power of 200 W for 1 h, with a 100 mm substrate-target separation. In order to obtain different preferred orientation samples by oxygen regulation, films preparation was carried out under pure Ar atmosphere and Ar:O2 = 10:1 atmosphere, respectively.

The as-deposited SnO2 films were then annealed at a rate of 1 ℃/s under the vacuum of 2.5 × 10-3 Pa through rapid thermal processing equipment (RTP-500 V). The temperature was stayed at 110 ℃ for 5 min and 200 ℃ for 5 min to remove residual adsorbate on the surface, and was maintained 400 ℃ for 1 h.

2.2. Characterization methods

The phase and micro-structure of the films were analyzed by X-ray diffraction (XRD, X’Pert Pro, Netherlands, Cu-Kα, wavelength = 1.5406 Å) with a grazing angle of 0.3°. Surface morphology was observed using a field-emission Scanning Electron Microscope (FESEM, Sigma HD) at 5 kV and atomic force microscope (AFM, Bruker multimode 8) by ScanAsyst mode. Information of lattice structure and oxygen vacancy was characterized by Raman spectra (XploRA PLUS, He-Ne laser transmitter) at an excitation wavelength of 532 nm. The steady state photoluminescence (PL) spectra were recorded at room temperature on fluorescence spectrophotometer (Perkin-Elmer LS55) using a 320 nm excitation light. The deionized water wetting properties of the samples were studied by contact angle measuring instrument (JC2000D1), and electrical properties of the samples are obtained by Hall effect measurement system (Hall 8800, magnetic field strength 6000 G).

2.3. Gas-sensing measurement

The SnO2 gas sensitive films were directly assembled and connected to the electrochemical workstation through aluminum oxide sheet and Pt wire. All the measurements of gas sensitivity were carried out in an electric furnace device equipped with a set of gas-mixing counter. At 300-400 ℃, the "I-t curve" (I: current; t: time) is derived from the current and test time with 2000 ppm target gas (H2, CO) in the carrier gas (99.99% N2, air) at a constant voltage of 5 V through an electrochemical workstation (CHI660B, Chenhua Instruments Inc). The resistance was determined by the fixed voltage and measured current according to Ohm's law. Response sensitivity of the gas sensors in the test is defined [16]:

R= Ra/Rg (1)

where Ra is the resistance of the sensing material in the carrier gas and Rg is the resistance in the target gas atmosphere. The response/recovery time is the time required for the material resistance value to pass from the initial/stable value to the stable/initial value of 90% after the reductive gas is introduced/disconnected.

3. Results and discussion

3.1. Texture and microstructure

Fig. 1(a) shows the XRD pattern of the as-prepared and annealed SnO2 films by RF magnetron sputtering. All the diffraction peaks can be identified as the tetragonal rutile structure SnO2, and are consistent with the standard JCPDF card (IGDO 01-070-4177). No any peaks which belong to the impurity phase are observed. The oxygen-regulated SnO2 films have the different texture, (101) preferred orientation films deposited at the pure Ar atmosphere and (110) preferred orientation films deposited at the Ar atmosphere with additional oxygen. Besides, the shape of the weaker peak (200) is not significantly different between the two kinds of as-prepared samples as show in Fig. 1(a), and the comparison of gas sensitivity of the two preferred crystal planes would not be affected.

Fig. 1.   (a) XRD patterns of the as-prepared and annealed SnO2 films, (b) Raman spectra, (c) resistivity and (d) PL spectra of the SnO2 films with (101) and (110) preferred orientations.

Due to the fact that the lighter O atom has a larger mean free path than the Sn atom, oxygen is easily lost from the plasma and target itself, which in turn results in O-poor state for the (101) orientation-preferred SnO2 film [17]. The lost Oxygen during film formation can be compensated by adding oxygen in the atmosphere according to the reactive sputtering principle. The ratio of Sn:O in (110) oriented SnO2 film deposited under the oxygenic atmosphere is approximately stoichiometric. As this case, the more oxygen is aggregated on the surface of the target and deteriorates the conductance. While the effective potential difference on target becomes smaller, the sputtering ability is weakened duo to the introduction of oxygen. As a result, the depositing rate (133.2 nm/h) in the oxygenic atmosphere is much lower than that (324.5 nm/h) in the pure argon atmosphere.

According to the calculation of the surface energy of the low crystal face index of the tetragonal rutile structure SnO2 [18,19], the surface energy of the polar unsaturated (101) planes is higher than that of non-polar saturated (110) planes. Conversely, the (101) crystal planes is likely to be formed under thermodynamic conditions in the absence of oxygen. Meanwhile, the (110) crystal planes has a high surface energy which makes it difficult to form, as described in the previous report [20]. Furthermore, the films in oxygen-added atmosphere can grow along the (110) orientation with lower surface energy by self-relaxation and diffusion at the conditions of weak sputtering capability. However, the sputtering intensity is relatively strong and the relaxation ability is insufficient in the pure Ar atmosphere. Therefore, (101) films are formed under the oxygen-free atmosphere, and (110) films are more easily formed under the oxygenic atmosphere.

After annealing, both (101) and (110)-oriented films show the decline in the relative intensity of the diffraction peaks, but their preferred orientations remain unchanged in the XRD pattern. The films are de-oxidized at high temperature and high vacuum to generate a large number of oxygen vacancies, which increases the lattice distortion of the films and causes the more diffused diffraction peaks of SnO2 after annealing. Moreover, the diffraction peak of the (101)-oriented film annealed under oxygen regulation is much more diffusive than that of the (110) film, which indicates that the film of (101) orientation is more sensitive to oxygen.

As a typical n-type gas sensing material, SnO2 usually has many kinds of defects such as oxygen vacancies, interstitial Sn atoms in the lattice and so on. The number and density of defects caused by oxygen vacancies have a great influence on the gas sensitivity of SnO2 [21]. Raman spectrum which reflects internal structural information in a SnO2 film can be usually used to detect oxygen vacancies from phonon vibrations [22]. Fig. 1(b) shows the Raman spectra of different orientation-preferred films with the rutile structure. Raman peaks near the wavenumber of 443 cm-1 and 775 cm-1 are identified as phonon vibrations in A2g and B2g modes. A2g mode is attributed to the microstructure of SnO2 nanocrystallites and might constitute a kind of vibration mode according to the Matossi force constant model [6]. B2g mode is the typical first-order Raman vibration mode of SnO2. It is caused by the codirectional vibration of Sn and O atoms in a plane perpendicular to the c-axis [23]. The peak at 572 cm-1 originates from the plane oxygen vacancy and reflects relevant structural information of the samples [24]. Particularly, the asymmetrically widened diffuse Raman peak around 572 cm-1 suggests that this peak is composed of the binary structure information of the Raman frequency shift at 572 cm-1 and 630 cm-1, and the peak at 630 cm-1 on the right shoulder is referred to as A1g. The A1g with non-degenerate mode is generated by an opposite vibration of Sn—Sn and O—O in the plane vertical to c-axis. This vibration is affected by Sn and O content in the samples.

As shown in Fig. 1(b), when compared to other samples, the Raman peak around 572 cm-1 of the as-deposited (110) SnO2 film is significantly broadened, which indicates that the intensity of A1g vibration mode at 630 cm-1 is strong. Besides, it also demonstrates that the ratio of Sn:O in the (110)-oriented film is closer to the stoichiometry and the corresponding sample has fewer oxygen vacancies. In addition, it is known that the electrical properties of the SnO2 film are affected by the concentration of free electrons provided by oxygen vacancies. The as-deposited (110) film with higher resistivity presented in Fig. 1(c) also points to fewer oxygen vacancies in the crystal lattice. The number of oxygen vacancy in both (110) and (101)-oriented films increases after annealing, which is also reflected in the electrical characteristic, as shown in Fig. 1(c).

Fig. 1(d) shows the photoluminescence (PL) spectra with the emission peaks around 380-390 nm, 510-528 nm, and 710 nm of different SnO2 films. The strongest peak at $\widetilde{3}$90 nm (about 3.26 eV) is attributed to the electronic transition between the valence band and the conduction band. This value is lower than that of the intrinsic SnO2 at room temperature (3.6 eV) [25], suggesting that the films in the present work are higher metallic. The peaks observed at the wavelength of about 510 and 528 nm (2.43 eV and 2.34 eV) reflect defect energy level in the SnO2, which is related to the formation of oxygen vacancies [26,27]. Compared with the as-deposited (110) SnO2 film, the as-deposited (101) film has almost no shift in the peak of 390 nm and 710 nm, while the peak at 528 nm is significantly shifted to 510 nm. The red-shift of this peak results from the reduction of doping band gap. It confirms the (101)-oriented film has more oxygen vacancies than the (110)-oriented film, which agrees well with the results of Raman spectrum detection.

The peaks at 390 nm and 528 nm of the (101)-oriented film show a clear blue shift after annealing, because the oxygen in O-poor (101)-oriented film is further removed at high temperature and high vacuum. More oxygen vacancies lead to an increase in the ratio of Sn:O in the sample. When the relative content of Sn increases to form an overweight self-doping, according to Burstein-Moss effect, the optical band gap is broadened and the absorption limit is blue-shifted as shown in Fig. 1(d). Meanwhile, the (110)-oriented film has little change in the peak position at 390 nm, but the peak at 510 nm is red-shifted under the same annealing conditions. That is because doping band gap becomes narrower in the light of the appropriate increase in oxygen vacancies caused by annealing. It seems that the (101)-oriented film obtained under an oxygen-poor atmosphere is also easily de-oxidized to form oxygen vacancies, which was consistent with XRD analysis.

The peak at 710 nm is usually attributed to surface defects at grain boundaries [28]. Based on surface morphology, the (110)-oriented film has a larger density of grain boundaries from smaller grain size and greater roughness, so that its fluorescence emission intensity is stronger than that of (101)-oriented film. Furthermore, higher concentration of oxygen vacancy should be responsible for the increase in fluorescence intensity of SnO2 after annealing.

3.2. Morphology and wettability

The AFM topography of films with (110) and (101) preferred orientations demonstrates that the (110)-oriented film has greater roughness than (101)-oriented film as shown in Fig. 2(a)-(d), which can be explained by the growth mode of thin films [29]. At the initial stage of films formation, grain nucleation are island-shaped [29,30], and grains tend to grow on the original cluster rather than the gap between the cluster and the cluster at weak sputtering ability. Only as the sputtering energy becomes larger the grains can extend to fill the gap. Therefore, in the oxygenic atmosphere, the weak sputtering ability results in the preferred growth of (110) crystal plane with small grains and great roughness, which brings a large specific surface area. During the annealing process, the diffusion of atoms make the grain size increase, and the grain end of the film surface changes from sharp-angled to hemispherical shape based on the reduction of surface energy. The surface morphology differences between (110) and (101)-oriented SnO2 films were also reflected in SEM as presented in Fig. 2(e)-(h). The film with (110) preferred orientation has a more easily observed surface morphology in SEM compared to the film with (101) preferred orientation, suggesting an obvious contrast and a greater roughness on the (110) film.

Fig. 2.   AFM topographies of (a) as-prepared (101) SnO2 film, (b) as-prepared (110) SnO2 film, (c) annealed (101) SnO2 film, (d) annealed (110) SnO2 film, and images of SEM and wetting angles of (e) as-prepared (101) SnO2 film, (f) as-prepared (110) SnO2 film, (g) annealed (101) SnO2 film, (h) annealed (110) SnO2 film.

The wetting properties of the samples are shown in Fig. 2(e)-(h). The (110)-oriented film has a greater wetting angle 97.3° compared to 90.3° of (101)-oriented film from Fig. 2(e) and (f). It reflects that (110) plane of SnO2 is a non-polar saturated and thermodynamic stable plane while (101) plane is a polar unsaturated plane, based on Young-Dupro formula that the larger wetting angle means the lower surface free energy of film. From Fig. 2(g)-(h), the annealed films with (110) and (101) orientation have a larger wetting angle than the as-prepared films. At high temperature, the atoms with enhanced activity diffuse to release the internal stress of the film and form a hemispherical end of the grain, by which the annealed film has lower surface energy.

3.3. Gas-sensing properties

3.3.1. Gas response of (101) and (110) SnO2 films

In order to study the gas sensing properties of the different crystalline orientations, and the gas response of the (101) and (110)-oriented SnO2 films at 300-400 °C were tested as shown in Fig. 3(a). Similar to references [14,20,33], the SnO2 films with different preferred orientations have the optimum temperature of gas response, which is attributed to competition between adsorption characteristics and reaction activity. It can be seen that H2 optimum sensitivity (R = 9.6) at 350 ℃ of the as-prepared (110) SnO2 film with a shorter response time is twice more than that (R = 4.1) at 300 ℃ of the as-prepared (101) SnO2 film. The (110)-oriented SnO2 films have high gas sensitivity, while the (101)-oriented film is likely to present lower response temperature. This difference provides a guide to prepare and operate SnO2 sensor with highly gas-sensing selectivity.

Fig. 3.   H2 (a) and CO (b) response of SnO2 films with (101) and (110) preferred orientation at different temperatures and atomic configuration of SnO2 at (101) orientation (c) and (110) orientation (d). Here O3c is 3-Oxygen coordination, Sn3c is 3-tin coordination, OBridging is Bridging oxygen, OPlane is plane oxygen, Sn5c is 5-tin coordination, Sn6c is 6-tin coordination.

The atomic configuration and adsorption sites of the O-poor (101) and O-rich (110) orientated SnO2 are shown in Fig. 3(c) and (d). The O-poor (101)-oriented SnO2 consists of alternating layers of tin atoms and oxygen atoms, and the main gas adsorption sites are O3c and Sn3c [31]. The main gas adsorption sites of the O-rich (110)-oriented SnO2 composed of OBridging, OPlane, Sn5c and Sn6c, in which OBridging and Sn6c are the optimal adsorption sites of H2 on the surface of SnO2 according to the calculation of adsorption energy [[32], [33], [34]]. More different adsorption sites of the (110) crystal planes provide larger adsorption potential and a more rapid reaction to the target gas H2 when compared with the (101) crystal plane. Moreover, the as-prepared (110) SnO2 film with more oxygen content has a better response, suggesting that lattice oxygen plays an important role for H2 gas response. The exposed OBridging on the crystal surface of O-rich (110)-oriented SnO2 film provides more favorable adsorption sites of H2, while the exposed Sn3c on the crystal surface of O-poor (101)-oriented SnO2 shields adsorption action of O3c. It is worth mentioning that lots of grain boundaries and great roughness ought to make a non-negligible contribution to high H2 gas response of the as-prepared (110) SnO2 film. Similarly, the response selectivity of different crystal planes is also reflected in the detection of CO as shown in Fig. 3(b), which will be explained in detail later.

Fig. 4(a) shows that the gas-sensitive response of (110) and (101)-oriented SnO2 films dropped after annealing. Removal of oxygen declined gas sensitivity of the sample, which indicates the leading role of lattice oxygen in the SnO2 during the gas-sensitive reactions. In addition, the decrease of specific surface area resulted from the morphology change of the annealed SnO2 film also weakens the response of gas sensing. The exposed OBridging on the crystal surface of O-rich (110)-oriented SnO2 film is more enough than the exposed OPlane on the crystal surface of O-poor (101)-oriented SnO2 film. Much more oxygen in lost during annealing process, which makes hydrogen adsorption sites decrease the greater magnitude. Therefore, the reduction in the H2 response of (110) films is much more obvious than that of (101) films.

Fig. 4.   (a) Response comparison of (101) and (110) orientation-preferred SnO2 films before and after annealing and (b) response of H2 at (101) and (110) orientation-preferred SnO2 films in pure N2 and air.

3.3.2. Gas selectivity

In order to investigate the gas selectivity of SnO2 films with (101) and (110) preferred orientations, the reductive CO similar to H2 was selected as the comparative atmosphere for study. Fig. 3(a) and (b) shows the response of different oriented films to two kinds of target gases, and H2 selectivity of the samples can be mathematically defined:

S(H2/CO)= R(H2)/R(CO) (2)

where R(H2) and R(CO) represent the response sensitivity of the SnO2 film to H2 and CO, respectively. It can be seen that the (110)-oriented film has better selectivity (S = 6.5) compared to that of the (101)-oriented film (S = 1.7). According to the literature [[33], [34], [35]], the optimal adsorption site of H2 gas is O in crystal lattice and that of CO gas is Sn in crystal lattice, and H2 relative to CO is more easily adsorbed on the O-rich (110) plane. Moreover, when each of the H2 and CO molecules is adsorbed, the electron losses are 0.08e and 0.06e, respectively. H2 provides more equivalent electrons than CO, making the higher conductivity of SnO2 sensor. Therefore, the (110)-oriented film has high H2 response and better selectivity between H2 and CO. However, the O-poor (101)-oriented film has more oxygen vacancies, and the effect of Sn adsorption site becomes more pronounced, which causes an increase in the response to CO, making the sample less selective for H2 and CO. Due to oxygen lost during annealing process, there is a decline in the selectivity of both (101) and (110)-oriented SnO2 film, which further illustrates that the selectivity of different preferred orientations is also affected by oxygen regulation.

3.3.3. Gas-sensitive mechanism

Many classical models have been established to explain the gas sensitivity mechanism [[35], [36], [37]], such as surface-space charge layer model, contact grain boundary barrier model and molecular reaction model. All these models indicate the crucial role of adsorbed oxygen (Oads). The change in conductivity of the sample is due to adsorption or desorption of the adsorbed oxygen under the carrier gas. However, the foregoing discussion of the gas-sensitivity of SnO2 films with different preferred orientations reflects the important role of lattice oxygen rather than adsorbed oxygen. In order to clarify the gas-sensitive mechanism, H2 gas sensitivity of (101) and (110) orientation-preferred SnO2 films in pure N2 and air atmosphere was detected as shown in Fig. 4(b).

Based on the adsorbed oxygen mechanism, the gas-sensitive response value detected in the air should be higher than that in the pure N2 atmosphere. However, contrary to the above theoretical prediction, our results is that H2 sensitivity of SnO2 films in N2 as carrier gas is higher than that in air, indicating that the oxygen-free atmosphere is more favorable for the conductivity change of SnO2 films. Therefore, the gas-sensitive mechanism of SnO2 films with (101) and (110) preferred orientation should be dominated by lattice oxygen. The reductive target gas combines with the lattice oxygen to release electrons, which increases the conductivity of the SnO2 sensor. The conductivity returns to the initial value after desorbing target gas, which is a dynamic reversible process. The response of SnO2 films in air is lower than that in N2, that is to say, the adsorbed oxygen may have negative effects on the gas sensitivity of the film. The adsorption site of the target gas could be occupied and blocked by the adsorbed oxygen, and the adsorbed oxygen shields and weakens the adsorption of H2. In summary, the sensitivity and selectivity of SnO2 films with different preferred orientations together with the further gas-sensing tests under different atmosphere demonstrated that gas sensing mechanism was dominated by lattice oxygen rather than adsorbed oxygen.

4. Conclusions

1) The SnO2 films with (101) and (110) preferred orientation were obtained by sputtering under pure and oxygen-added Ar atmosphere, respectively. The O-poor (101)-oriented SnO2 film has a larger surface energy and a narrower doped band gap, while the O-rich (110)-oriented SnO2 film has a smaller grain and a higher roughness. Compared with the (101)-oriented SnO2 films, the (110)-oriented SnO2 film has higher H2 gas sensitivity and better selectivity between H2 and CO, 9.6 and 6.5 in optimum, respectively.

2) After annealing in vacuum, the surface energy of both (101)-oriented and (110)-oriented SnO2 films reduces with the increased grain size and rounded grain shape. As lattice oxygen is depleted, the (110)-oriented SnO2 film exhibits a larger reduction of the gas-sensitive response than the (101)-oriented SnO2 films.

3) Combined with a stronger gas response in pure N2 than air atmosphere, the comparison of gas-sensitive properties of different crystalline orientation based on oxygen regulation suggests that the gas-sensitive mechanism of SnO2 films is dominated by lattice oxygen. This provides a theoretical basis for further gas-sensitive studies of such films.

Acknowledgments

This work was financially supported by the National Natural Science Foundation of China (Nos. 51502126 and 51874169) and the Natural Science Foundation of Liaoning Province (No. 20180550802).


/