High Sensitivity Low-Temperature Hydrogen Sensors Based on SnO 2 / κ ( ε )-Ga 2 O 3 :Sn Heterostructure

: The structural and gas-sensitive properties of n - N SnO 2 / κ ( ε )-Ga 2 O 3 :Sn heterostructures were investigated in detail for the ﬁrst time. The κ ( ε )-Ga 2 O 3 :Sn and SnO 2 ﬁlms were grown by the halide vapor phase epitaxy and the high-frequency magnetron sputtering, respectively. The gas sensor response and speed of operation of the structures under H 2 exposure exceeded the corresponding values of single κ ( ε )-Ga 2 O 3 :Sn and SnO 2 ﬁlms within the temperature range of 25–175 ◦ C. Meanwhile, the investigated heterostructures demonstrated a low response to CO, NH 3 , and CH 4 gases and a high response to NO 2 , even at low concentrations of 100 ppm. The current responses of the SnO 2 / κ ( ε )-Ga 2 O 3 :Sn structure to 10 4 ppm of H 2 and 100 ppm of NO 2 were 30–47 arb. un. and 3.7 arb. un., correspondingly, at a temperature of 125 ◦ C. The increase in the sensitivity of heterostructures at low temperatures is explained by a rise of the electron concentration and a change of a microrelief of the SnO 2 ﬁlm surface when depositing on κ ( ε )-Ga 2 O 3 :Sn. The SnO 2 / κ ( ε )-Ga 2 O 3 :Sn heterostructures, having high gas sensitivity over a wide operating temperature range, can ﬁnd application in various ﬁelds.


Introduction
Sustainable development in terms of preserving the environment requires employment of a great number of sensors: biosensors, image sensors, motion sensors, and chemical sensors for indoor and outdoor as well as for industry-relevant gas surveillance and control.Wide bandgap metal oxide semiconductors tin dioxide (SnO 2 ) and gallium oxide (Ga 2 O 3 ) are of high interest for the development of gas sensors and transparent contacts, finding applications in a number of devices [1][2][3][4][5][6].Heterostructures based on metal oxide semiconductors allow the advantages of each component to be combined in a single structure [7].Thus, superior gas-sensitive characteristics can be achieved for heterostructures compared to single semiconductors.It is reasonable to combine semiconductors with high catalytic activity and concentration of electrons, involved in the physico-chemical processes at chemisorption of gas molecules on the semiconductor surface.SnO 2 is one of the most studied metal oxide semiconductors for gas sensor applications [1] primarily due to its high catalytic activity, which leads to a high gas sensitivity compared to other metal oxides.Chemisorption of gas molecules on the SnO 2 surface occurs with the involvement of free electrons.However, pure SnO 2 does not have a high electron concentration.Localization of electrons in this semiconductor can be achieved by forming heterostructures.Ga 2 O 3 with an electron affinity χ = 4.0 eV can be paired with SnO 2 , which is characterized by χ = 5.32 eV [8], to form such a heterostructure.In turn, Ga 2 O 3 needs to be doped to achieve the required concentration of free electrons.In this case, one can expect an increase in the sensitivity of such heterostructure to gases as compared to pure SnO 2 and Ga 2 O 3 films.
Gallium oxide has several polymorphs [9][10][11][12] namely α, β, γ, δ, and κ(ε).Metastable κ(ε)-Ga 2 O 3 polymorph is of particular interest for the development of electronic devices due to its fundamental properties [13] such as the thermal stability up to 700 • C; the high bandgap E g of 4.5-5.0eV; availability of the ferroelectric properties; the high symmetry of a crystal lattice.κ(ε)-Ga 2 O 3 is a novel material in terms of sensors, since its gas sensitivity was researched for the first time in 2022 [14].We have demonstrated that κ(ε)-Ga 2 O 3 :Sn films grown by the halide vapor phase epitaxy (HVPE) have a low resistance (i.e., high electron concentration), stable characteristics in the temperature range from 20 • C to 500 • C, and exhibit sensitivity to H 2 at room temperature (RT) [14].In addition, the κ(ε)-Ga 2 O 3 polymorph meets the conditions of heteroepitaxy on a commercially available (0001) Al 2 O 3 substrate better than monoclinic β-Ga 2 O 3 [15].Thus, doped κ(ε)-Ga 2 O 3 :Sn can be chosen to pair with SnO 2 to form a heterostructure.
The purpose of this work is to gain insight into the gas-sensitive properties of SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructures.

Materials and Methods
The following films were deposited on (0001) single crystal Al 2 O 3 substrates: κ(ε)-Ga 2 O 3 :Sn and SnO 2 thin films as well as SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure.The process of the κ(ε)-Ga 2 O 3 :Sn films growth was multistage.In the first stage, a 3-µm-thick semiinsulating (SI) GaN layer was deposited on the Al 2 O 3 substrate by gas phase deposition employing a homemade reactor.This layer served as a template for the κ(ε)-Ga 2 O 3 :Sn film growth.In the second stage, a 1-µm-thick κ(ε)-Ga 2 O 3 layer in situ doped by Sn was deposited on the SI-GaN layer by HVPE using a hot-wall homemade reactor.Gaseous gallium chloride and oxygen were utilized as precursors.The doping of the κ(ε)-Ga 2 O 3 films was carried out during the growth by adding tin.The HVPE growth temperature of the κ(ε)-Ga 2 O 3 :Sn film was 600 • C. The analysis of current-voltage (I-V) and capacitance-voltage (C-V) characteristics applied at this stage showed that the effective donor concentration N d of the films was 5.13 × 10 20 cm −3 .120-nm-thick pure SnO 2 thin films were deposited by means of magnetron sputtering of an Sn (5N) target in an oxygen-argon plasma on Al 2 O 3 and κ(ε)-Ga 2 O 3 :Sn.An Edwards A-500 (Edwards, USA) setup was employed.To prepare SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructures, SnO 2 thin films were deposited through a mask with square-shaped slots of 1 mm × 1 mm.The temperature of substrates during the deposition of the film was RT.The working pressure and power were kept at 7 × 10 −3 mbar and 70 W, respectively.The oxygen concentration in the O 2 +Ar mixture was 56.1 ± 0.5 vol.%.The as-deposited SnO 2 films were annealed ex situ at T = 600 • C; for 4 hours in air.The estimates showed that the N d value of these films was 5.26 × 10 17 cm −3 .Pt contacts were deposited on the κ(ε)-Ga 2 O 3 :Sn and SnO 2 films (see Figure 1) by means of the magnetron sputtering.Pt contacts were chosen on the basis of their high stability at high temperatures and under exposure to various gases, which are of natural surroundings-and industrial relevance.as-deposited SnO2 films were annealed ex situ at T = 600 °С for 4 hours in air.The estimates showed that the Nd value of these films was 5.26 × 10 17 cm −3 .Pt contacts were deposited on the κ(ε)-Ga2O3:Sn and SnO2 films (see Figure 1) by means of the magnetron sputtering.Pt contacts were chosen on the basis of their high stability at high temperatures and under exposure to various gases, which are of natural surroundings-and industrial relevance.X-ray diffraction (XRD) analysis of the samples was performed at DRON-6 diffractometer (Bourevestnik, Petersburg, Russia) equipped with a copper anode (CuKα1, λ = 1.5406Å).The XRD patterns were registered in θ-2θ scanning mode.The phase composition of the samples was identified by the position of the reflection peaks.XRD θ-2θ curves were processed using the Scherrer method [20] to determine the characteristic size of the block in the direction perpendicular to the plane of epitaxial growth.
The chemical composition of the samples was studied by X-ray photoelectron spectroscopy (XPS).The XPS measurements were carried out using a hemispherical analyzer included in the ESCALAB 250Xi (Thermo Fisher Scientific, Waltham, MA, USA) laboratory spectrometer.The measurements were carried out using a monochromatized AlKa radiation (hv = 1486.6eV).Survey and core (O1s, Sn3d, Ga3d) photoemission (PE) spectra were recorded at the analyzer transmission energy of 100 and 50 eV, respectively.The film's surface was irradiated with argon ions at an average energy of 3 eV for 60 s before XPS measurements to remove adsorbed atoms and molecules of contaminants.The analysis of the core spectra was processed employing the Avantage Data System software.
Measurement of transmission spectra was carried out using an Ocean Optics (Ocean Insight, Orlando, FL, USA) spectrometric system to determine the Eg of SnO2 in the wavelength range of λ = 300-600 nm.The transmission spectrum of κ(ε)-Ga2O3:Sn films was measured using a UV-VIS two-beam SPECORD (Analytik Jena, Jena, Germany) spectrophotometer in the range of λ = 230-360 nm.
A high-resolution field emission scanning electron microscope (FESEM) Apreo 2S (Thermo Fisher Scientific, USA) operating at an accelerating voltage of 5 kV was employed to study the microrelief of the film surfaces with a high resolution.
Gas sensing measurements of the samples were performed in a dedicated sealed chamber with a volume of 100 cm 3 , equipped with a micro-probe Nextron MPS-CHH station (Nextron, Korea).A ceramic-type heater, installed in the sealed chamber, was used to heat the samples.The accuracy of temperature T control was ±0.1 °C.The experiments were carried out under dark conditions.Streams of pure dry air or gas mixture of pure dry air + H2 were pumped through the chamber to measure the gas sensing characteristics of the samples.The H2 concentration in the mixture was controlled by a gas mixing and delivery system Microgas F-06 (Intera, Moscow, Russia).A special generator (Khimelektonika SPE, Moscow, Russia) was used to produce pure dry air.The total flow rate of the gas mixtures through the chamber was 1000 sccm.The relative error of the gas mixture flow rate did not exceed 1.5 %.A Keithley 2636 A (Keithley, Solon, OH, USA) source meter was utilized to measure the time dependences of the current I and the I-V characteristics of the samples.An E4980A RLC-meter (Agilent, Santa Clara, CA, USA) was applied to measure the С-V dependences.Additionally, gas sensing measurements of the samples were carried out under exposure to NH3, CH4, CO, NO2, and O2.A mixture X-ray diffraction (XRD) analysis of the samples was performed at DRON-6 diffractometer (Bourevestnik, Petersburg, Russia) equipped with a copper anode (CuK α1 , λ = 1.5406Å).The XRD patterns were registered in θ-2θ scanning mode.The phase composition of the samples was identified by the position of the reflection peaks.XRD θ-2θ curves were processed using the Scherrer method [20] to determine the characteristic size of the block in the direction perpendicular to the plane of epitaxial growth.
The chemical composition of the samples was studied by X-ray photoelectron spectroscopy (XPS).The XPS measurements were carried out using a hemispherical analyzer included in the ESCALAB 250Xi (Thermo Fisher Scientific, Waltham, MA, USA) laboratory spectrometer.The measurements were carried out using a monochromatized AlK a radiation (hv = 1486.6eV).Survey and core (O1s, Sn3d, Ga3d) photoemission (PE) spectra were recorded at the analyzer transmission energy of 100 and 50 eV, respectively.The film's surface was irradiated with argon ions at an average energy of 3 eV for 60 s before XPS measurements to remove adsorbed atoms and molecules of contaminants.The analysis of the core spectra was processed employing the Avantage Data System software.
Measurement of transmission spectra was carried out using an Ocean Optics (Ocean Insight, Orlando, FL, USA) spectrometric system to determine the E g of SnO 2 in the wavelength range of λ = 300-600 nm.The transmission spectrum of κ(ε)-Ga 2 O 3 :Sn films was measured using a UV-VIS two-beam SPECORD (Analytik Jena, Jena, Germany) spectrophotometer in the range of λ = 230-360 nm.
A high-resolution field emission scanning electron microscope (FESEM) Apreo 2S (Thermo Fisher Scientific, USA) operating at an accelerating voltage of 5 kV was employed to study the microrelief of the film surfaces with a high resolution.
Gas sensing measurements of the samples were performed in a dedicated sealed chamber with a volume of 100 cm 3 , equipped with a micro-probe Nextron MPS-CHH station (Nextron, Busan, Republic of Korea).A ceramic-type heater, installed in the sealed chamber, was used to heat the samples.The accuracy of temperature T control was ±0.1 • C. The experiments were carried out under dark conditions.Streams of pure dry air or gas mixture of pure dry air + H 2 were pumped through the chamber to measure the gas sensing characteristics of the samples.The H 2 concentration in the mixture was controlled by a gas mixing and delivery system Microgas F-06 (Intera, Moscow, Russia).A special generator (Khimelektonika SPE, Moscow, Russia) was used to produce pure dry air.The total flow rate of the gas mixtures through the chamber was 1000 sccm.The relative error of the gas mixture flow rate did not exceed 1.5%.A Keithley 2636 A (Keithley, Solon, OH, USA) source meter was utilized to measure the time dependences of the current I and the I-V characteristics of the samples.An E4980A RLC-meter (Agilent, Santa Clara, CA, USA) was applied to measure the C-V dependences.Additionally, gas sensing measurements of the samples were carried out under exposure to NH 3 , CH 4 , CO, NO 2 , and O 2 .A mixture of N 2 + O 2 was used to study the sensitivity of samples to O 2 .To study the effect of relative humidity (RH) on the response of the samples to H 2 , the pure dry air in one of the channels was passed through a bubbler with distilled water.Then it entered the homogenizer, where it was mixed with the pure dry air and/or pure dry air + H 2 mixture streams from the other channels.Varying the ratio of flows through the channels, we set the desired level of RH in the measuring chamber.An HIH 4000 Honeywell capacitive sensor with an absolute error of ±3.5% was used to measure the RH.Just prior to these measurements, all the samples were subjected to heat treatment at T = 500 • C for 90 s in pure dry air to stabilize the contact properties and regenerate the surface.0002), (0004), and (0006) reflections of the GaN template (AMCSD # 99-101-0461).Peaks corresponding to SnO 2 could not be distinguished due to possible overlapping by neighboring reflections of other phases.Thus, the (101) reflection of SnO 2 (AMCSD no.99-100-8661) is close to the (0002) one of GaN, and the (111) reflection of SnO 2 is close to the (004) one of κ(ε)-Ga 2 O 3 .The auxiliary vertical red lines of equal intensity depicted in Figure 2a are the tabular values of the SnO 2 reflection positions.In addition, difficulties in SnO 2 peaks identification may be caused by the low film thickness and the developed microrelief of the surface.Finally, the possible low crystallinity of the SnO 2 phase may be the reason for the absence of sharp peaks on the XRD pattern.In this case, broad humps of a low intensity may be present.

Structural Properties
κ(ε)-Ga 2 O 3 :Sn and SnO 2 films are characterized by direct optical transitions according to the analysis of transmission spectra (see Figure 2b), where α is the absorption coefficient.E g values were graphically calculated and proved to be equal to 4.61 ± 0.01 eV and 3.76 ± 0.01 eV for the κ(ε)-Ga 2 O 3 :Sn and SnO 2 films, respectively.
According to XPS analysis, the composition of the SnO 2 film includes Sn and O elements only.However, carbon (C) as a common contaminant was also observed in the subsurface layer a few nanometers thick.C atoms completely disappear after argon-etching for 60 s.Ga, Sn, O, and C lines were observed in the survey PE spectra of κ(ε)-Ga 2 O 3 :Sn film.The Sn concentration in this film appeared to be about 3 at.%, which indicates a high level of doping.Thus, the chemical analysis has shown that there are no third-party impurities in the composition of κ(ε)-Ga 2 O 3 :Sn and SnO 2 films, which confirms the high purity of the deposited films.
The analysis of the chemical state of Sn based on the Sn3d 5/2 PE line revealed the energy position of the main maximum of Sn at 486.5 and 486.3 eV (Figure 2c).The obtained values are in good agreement with the literature data [21,22] and correspond to the higher oxidation state of Sn-SnO 2 oxide.A lower value of the SnO 2 energy position for the κ(ε)-Ga 2 O 3 :Sn film indicates the effect of Ga 2 O 3 on the charge state of SnO 2 .Previously, we have observed a similar effect of the Sn3d 5/2 PE line shift to low binding energies of the SnO 2 film doped with rare-earth elements and platinum group metals [21,22].Analysis of the Ga chemical state in the κ(ε)-Ga 2 O 3 :Sn film based on the Ga3d PE line showed that Ga corresponds to the higher Ga 2 O 3 oxide [23] (Figure 2d).
FESEM images of the SnO 2 films surface deposited on Al 2 O 3 substrates and κ(ε)-Ga 2 O 3 :Sn film are displayed in Figure 2e,f, respectively.The microrelief of the SnO 2 film on Al 2 O 3 (Figure 2e) contains small spherical grains with a diameter of ~35 nm and large agglomerates with a characteristic size of ~300 nm.Whereas the microrelief of the SnO 2 film deposited on a κ(ε)-Ga 2 O 3 :Sn one (see Figure 2f) is represented by small grains with a diameter of ~35 nm only.The formation of large agglomerates for these structures was not observed.κ(ε)-Ga2O3:Sn and SnO2 films are characterized by direct optical transitions according to the analysis of transmission spectra (see Figure 2b), where α is the absorption coefficient.Eg values were graphically calculated and proved to be equal to 4.61 ± 0.01 eV and 3.76 ± 0.01 eV for the κ(ε)-Ga2O3:Sn and SnO2 films, respectively.
According to XPS analysis, the composition of the SnO2 film includes Sn and O elements only.However, carbon (C) as a common contaminant was also observed in the subsurface layer a few nanometers thick.C atoms completely disappear after argon-etching for 60 s.Ga, Sn, O, and C lines were observed in the survey PE spectra of κ(ε)-Ga2O3:Sn film.The Sn concentration in this film appeared to be about 3 at.%, which indicates a high level of doping.Thus, the chemical analysis has shown that there are no third-party impurities in the composition of κ(ε)-Ga2O3:Sn and SnO2 films, which confirms the high purity of the deposited films.
The analysis of the chemical state of Sn based on the Sn3d5/2 PE line revealed the energy position of the main maximum of Sn at 486.5 and 486.3 eV (Figure 2c).The ob-   Exposure to H2 leads to a reversible increase in the I through heterostructures at T = 25-200 °C. Figure 3b shows the change in the I-V characteristics of the SnO2/κ(ε)-Ga2O3:Sn heterostructure when exposed to 10 4 ppm of H2 at T = 150 °C.The  To assess the effect of H 2 on the I through the SnO 2 /κ(ε)-Ga 2 O 3 :Sn structures, the current response S I was calculated based on the experimental I-V characteristics by the following ratio: where I H is the current of the charge carrier through the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure in the gas mixture of pure dry air + H 2 ; I air is the current of the charge carrier through the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure in pure dry air.The S I values calculated on the basis of the experimental I-V characteristics and time dependences of currents at a fixed U (see Figure 4a) coincide.The S I value depends on the magnitude and direction of the applied voltage (see Figure 3c).[14,25].High sensitivity to H 2 at moderate temperatures (T = 300 • C) is characteristic of the SnO 2 thin films.Low S I for the κ(ε)-Ga 2 O 3 :Sn films are caused by a significant influence of the bulk conductivity G b , which does not depend on the charge state of the surface.The dependence of the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructures response to H 2 on temperature is characterized by a maximum at T = 125 • C.These samples demonstrate the highest S I in the range of T = 75-125 • C.
The experimental results displayed in Figure 4c prove that the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructures are characterized by the high speed of operation compared to the SnO 2 thin films when exposed to H 2 .The response t res and recovery t rec times were calculated to assess the speed of operation by the method described in ref. [9].The calculated t res and t rec values can only be used to compare the speed of sensors operation at similar experimental conditions.t res and t rec decrease exponentially with T. t rec and t res + t rec of the SnO 2 /κ(ε)-Ga 2 O 3 :Sn structures are significantly lower than those of SnO 2 thin films at T = 25-200 • C. SnO 2 films are characterized by low t res .The speed of operation for the κ(ε)-Ga 2 O 3 :Sn films was not evaluated due to their low responses at T > 50 • C.These samples are of interest for developing room temperature H 2 sensors.The t res and t rec of these films under exposure to 10 4 ppm of H 2 at T = 25 • C are 349.2s and 379.6 s, respectively.Obviously, SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructures are the most interesting for highly sensitive H 2 sensors with high speed of operation and low operating temperatures.Therefore, our further attention will be focused on these structures.
The dependence of the SnO 2 /κ(ε)-Ga 2 O 3 :Sn structure response on the H 2 concentration n H2 is linear (Figure 4d,e) in the n H2 range of 100-30000 ppm.The I air and I H of the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure decrease by 30% and 28%, respectively (see Figure 4f), during a cyclic exposure to H 2 (five cycles).At the same time, the current response decreased by only 17%.The observed decrease in response during cyclic exposure to H 2 is caused by the manifestation of chemisorbed hydrogen atoms with high binding energy.The temperature of 125 • C is not sufficient for the complete desorption of these hydrogen atoms from the semiconductor surface.Short-term heating of the structure at high temperatures can be used to regenerate the surface of semiconductors and for full desorption of H atoms [26].The results of the long-term tests of the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructures at T = 125 • C and when exposed to 10 4 ppm of H 2 demonstrated opposite changes of S I .The samples after the experiments were stored in sealed packages.The long-term tests lasted 8 weeks with an interval between the experiments of 7-8 days.Just prior to each measurement the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructures were subjected to heat at T = 500 • C for 90 s.There were increases in response from ~30 arb.un. to 47 arb.un.during the long-term tests.Response increases mostly due to a decrease in I air .The most significant changes in response were in the first 4 weeks of testing.
The responses of the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure to NO 2 , CH 4 , NH 3 , CO, and O 2 gases at T = 125 • C was measured to evaluate its selectivity (Figure 4g).Noteworthy, is that I through heterostructure increases reversibly when exposed to 10 4 ppm of CH 4 , NH 3 h and CO.The response to these gases has been calculated by equation (1).The I H was replaced by I g , where I g is the charge carrier current through the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure in the gas mixture of pure dry air + reducing gas (CH 4 , NH 3 , or CO).The responses to CH 4 , NH 3 , and CO are insignificant compared to the S I to H 2 , which equates to 29.92-46.98arb.un. at T = 125 • C and n H2 = 10 4 ppm.
It was found, that the I value of the SnO 2 /κ(ε)-Ga 2 O 3 :Sn reversibly decreases when exposed to NO 2 and O 2 .The responses to NO 2 (S NO2 ) and O 2 (S O2 ) have been calculated by the following equations, correspondently: where  4h).The most significant decrease in S I occurs when RH increases from 0 to 34 %.In the range of RH = 34-90.0%,the response varies slightly.
Furthermore, the effect of 10 4 ppm of H 2 on the C-V characteristics of the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructures at T = 125 • C and signal frequencies f = 1 kHz, 10 kHz, and 1 MHz have been studied.The results are illustrated in Figure 4i,j.Evidently, exposure to H 2 leads to a reversible increase in the electrical capacity of the structures.The capacitive response S C has been calculated by the following equation: where C H is the electrical capacitance of SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure in the gas mixture of pure dry air + H 2 ; and C air is the electrical capacitance of structures in pure dry air.Visibly, the S C (see Figure 4i) is significantly lower than the S I (Figure 3c).At f = 10 kHz and 1 MHz the capacitive response varies weakly.The highest S C value is observed in the range of U = 0.95-5.00V at f = 1 kHz and has a maximum at U = 2.9 V.

The Mechanism of the Sensory Effect
Initially, the resistance of the Pt/κ(ε)-Ga 2 O 3 :Sn interface is low and the Pt/SnO 2 contact is ohmic.The change in the potential barrier at Pt/κ(ε)-Ga 2 O 3 :Sn and Pt/SnO 2 interfaces upon exposure to gases can be neglected.The observed high responses of the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure at T = 25-175 • C; are due to the formation of the n-N isotypic heterojunction, where SnO 2 is the base.
Diffusion of H atoms up to the SnO 2 /κ(ε)-Ga 2 O 3 :Sn interface at T = 25-175 • C; is unlikely.Changes of I and C upon exposure to H 2 occur mainly due to the chemisorption of gas molecules on semiconductor's surface.In the air atmosphere within a temperature range of T = 25-175 • C, the oxygen chemisorbs mainly in a molecular form on metal oxide semiconductors surface and captures electrons from their conduction band [25,27].The reaction of reversible chemisorption of oxygen molecules can be represented as follows: where S a is a free adsorption center; e is the electron charge; O 2 − (c) is the chemisorbed oxygen ion.As a result of reaction ( 4), an electron-depleted region is formed in the nearsurface part of the semiconductor.A negative charge on the surface causes the upward of energy bands bending at eV s , where V s is the surface potential and eV s ~Ni 2 , where N i is the surface density of chemisorbed oxygen ions.In our case, the Debye length L D for SnO 2 exceeds the grain size and the effect of grain boundaries on the transport of charge carriers in a semiconductor can be ignored.Oxygen chemisorption weakly changes the electrical conductivity of the κ(ε)-Ga 2 O 3 :Sn films due to the low contribution of the surface conductivity G s to the total conductivity of G t .Thus, changes in the current during the chemisorption of gases are mainly due to changes in the concentration of charge carriers in the SnO 2 .
G t = G b + G s and the relationship between G t and eV s in our case is described by the following equation [28]: where D is the SnO 2 film thickness; k is the Boltzmann constant.An increase in the eV s due to the oxygen molecule's chemisorption on the SnO 2 surface leads to a drop of G t .
The increase in G t when exposed to H 2 is caused by the interaction of H 2 molecules with previously chemisorbed O 2 − (c) on the SnO 2 surface.This interaction can be represented as follows: As a result of reaction (7), a neutral H 2 O molecule is formed and desorbed, an electron returns to the conduction band of SnO 2 , the eV s decreases and the finally G t increases.When exposed to NO 2 the following reactions take place [29]: NO 2 molecules chemisorb onto free adsorption centers and capture electrons from the SnO 2 conduction band.Meanwhile, eV s is proportional to (N i + N NO2 ) 2 in the mixtures of air + NO 2 , where N NO2 is the surface density of chemisorbed NO 2 − ions [25,27].An additional negative charge on the surface of the SnO 2 film leads to a greater decrease in G t .Further, NO 2 − ions are dissociated to form chemisorbed O − (c) ions and gaseous NO molecules.In the low temperature region, atomic O − (c) ions associated to O 2 − (c) form and a free electron e − , which returns to the conduction band of the semiconductor.
The κ(ε)-Ga 2 O 3 :Sn film is a source of electrons that are involved in reactions ( 5) and ( 8) with the gas molecules on the SnO 2 surface.This causes a high response of heterostructure at T = 75-175 • C. The base region (SnO 2 ) is filled with electrons from the κ(ε)-Ga 2 O 3 :Sn film with T rising, G b of SnO 2 increases and the response decreases.SnO 2 films deposited on κ(ε)-Ga 2 O 3 :Sn are characterized by the absence of large agglomerates (see Figure 2e,f).This leads to an increase in the specific surface area of SnO 2 and the surface density of adsorption centers for gas molecules.

Conclusions
The structural and gas-sensitive properties of the n-N SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructures were investigated for the first time.The κ(ε)-Ga 2 O 3 :Sn and SnO 2 films were obtained by the halide vapor phase epitaxy and the high-frequency magnetron sputtering, respectively.The κ(ε)-Ga 2 O 3 :Sn crystalline film has a bandgap of 4.61 ± 0.01 eV.The SnO 2 nanocrystalline film has a bandgap of 3.76 ± 0.01 eV and is characterized by a developed microrelief of the surface, represented by grains with a size of ~35 nm.Exposure to H 2 leads to an increase in electrical current and capacitance of SnO 2 /κ(ε)-Ga 2 O 3 :Sn structures.The current response of heterostructures to H 2 significantly exceeds the capacitive one.Gas sensor response and speed of operation of the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure under H 2 exposure overperform those of the single κ(ε)-Ga 2 O 3 :Sn and SnO 2 films in the temperature range of 25-175 • C.This heterostructure demonstrates a low response to CO, NH 3 , and CH 4 and a high response to NO 2 even at low concentrations.The current responses of SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure to 10 4 ppm of H 2 and 100 ppm of NO 2 at 125 • C were 30-47 A.U. and 3.7 A.U., correspondingly.The sensory effect is realized mainly due to the chemisorption of gas molecules on the SnO 2 surface, which is the base region of the heterostructure.The κ(ε)-Ga 2 O 3 :Sn film is a source of electrons that are involved in reactions with gas molecules on the SnO 2 film surface.The SnO 2 film deposited on the κ(ε)-Ga 2 O 3 :Sn film is characterized by a more developed surface microstructure.This leads to an increase in the surface density of adsorption centers for gas molecules.The advantages of the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure for gas sensors are shown, the main one being high sensitivity at relatively low operating temperatures.Doubtfully, this structure has every chance of being the base of the sensor.

Figure
Figure2aillustrates the θ-2θ XRD pattern of the SnO 2 /Ga 2 O 3 heterostructure deposited on an Al 2 O 3 substrate via a GaN template.The peaks at 2θ = 41.8 • and 90.9 • are associated with the (0006) and (0 0 0 12) reflections of the Al 2 O 3 substrate (ICDD # 00-042-1468).A series of peaks at 2θ = 19.2• , 39.0 • , 60.0 • , 83.6 • , and 112.7 • correspond to the 002, 004, 006, 008, and 0 0 10 planes of the κ(ε)-Ga 2 O 3 phase.(The calculation was made on the basis of the Bragg equation for the case of CuK α1 anode (λ = 1.5406Å).The peaks at 2θ = 34.7 • , 73.0 • , and 126.1 • are due to the (0002), (0004), and (0006) reflections of the GaN template (AMCSD # 99-101-0461).Peaks corresponding to SnO 2 could not be distinguished due to possible overlapping by neighboring reflections of other phases.Thus, the (101) reflection of SnO 2 (AMCSD no.99-100-8661) is close to the (0002) one of GaN, and the (111) reflection of SnO 2 is close to the (004) one of κ(ε)-Ga 2 O 3 .The auxiliary vertical red lines of equal intensity depicted in Figure2aare the tabular values of the SnO 2 reflection positions.In addition, difficulties in SnO 2 peaks identification may be caused by the low film thickness and the developed microrelief of the surface.Finally, the possible low crystallinity of the SnO 2 phase may be the reason for the absence of sharp peaks on the XRD pattern.In this case, broad humps of a low intensity may be present.κ(ε)-Ga 2 O 3 :Sn and SnO 2 films are characterized by direct optical transitions according to the analysis of transmission spectra (see Figure2b), where α is the absorption coefficient.E g values were graphically calculated and proved to be equal to 4.61 ± 0.01 eV and 3.76 ± 0.01 eV for the κ(ε)-Ga 2 O 3 :Sn and SnO 2 films, respectively.According to XPS analysis, the composition of the SnO 2 film includes Sn and O elements only.However, carbon (C) as a common contaminant was also observed in the subsurface layer a few nanometers thick.C atoms completely disappear after argon-etching for 60 s.Ga, Sn, O, and C lines were observed in the survey PE spectra of κ(ε)-Ga 2 O 3 :Sn film.The Sn concentration in this film appeared to be about 3 at.%, which indicates a high level of doping.Thus, the chemical analysis has shown that there are no third-party impurities in the composition of κ(ε)-Ga 2 O 3 :Sn and SnO 2 films, which confirms the high purity of the deposited films.The analysis of the chemical state of Sn based on the Sn3d 5/2 PE line revealed the energy position of the main maximum of Sn at 486.5 and 486.3 eV (Figure2c).The obtained values are in good agreement with the literature data[21,22] and correspond to the higher oxidation state of Sn-SnO 2 oxide.A lower value of the SnO 2 energy position for the κ(ε)-Ga 2 O 3 :Sn film indicates the effect of Ga 2 O 3 on the charge state of SnO 2 .Previously, we have observed a similar effect of the Sn3d 5/2 PE line shift to low binding energies of the SnO 2 film doped with rare-earth elements and platinum group metals[21,22].Analysis of the Ga chemical state in the κ(ε)-Ga 2 O 3 :Sn film based on the Ga3d PE line showed that Ga corresponds to the higher Ga 2 O 3 oxide[23] (Figure2d).FESEM images of the SnO 2 films surface deposited on Al 2 O 3 substrates and κ(ε)-Ga 2 O 3 :Sn film are displayed in Figure2e,f, respectively.The microrelief of the SnO 2 film on Al 2 O 3 (Figure2e) contains small spherical grains with a diameter of ~35 nm and large agglomerates with a characteristic size of ~300 nm.Whereas the microrelief of the SnO 2 film deposited on a κ(ε)-Ga 2 O 3 :Sn one (see Figure2f) is represented by small grains with a diameter of ~35 nm only.The formation of large agglomerates for these structures was not observed.

3. 2 .
Gas-Sensitive Properties of the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure The I-V characteristics of the SnO 2 thin films equipped with Pt contacts are linear in the range of applied voltages U = −40-40 V at RT as well as at higher T. Contrary to this, the I-V characteristics of the κ(ε)-Ga 2 O 3 :Sn films equipped with Pt contacts are nonlinear.The dependence of ln(I) on U 1/4 is linear, indicating the presence of a Schottky barrier at the Pt/κ(ε)-Ga 2 O 3 :Sn interface [24].The I value through the Pt/κ(ε)-Ga 2 O 3 :Sn/Pt structures exceeds 0.1 A at U > 12 V which leads to the samples self-heating.The SnO 2 /κ(ε)-Ga 2 O 3 :Sn structure equipped with Pt contacts is a n-N isotype heterojunction and the Schottky barriers are connected in series.The I-V characteristics of such heterostructures are nonlinear and asymmetric as can be seen in Figure 3a.The I(U = 4 V)/I(U = −4 V) ratio reaches the value of ~2 × 10 3 at T = 25 • C, then drops by half as T increases to 150 • C. The increase in reverse current with T rising is significantly higher than the increase in forward current.The forward-bias region of the I-V characteristics is approximated by the following function: I f = A 1 × exp(B 1 U), where I f is a forward current; and A 1 and B 1 are the constants: A 1 = (3.0 ± 0.4) × 10 −6 A and B 1 = 0.88 ± 0.02 V −1 at T = 25 • C. The reverse-bias region of the I-V characteristics can be approximated by a similar function of I r = A 2 × exp(B 2 |U|), where I r is a reverse current; A 2 and B 2 are the constants: A 2 = (2.0 ± 0.4) × 10 −9 A and B 1 = 0.89 ± 0.04 V −1 at T = 25 • C. The forwardbias mode of the structure corresponds to the application of a positive potential to the SnO 2 /Pt interface.The SnO2/κ(ε)-Ga2O3:Sn structure equipped with Pt contacts is a n-N isotype het-erojunction and the Schottky barriers are connected in series.The I-V characteristics of such heterostructures are nonlinear and asymmetric as can be seen in Figure3a.The I(U = 4 V)/I(U = −4 V) ratio reaches the value of ~2 ×10 3 at T = 25 °C, then drops by half as T increases to 150 °C.The increase in reverse current with T rising is significantly higher than the increase in forward current.The forward-bias region of the I-V characteristics is approximated by the following function: If = A1×exp(B1U), where If is a forward current; and A1 and B1 are the constants: A1 = (3.0 ± 0.4) × 10 −6 A and B1 = 0.88 ± 0.02 V −1 at T = 25 °C.The reverse-bias region of the I-V characteristics can be approximated by a similar function of Ir = A2 × exp(B2|U|), where Ir is a reverse current; A2 and B2 are the constants: A2 = (2.0 ± 0.4) × 10 −9 A and B1 = 0.89 ± 0.04 V −1 at T = 25 °C.The forward-bias mode of the structure corresponds to the application of a positive potential to the SnO2/Pt interface.

Figure 3 .
Figure 3. I-V characteristics of the SnO2/κ(ε)-Ga2O3:Sn heterostructure at T = 25 °C in pure dry air in semi-logarithmic coordinates (a), at T = 150 °C under exposure to 10 4 ppm of H2 (b).The insertions show these I-V characteristics in linear coordinates.Dependence of the response to 10 4 ppm H2 on the applied voltage at different temperatures (c), the insertion shows the response to 10 4 ppm H2 on the applied voltage at T = 25 °C and 200 °C.

Figure 3 .
Figure 3. I-V characteristics of the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure at T = 25 • C in pure dry air in semi-logarithmic coordinates (a), at T = 150 • C under exposure to 10 4 ppm of H 2 (b).The insertions show these I-V characteristics in linear coordinates.Dependence of the response to 10 4 ppm H 2 on the applied voltage at different temperatures (c), the insertion shows the response to 10 4 ppm H 2 on the applied voltage at T = 25 • C and 200 • C. Exposure to H 2 leads to a reversible increase in the I through heterostructures at T = 25-200 • C. Figure 3b shows the change in the I-V characteristics of the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure when exposed to 10 4 ppm of H 2 at T = 150 • C. The type of the functions, approximating the forward and reverse branches of the I-V characteristics, does not change with increasing T to 200 • C and under exposure to 10 4 ppm of H 2 .The A 1 and A 2 increase, but B 1 and B 2 values decrease with T. The A 1 and A 2 values increase, whereas B 1 and B 2 practically do not change when exposed to H 2 .Table 1 shows the A 1 , A 2 , B 1, and B 2 values at T = 150 • C and under exposure to 10 4 ppm of H 2 .

Figure 4 .
Figure 4. Gas-sensitive properties of n-N SnO2/κ(ε)-Ga2O3:Sn heterostructure and other samples: (a) time dependences of current upon exposure to 10 4 ppm of H2; (b) temperature dependences of responses to 10 4 ppm of Н2; (c) temperature dependences of response and recovery times upon exposure to 10 4 ppm of Н2 for different samples; (d) time dependence of current upon exposure to different H2 concentration; (e) dependence of response on H2 concentration; (f) time dependence of current upon cyclic exposure to 10 4 ppm of H2; (g) responses to fixed concentrations of NO2, CH4, NH3, CO, H2, and O2; (h) effect of the relative humidity on responses to 10 4 ppm of H2; (i) dependences of the capacitive response on applied voltage upon exposure to 10 4 ppm of H2 at T = 125 °C and different frequencies; (j) effect of 10 4 ppm of H2 on C-V characteristics at T = 125 °C and different frequencies; dependences in (d-h) were measured at T = 125 °C and U = 0.5 V.
NO 2 is the charge carrier current through the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure in the gas mixture of pure dry air + NO 2 ; I N is the charge carrier current through the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure in the nitrogen atmosphere; I O 2 is the charge carrier current through the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure in the gas mixture of N 2 + O 2 .S I ratio when exposed to 100 ppm of H 2 at T = 125 • C happened to be 26.7 times lower than those for 100 ppm of NO 2 (Figure4g).The SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure also demonstrated relatively high response to O 2 .The response to O 2 appears to be higher than to CH 4 , NH 3 , or CO at same concentration values.Hence, we have shown that SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure is also attractive for developing highly sensitive NO 2 and O 2 sensors operating at low temperatures.An increase in RH leads to a drop in the response of the SnO 2 /κ(ε)-Ga 2 O 3 :Sn heterostructure to H 2 (Figure

Table 2 .
Gas-sensitive characteristics of structures based on Ga 2 O 3 polymorphs.

Table 3 .
Gas-sensitive characteristics of heterostructures based on different metal oxide semiconductors.