Drop-Dry Deposition of SnO2 Using a Complexing Agent and Fabrication of Heterojunctions with Co3O4

The drop-dry deposition (DDD) is a simple chemical technique of thin film deposition, which can be applied to metal oxides. The deposition solution is an aqueous solution including a metal salt and an alkali. However, some metal ions react spontaneously with water and precipitate. This work is the first attempt to use complexing agents in DDD to suppress the precipitation. SnO2 thin films are fabricated using DDD with Na2S2O3 as a complexing agent and via annealing in air. The results of the Auger electron spectroscopy measurement show that the O/Sn composition ratio of the annealed films approached two, indicating that the annealed films are SnO2. The photoelectrochemical measurement results show that the annealed films are n-type. Co3O4/SnO2 heterojunction is fabricated using p-type Co3O4 films which are also deposited via DDD. The heterojunction has rectification and photovoltaic properties. Thus, for the first time, a metal oxide thin film was successfully prepared via DDD using a complexing agent, and oxide thin film solar cells are successfully prepared using only DDD.


Introduction
Solar energy is an inexhaustible renewable energy source. Photovoltaic technology that uses solar energy is an environment-friendly and promising technology. In addition to crystalline silicon solar cells, thin-film materials such as CdTe [1], GaAs [2], and Cu(In,Ga)Se 2 (CIGS) [3] have been widely studied. In this study, we focus on metal oxide materials.
Metal oxides are an important branch of inorganic materials. Their unique physical and chemical properties lead to a wide range of applications, including electronic devices, batteries, catalysts, and sensors. Many metal oxide materials consist of non-polluting and abundant elements, meeting the requirements of inexpensive production. From an optical point of view, there are many metal oxide materials suitable for photovoltaic applications. All-oxide photovoltaics cells are promising to realize extremely cheap photovoltaic systems.
The methods of preparing SnO 2 thin films include chemical bath deposition (CBD) [22,23], magnetron sputtering [11,13], spray pyrolysis [12], sol-gel technique [24][25][26], ECD [18,27,28], etc. In this work, we prepare tin hydroxide thin films using the drop-dry deposition (DDD) method [29][30][31], and transform it into SnO 2 films via annealing. DDD is a method of preparing thin films by dropping the deposition solution on a substrate and then drying it. The apparatus is simple, requiring only a pipette and a heating plate. Physical and chemical vapor depositions mostly require vacuum or leak-tight reactors, and the equipment is expensive. Compared with these methods, the equipment required for DDD is simple and easy to operate. Compared with other chemical liquid phase deposition methods, DDD has some advantages, as described below.
So far, Mg(OH) 2 , Co 3 O 4 and NiO have been successfully fabricated using DDD [29][30][31]. For example, we prepared a 0.5 µm thick Co(OH) 2 thin films via DDD with a total deposition time of about 30 min [30]. The deposition time in DDD is much shorter than for CBD; the deposition time is usually several hours in CBD [32]. Spray pyrolysis is another popular chemical technique, but it requires an atomizer and high temperatures (300-450 • C) [33]. For DDD, the equipment is simpler (just a pipette), and thin films can be deposited at lower temperatures (60 • C). The experimental equipment for DDD is also simpler than those of the spin-coating: spin-coating requires a spin coater. In general, the advantages of DDD are simple equipment, easy operation, low cost, low deposition temperature, and short deposition time.
In the previous studies on DDD, the deposition solution used was an aqueous solution including a metal salt (e.g., Co(NO 3 ) 2 ) and an alkali (e.g., NaOH). The metal hydroxide is generated in the solution, and the solution is saturated or nearly-saturated without precipitation by controlling the concentration of the chemicals and the pH in the solution. However, the deposition solution of tin hydroxide cannot be prepared similarly. When a tin salt (e.g., SnSO 4 ) is dissolved in water, hydrolysis occurs and tin hydroxide precipitates even without adding alkali [34], making the solution unusable for deposition.
To prevent precipitation in the solution, a complexing agent is added to the DDD solution for the first time in this work. If Sn 2+ ion forms a complex, the hydrolysis reaction will be suppressed. There are many types of complexing agents, and chemicals such as EDTA and tartaric acid are commonly used in metal plating. Studies on complexes based on Na 2 S 2 O 3 have also been reported [35]. In this work, tartaric acid and Na 2 S 2 O 3 are used to prepare the deposition solutions.
In this work, we also fabricate Co 3 O 4 /SnO 2 heterojunction solar cells with Co 3 O 4 prepared via DDD. Co 3 O 4 is a p-type metal oxide semiconductor with two direct bandgaps of 1.5 eV and 2.0 eV, and is well suited for an absorber material in solar cells [36]. Co 3 O 4 has already been used in photovoltaic devices [37][38][39][40]. So far, Co 3 O 4 /SnO 2 heterojunctions have been mainly used in gas sensors [41][42][43][44] and catalysts [45][46][47], whereas we attempt to apply Co 3 O 4 /SnO 2 heterojunctions in solar cells for the first time. This paper is also the first report of solar cells fabrication using only DDD. As shown below, photovoltaic output is confirmed for the Co 3 O 4 /SnO 2 heterojunction, which demonstrates that DDD can be actually used for solar cell fabrication.

Experimental Design
The deposition process of DDD is shown in Figure 1. In the deposition solution, tin(II) sulfate (SnSO 4 , minimum 93% purity, Kanto Chemical Co., Inc., Tokyo, Japan) was used as the Sn 2+ source, and sodium hydroxide (NaOH, minimum 97% purity, Kanto Chemical Co., Inc., Tokyo, Japan) was added to adjust the solution pH. The SnSO 4 concentration was 10 or 20 mM. Tartaric acid (L(+)-Tartaric acid, minimum 99% purity, Kanto Chemical Co., Inc., Tokyo, Japan) or sodium thiosulfate (Na 2 S 2 O 3 , minimum 97% purity, Kanto Chemical Co., Inc., Tokyo, Japan) was added as a complexing agent. The specific deposition solution conditions with Na 2 S 2 O 3 addition are shown in Table 1. Fluorine-doped tin oxide (FTO)-coated glass (Furuuchi Chemical Co., Ltd., Tokyo, Japan), indium tin oxide (ITO)coated glass (Furuuchi Chemical Co., Ltd., Tokyo, Japan) and alkali-free glass (Corning Co., Corning, NY, USA) sheets were selected as the substrates, and the deposition area was 1.8 × 1.8 cm 2 . Briefly, 0.1 mL of the solution was dropped each time, the heating temperature was 60 • C, and the deposition cycles were 5 times. After depositing the thin Materials 2023, 16, 5273 3 of 13 film, the sample was placed in a tube furnace and annealed at 200 or 400 • C for 1 h in air to convert the as-deposited film into a SnO 2 film. For resistivity measurements, inter-digitpattern indium electrodes (0.1 × 0.1 cm 2 ) were fabricated via vacuum evaporation on the films deposited on the alkali-free glass substrate. (Corning Co., Corning, NY, USA) sheets were selected as the substrates, and the deposition area was 1.8 × 1.8 cm 2 . Briefly, 0.1 mL of the solution was dropped each time, the heating temperature was 60 °C, and the deposition cycles were 5 times. After depositing the thin film, the sample was placed in a tube furnace and annealed at 200 or 400 °C for 1 h in air to convert the as-deposited film into a SnO2 film. For resistivity measurements, inter-digit-pattern indium electrodes (0.1 × 0.1 cm 2 ) were fabricated via vacuum evaporation on the films deposited on the alkali-free glass substrate.   Figure 2 shows the schematic and photograph of the Co3O4/SnO2 heterojunction. The Co3O4/SnO2 heterojunctions were fabricated as follows: First, the SnO2 film with an area of 1.8 × 1.8 cm 2 was fabricated on the ITO substrate via DDD and annealing (at 200 °C, as described below). Then, a Co3O4 film was fabricated on it via DDD and annealing at 200 °C; the preparation of Co3O4 is the same as reported in ref. [30]. The deposition area of Co3O4 was limited to 0.8 × 0.8 cm 2 , and the deposition solution contained 20 mM of Cobalt(II) nitrate hexahydrate (Co(NO3)2·6H2O, minimum 97% purity, Kanto Chemical Co., Inc., Tokyo, Japan) and 10 mM of NaOH. In order to avoid the Co3O4 layer from becoming too thick, the deposition cycles were 3 times instead of 5 times. For the current densityvoltage (J-V) characterization, indium electrodes (0.1 × 0.1 cm 2 ) were fabricated on the heterojunction via vacuum evaporation.   Figure 2 shows the schematic and photograph of the Co 3 O 4 /SnO 2 heterojunction. The Co 3 O 4 /SnO 2 heterojunctions were fabricated as follows: First, the SnO 2 film with an area of 1.8 × 1.8 cm 2 was fabricated on the ITO substrate via DDD and annealing (at 200 • C, as described below). Then, a Co 3 O 4 film was fabricated on it via DDD and annealing at 200 • C; the preparation of Co 3 O 4 is the same as reported in ref. [30]. The deposition area of Co 3 O 4 was limited to 0.8 × 0.8 cm 2 , and the deposition solution contained 20 mM of Cobalt(II) nitrate hexahydrate (Co(NO 3 ) 2 ·6H 2 O, minimum 97% purity, Kanto Chemical Co., Inc., Tokyo, Japan) and 10 mM of NaOH. In order to avoid the Co 3 O 4 layer from becoming too thick, the deposition cycles were 3 times instead of 5 times. For the current density-voltage (J-V) characterization, indium electrodes (0.1 × 0.1 cm 2 ) were fabricated on the heterojunction via vacuum evaporation.
The optical properties, structure and morphology of the films were characterized through transmittance measurements (Jasco V-570 UV/VIS/NIR spectrometer, Tokyo, Japan), X-ray diffraction (XRD) measurements (SmartLab SE X-ray diffractometer, Tokyo, Japan), scanning electron microscope (SEM) images and Auger electron spectroscopy (AES) measurements (JEOL JAMP-9500F, Akishima, Japan). All equipment was calibrated before use. Photoelectrochemical (PEC) measurement was carried out in a three-electrode system with a Ag/AgCl reference electrode. A 100 mM sodium sulfate (Na 2 SO 4 , minimum 99% purity, Kanto Chemical Co., Inc., Tokyo, Japan) solution was used as the electrolyte. The samples were intermittently irradiated with 100 mW/cm 2 light (ABET Technologies 10500 Sun Simulator, Milford, CT, USA) at 5 s intervals and scanned for sample potentials in the range of −1 to 0 V and 0 to 1 V at a scan rate of 5 mV/s. The optical properties, structure and morphology of the films were characterized through transmittance measurements (Jasco V-570 UV/VIS/NIR spectrometer, Tokyo, Japan), X-ray diffraction (XRD) measurements (SmartLab SE X-ray diffractometer, Tokyo, Japan), scanning electron microscope (SEM) images and Auger electron spectroscopy (AES) measurements (JEOL JAMP-9500F, Akishima, Japan). All equipment was calibrated before use. Photoelectrochemical (PEC) measurement was carried out in a three-electrode system with a Ag/AgCl reference electrode. A 100 mM sodium sulfate (Na2SO4, minimum 99% purity, Kanto Chemical Co., Inc., Tokyo, Japan) solution was used as the electrolyte. The samples were intermittently irradiated with 100 mW/cm 2 light (ABET Technologies 10500 Sun Simulator, Milford, CT, USA) at 5 s intervals and scanned for sample potentials in the range of −1 to 0 V and 0 to 1 V at a scan rate of 5 mV/s.

Deposition Reactions
When 10 mM of tartaric acid was added to the solution containing 10 mM of SnSO4, no precipitation occurred, and the solution was completely transparent. White deposits were obtained through dropping and drying of the solution on the substrate, but all the deposits were dissolved in water during the rinsing process. Thus, tin hydroxide or oxide was not successfully formed as a thin film. This occurred because complexes with tartaric acid were stable and were not decomposed to generate hydroxide during the drying process. The substance remaining on the substrate would be the complex that is easily soluble in water, and thus was dissolved during the rinsing process.
The complex with Na2S2O3 is expected to be less stable than that with tartaric acid, since Na2S2O3 is in fact not common as a complexing agent in metal plating. Solutions with different amounts of Na2S2O3 and NaOH were prepared and the color and precipitation of the solutions are given in Table 1.
For SnSO4 concentration of 10 mM, the solution with 200 mM of Na2S2O3 was colorless and transparent, while the solution with 100 mM of Na2S2O3 was light white. There was no precipitation for these two cases. When the Na2S2O3 amount was further reduced to 50 mM, the color of the solution deepened and became yellowish-white. Since the solution did not exhibit precipitation, it is considered that S2O3 2− and Sn 2+ successfully formed the complex, as reported in ref. [35]: The solutions of conditions I to III were dropped and dried on the substrate. The deposit under condition III (200 mM of Na2S2O3) was removed during the rinsing process. In contrast, for conditions I and II (50 mM and 100 mM of Na2S2O3), a white film remained on the substrate after rinsing. The film thickness was about 0.2 µm. Under condition IV (SnSO4 amount increased to 20 mM), the solution appeared to be yellow and exhibited precipitation, and thus, could not be used for depositing a thin film.

Deposition Reactions
When 10 mM of tartaric acid was added to the solution containing 10 mM of SnSO 4 , no precipitation occurred, and the solution was completely transparent. White deposits were obtained through dropping and drying of the solution on the substrate, but all the deposits were dissolved in water during the rinsing process. Thus, tin hydroxide or oxide was not successfully formed as a thin film. This occurred because complexes with tartaric acid were stable and were not decomposed to generate hydroxide during the drying process. The substance remaining on the substrate would be the complex that is easily soluble in water, and thus was dissolved during the rinsing process.
The complex with Na 2 S 2 O 3 is expected to be less stable than that with tartaric acid, since Na 2 S 2 O 3 is in fact not common as a complexing agent in metal plating. Solutions with different amounts of Na 2 S 2 O 3 and NaOH were prepared and the color and precipitation of the solutions are given in Table 1.
For SnSO 4 concentration of 10 mM, the solution with 200 mM of Na 2 S 2 O 3 was colorless and transparent, while the solution with 100 mM of Na 2 S 2 O 3 was light white. There was no precipitation for these two cases. When the Na 2 S 2 O 3 amount was further reduced to 50 mM, the color of the solution deepened and became yellowish-white. Since the solution did not exhibit precipitation, it is considered that S 2 O 3 2− and Sn 2+ successfully formed the complex, as reported in ref. [35]: The solutions of conditions I to III were dropped and dried on the substrate. The deposit under condition III (200 mM of Na 2 S 2 O 3 ) was removed during the rinsing process. In contrast, for conditions I and II (50 mM and 100 mM of Na 2 S 2 O 3 ), a white film remained on the substrate after rinsing. The film thickness was about 0.2 µm. Under condition IV (SnSO 4 amount increased to 20 mM), the solution appeared to be yellow and exhibited precipitation, and thus, could not be used for depositing a thin film.
Under conditions I-IV, the solution is acidic. NaOH was added to adjust the pH of the solution (conditions V and VI). After adding 2.5 mM of NaOH (condition V), the pH of the solution was similar to that of condition IV, and the solution was yellow, showing precipitates. The pH of the solution was 11.7 after adding 25 mM of NaOH (condition VI) and the color of the solution was initially white and it changed to dark after several minutes. The color change of the alkaline solution may be due to the disproportionation reaction of the tin hydroxide to produce Sn(metal) [48], which causes the solution to become dark.
AES measurements were performed on the films prepared under conditions I and II, and the results are shown in Figure 3. In addition to Sn and O, the films prepared under acidic environment also contain a significant amount of S. It shows that in addition to tin hydroxide, the film also contained sulfur and/or sulfide. minutes. The color change of the alkaline solution may be due to the disproportionation reaction of the tin hydroxide to produce Sn(metal) [48], which causes the solution to become dark.
AES measurements were performed on the films prepared under conditions I and II, and the results are shown in Figure 3. In addition to Sn and O, the films prepared under acidic environment also contain a significant amount of S. It shows that in addition to tin hydroxide, the film also contained sulfur and/or sulfide. Based on the above results, the reactions during the deposition process could be considered as follows.
(a) Decomposition of the complex.
(d) Decomposition of S2O3 2− (release of S) [49]. In an acidic environment, reactions (a)-(e) can occur. Because [Sn(S2O3)2] 2− is not so stable, Sn 2+ ions will be released and react with water to generate tin hydroxide (reactions (a) and (b)). In reaction (d), S2O3 2− ion reacts with H2O or H + to release S. And in reactions (c) and (e), due to the disproportionation reaction, a part of tin and S further react to form SnSx. The final products will be tin hydroxide, SnSx, and S.
In an alkaline environment, reactions (a)-(c) and (f) will occur but not reactions (d) and (e). Due to the presence of a large amount of OH − , Sn 2+ ions released from [Sn(S2O3)2] 2− react with OH − to generate the white precipitate of tin hydroxide (reaction (f)). And in In an alkaline environment, reactions (a)-(c) and (f) will occur but not reactions (d) and (e). Due to the presence of a large amount of OH − , Sn 2+ ions released from [Sn(S 2 O 3 ) 2 ] 2− react with OH − to generate the white precipitate of tin hydroxide (reaction (f)). And in reaction (c), metallic Sn is generated, resulting in a dark color of the solution and the precipitates. The final products will be tin hydroxide and Sn(metal).
In the characterization described below, the following deposition conditions were adopted on the basis of the above analyses: 10 mM of SnSO 4 and 50 mM o fNa 2 S 2 O 3 with no NaOH added (condition I).

SnO 2 Film Properties
The film thickness was about 0.2 µm before annealing and it reduced to about 0.15 µm after annealing at 200 • C and 400 • C. This is mainly due to the conversion of hydroxides into SnO 2 . Figure 4 shows the SEM images of the films on the FTO substrate before and after annealing. Compared to the FTO substrate in Figure 4a, the film before annealing in Figure 4b clearly has more grains. In the SEM image of the SnO 2 thin film prepared at 60 • C using CBD, it was observed that the crystal grains were spherical [22,23]. In contrast, shape of the grains in Figure 4b were irregular. The thin films prepared via DDD contain not only SnO 2 but also substances such as SnS x , which may affect the formation of grains. As can be seen in Figure 4c,d, the size and number of these grains did not change significantly with the increase in annealing temperature.

SnO2 Film Properties
The film thickness was about 0.2 µm before annealing and it reduced to about 0.15 µm after annealing at 200 °C and 400 °C. This is mainly due to the conversion of hydroxides into SnO2. Figure 4 shows the SEM images of the films on the FTO substrate before and after annealing. Compared to the FTO substrate in Figure 4a, the film before annealing in Figure 4b clearly has more grains. In the SEM image of the SnO2 thin film prepared at 60 °C using CBD, it was observed that the crystal grains were spherical [22,23]. In contrast, shape of the grains in Figure 4b were irregular. The thin films prepared via DDD contain not only SnO2 but also substances such as SnSx, which may affect the formation of grains. As can be seen in Figure 4c,d, the size and number of these grains did not change significantly with the increase in annealing temperature.
Since the FTO substrate is also SnO2, the FTO substrate was not used for AES, XRD and PEC measurements, and instead the ITO substrate was used.  Figure 5 shows the AES measurement results after annealing. The S peak intensity decreased after the annealing at 200 °C, and it further decreased after 400 °C annealing. It is presumed that this phenomenon occurs because the S in the film reacts with oxygen in Since the FTO substrate is also SnO 2 , the FTO substrate was not used for AES, XRD and PEC measurements, and instead the ITO substrate was used. Figure 5 shows the AES measurement results after annealing. The S peak intensity decreased after the annealing at 200 • C, and it further decreased after 400 • C annealing. It is presumed that this phenomenon occurs because the S in the film reacts with oxygen in the air during annealing to produce gases such as SO 2 and then escapes. The Na signal is a result of contamination by the deposition solution. Na was not detected in the film before annealing (Figure 3) because the rinsing step removed Na from the surface of the film, while Na located deeper in the film was not removed and was exposed by the decomposition of substances such as SnS on the surface during annealing. In addition, due to the detection limit of AES (about 1%), there may also be a small amount of undetected Na in the films before annealing and after annealing at 200 • C. Taking the SnO 2 standard sample as a reference, the O/Sn composition ratio before and after annealing was calculated: the ratio is 1.11 before annealing and it is 1.65 and 2.15 after annealing at 200 and 400 • C, respectively, as shown in Table 2. As the temperature increases, O/Sn ratio approached two, which indicates that the films annealed at 400 • C are SnO 2 .
sition of substances such as SnS on the surface during annealing. In addition, due to the detection limit of AES (about 1%), there may also be a small amount of undetected Na in the films before annealing and after annealing at 200 °C. Taking the SnO2 standard sample as a reference, the O/Sn composition ratio before and after annealing was calculated: the ratio is 1.11 before annealing and it is 1.65 and 2.15 after annealing at 200 and 400 °C, respectively, as shown in Table 2. As the temperature increases, O/Sn ratio approached two, which indicates that the films annealed at 400 °C are SnO2.  The structural properties of the films were characterized via XRD measurements, as shown in Figure 6. However, we did not observe any peak other than those of the ITO substrate, as shown in Figure 6a. Comparison with the powder diffraction file ICDD 00-041-1445 for SnO2 also shows that the peaks in Figure 6a are not from SnO2. Therefore, the films before and after annealing may be amorphous. The other substances mixed in the films may also affect crystallization, just as for the grain formation (SEM results).  The structural properties of the films were characterized via XRD measurements, as shown in Figure 6. However, we did not observe any peak other than those of the ITO substrate, as shown in Figure 6a. Comparison with the powder diffraction file ICDD 00-041-1445 for SnO 2 also shows that the peaks in Figure 6a are not from SnO 2 . Therefore, the films before and after annealing may be amorphous. The other substances mixed in the films may also affect crystallization, just as for the grain formation (SEM results). The results of optical transmittance measurement are shown in Figure 7. The transmittance of the film in the visible region before annealing was 55-87%, which slightly decreases to 55-83% after annealing at 200 °C, and increases to 63-89% after annealing at 400 °C. The decrease in the visible region is due to scattering by the surface roughness and absorption by the narrow-bandgap SnSx phase [50,51] in the film. The literature value of the band gap is 3.6 eV for SnO2 [4,5] but edge around the there is no clear absorption corresponding wavelength (340 nm). This would indicate the amorphous nature of the film. The transmittance of the films annealed at 400 °C is higher than that of the films annealed at 200 °C. This could be because the narrow-bandgap SnSx phase in the film was converted to SnO2 through annealing [52,53]; the S content in the film was decreased after annealing, as can be seen from the AES results. Due to the roughness of the surface of The results of optical transmittance measurement are shown in Figure 7. The transmittance of the film in the visible region before annealing was 55-87%, which slightly decreases to 55-83% after annealing at 200 • C, and increases to 63-89% after annealing at 400 • C. The decrease in the visible region is due to scattering by the surface roughness and absorption by the narrow-bandgap SnS x phase [50,51] in the film. The literature value of the band gap is 3.6 eV for SnO 2 [4,5] but edge around the there is no clear absorption corresponding wavelength (340 nm). This would indicate the amorphous nature of the film. The transmittance of the films annealed at 400 • C is higher than that of the films annealed at 200 • C. This could be because the narrow-bandgap SnS x phase in the film was converted to SnO 2 through annealing [52,53]; the S content in the film was decreased after annealing, as can be seen from the AES results. Due to the roughness of the surface of films, the transmittance results obtained from measurements at different points of the same film have variations of 1-2%, but this error is not so serious to affect the above conclusions. mittance of the film in the visible region before annealing was 55-87%, which slightly decreases to 55-83% after annealing at 200 °C, and increases to 63-89% after annealing at 400 °C. The decrease in the visible region is due to scattering by the surface roughness and absorption by the narrow-bandgap SnSx phase [50,51] in the film. The literature value of the band gap is 3.6 eV for SnO2 [4,5] but edge around the there is no clear absorption corresponding wavelength (340 nm). This would indicate the amorphous nature of the film. The transmittance of the films annealed at 400 °C is higher than that of the films annealed at 200 °C. This could be because the narrow-bandgap SnSx phase in the film was converted to SnO2 through annealing [52,53]; the S content in the film was decreased after annealing, as can be seen from the AES results. Due to the roughness of the surface of films, the transmittance results obtained from measurements at different points of the same film have variations of 1-2%, but this error is not so serious to affect the above conclusions. The films prepared on the alkali-free glass substrate were subjected to resistivity measurement and the calculated resistivity values are listed in Table 2. It was confirmed that the alkali-free glass substrate itself is insulating. Current-voltage (I-V) measurements The films prepared on the alkali-free glass substrate were subjected to resistivity measurement and the calculated resistivity values are listed in Table 2. It was confirmed that the alkali-free glass substrate itself is insulating. Current-voltage (I-V) measurements were performed on three different points for each sample, and the average value of resistivity was calculated. The resistivity before annealing was 1.4 × 10 6 Ωcm and it decreased to 3.4 × 10 4 Ωcm after annealing at 200 • C, further decreasing to 1.8 Ωcm after annealing at 400 • C. Thus, as the annealing temperature increases, the film resistivity gradually decreases, which is consistent with the results of SnO 2 film prepared using other methods [54]. Improvement in crystallinity and/or increase in the amount of oxide vacancies or excess metal ions will lead to a decrease in film resistivity [55].
The PEC measurement results of the films before and after annealing are shown in Figure 8. During the positive scan, the properties of the film changed due to the reactions caused by the current, resulting in discontinuity at 0 V between the data of the positive and negative scans. Both before and after annealing, photo response was observed under positive bias, indicating that the films were all n-type. The photo response of the film annealed at 200 • C is significantly larger than those of the films before annealing and those annealed at 400 • C. Therefore, in the following heterojunction experiment, the film was annealed at 200 • C. SnO 2 films have so far been prepared using another simple chemical technique, called CBD [22,23]. Following are the differences in deposition results between CBD and DDD. The films prepared using CBD have higher transmittance (about 70-85%), but the resistivity of the films before annealing is very high (not measurable) and the resistivity after annealing at 450 • C is as high as 10 3 Ωcm. The SnO 2 films prepared using DDD have conductivity both before and after annealing, and the resistivity of the films after annealing at 400 • C is only 1.8 Ωcm. The transmittance in the visible field is about 50-80%. Thus, the DDD-SnO 2 films have high conductivity and low transmittance compared with the CBD-SnO 2 films. Figure 8. During the positive scan, the properties of the film changed due to the reactions caused by the current, resulting in discontinuity at 0 V between the data of the positive and negative scans. Both before and after annealing, photo response was observed under positive bias, indicating that the films were all n-type. The photo response of the film annealed at 200 °C is significantly larger than those of the films before annealing and those annealed at 400 °C. Therefore, in the following heterojunction experiment, the film was annealed at 200 °C. SnO2 films have so far been prepared using another simple chemical technique, called CBD [22,23]. Following are the differences in deposition results between CBD and DDD. The films prepared using CBD have higher transmittance (about 70-85%), but the resistivity of the films before annealing is very high (not measurable) and the resistivity after annealing at 450 °C is as high as 10 3 Ωcm. The SnO2 films prepared using DDD have conductivity both before and after annealing, and the resistivity of the films after annealing at 400 °C is only 1.8 Ωcm. The transmittance in the visible field is about 50-80%. Thus, the DDD-SnO2 films have high conductivity and low transmittance compared with the CBD-SnO2 films.

Co3O4/SnO2 Heterojunction
The thickness of the Co3O4/ SnO2 heterojunction has been measured to be about 0.15 µm for SnO2 and 0.2 µm for Co3O4. The XRD results of the annealed heterojunction are shown in Figure 6b. In addition to the peaks of the ITO substrate, peaks caused by the films were observed. Comparison with the powder diffraction file ICDD 00-009-0418 shows that these peaks are the (311), (400) and (440) peaks of Co3O4, respectively. It is consistent with the results obtained from our previous study on Co3O4 films [30]. Thus, Co3O4 was successfully deposited onto SnO2 to form the Co3O4/SnO2 heterojunction. Figure 9 shows the J-V characteristics of the Co3O4/SnO2 heterojunction in dark and under AM1.5 (100 mW/cm 2 ). In Figure 9a, rectification characteristics and photo-responsivity are observed. As shown in the enlarged figure (Figure 9b), its open-circuit voltage

Co 3 O 4 /SnO 2 Heterojunction
The thickness of the Co 3 O 4 /SnO 2 heterojunction has been measured to be about 0.15 µm for SnO 2 and 0.2 µm for Co 3 O 4 . The XRD results of the annealed heterojunction are shown in Figure 6b. In addition to the peaks of the ITO substrate, peaks caused by the films were observed. Comparison with the powder diffraction file ICDD 00-009-0418 shows that these peaks are the (311), (400) and (440) peaks of Co 3 O 4 , respectively. It is consistent with the results obtained from our previous study on Co 3 O 4 films [30]. Thus, Co 3 O 4 was successfully deposited onto SnO 2 to form the Co 3 O 4 /SnO 2 heterojunction. Figure 9 shows the J-V characteristics of the Co 3 O 4 /SnO 2 heterojunction in dark and under AM1.5 (100 mW/cm 2 ). In Figure 9a, rectification characteristics and photoresponsivity are observed. As shown in the enlarged figure (Figure 9b), its open-circuit voltage V oc is about 50 mV and its short-circuit current J sc is about 2.2 µA/cm 2 . The calculated energy conversion efficiency is about 2.3 × 10 −5 %. The above results indicate that Co 3 O 4 /SnO 2 heterojunctions can be used as diodes and solar cells. However, its leakage current is seriously large, which affects its performance in solar cells. Voc is about 50 mV and its short-circuit current Jsc is about 2.2 µA/cm 2 . The calculated energy conversion efficiency is about 2.3 × 10 −5 %. The above results indicate that Co3O4/SnO2 heterojunctions can be used as diodes and solar cells. However, its leakage current is seriously large, which affects its performance in solar cells.
(a) (b) As noted in Introduction, there are only a few reports on SnO2-based p-n heterojunction diodes; SnO2 has been mainly used as transparent electrodes, dye-adsorbed films of dye-sensitized solar cells or electron selective layers of perovskite solar cells [15][16][17][18][19][20][21]. SnO2based heterojunctions have been mostly used in catalysts and sensors [41][42][43][44][45][46][47]. In this study, it was shown for the first time that a Co3O4/SnO2 heterojunction is applicable in a solar cell. And due to the wide bandgap and transparency, SnO2 can be combined with a wide bandgap p-type semiconductor, such as NiO, to fabricate transparent solar cells. In addition, SnO2 prepared using DDD can also be used to prepare low-cost transparent elec- As noted in Introduction, there are only a few reports on SnO 2 -based p-n heterojunction diodes; SnO 2 has been mainly used as transparent electrodes, dye-adsorbed films of dye-sensitized solar cells or electron selective layers of perovskite solar cells [15][16][17][18][19][20][21]. SnO 2 -based heterojunctions have been mostly used in catalysts and sensors [41][42][43][44][45][46][47]. In this study, it was shown for the first time that a Co 3 O 4 /SnO 2 heterojunction is applicable in a solar cell. And due to the wide bandgap and transparency, SnO 2 can be combined with a wide bandgap p-type semiconductor, such as NiO, to fabricate transparent solar cells. In addition, SnO 2 prepared using DDD can also be used to prepare low-cost transparent electronic devices, such as transparent diodes [56] and transparent thin-film transistors (TFT) [57]. Those devices can be prepared on a non-conductive substrate, such as glass at low temperatures using DDD, which will drastically reduce the fabrication cost.
Although Co 3 O 4 /SnO 2 heterojunction solar cells prepared via DDD are not efficient at present, they can be applied in the field of internet of things (IoT) devices. It was reported that perovskite solar cells can power wireless devices (radio frequency identification), which required 10-45 µW of power [58]. Gas sensors based on the WS 2x Se 2-2x alloy requires only 0.75-21 µW of power to operate [59]. A self-powered visible-light-transparent CO 2 gas sensor based on a NiO/ZnO solar cell was also reported [60]. Such low-power IoT devices are increasingly used in daily life, and solar cells are useful as a power source even if they are not of high-efficiency.
For future research, the quality of the SnO 2 films should be improved for better performance of solar cells, and their applications need to be examined more, such as in transparent diodes, transistor, and solar cells.

Conclusions
In this study, tin hydroxide films were successfully prepared via DDD using Na 2 S 2 O 3 as the complexing agent and were then converted into SnO 2 films through annealing. The complexing agent was used for the first time to suppress spontaneous hydrolysis and precipitation in the deposition solution.
The transmittance of the annealed films was higher than 50% in the visual range. The results of the AES measurement showed that the O/Sn composition ratio of the annealed films approached two, indicating that the annealed films were SnO 2 . The PEC results showed that the annealed films were n-type, and the photo response of the films annealed at 200 • C was significantly larger than that of the films annealed at 400 • C. And the results of resistivity measurement showed that the films annealed at 400 • C have a lower resistivity of 1.8 Ωcm.
In this study, Co 3 O 4 /SnO 2 p-n heterojunctions were fabricated with Co 3 O 4 deposited via DDD, and rectification and photovoltaic properties were observed. It was proven that Co 3 O 4 /SnO 2 heterojunctions can potentially be used in solar cells. Although the energy conversion efficiency was still low, about 10 −5 %, the cell could be used in low-power IoT applications. Moreover, it is demonstrated for the first time that oxide thin film solar cells can be fabricated using only DDD. The metal oxide materials, such as SnO 2, are abundant, and the equipment required for DDD is very simple and low cost. Thus, this work could contribute to reducing the fabrication cost of various thin film devices, especially in the field of transparent electronics.

Conflicts of Interest:
The authors declare no conflict of interest.