Introduction

SnO2 is an n-type semiconductor with a wide bandgap (>3 eV), and has been utilized in various technical fields. It is one of the most popular materials for gas sensors,1,2 and doped thin films can achieve high conductivity and thus are widely used as transparent electrodes.3,4 It is also used as an electrode material for dye-sensitized solar cells5,6 and perovskite solar cells.7,8 Moreover, with its n-type conductivity, SnO2 can be used for constructing a pn heterostructure with a p-type semiconductor. SnO2-based composite and core–shell structures were shown to be useful for chemical sensors.9,10,11,12 In addition, since tin is an inexpensive and nontoxic material, SnO2-based heterostructures can be advantageous for solar cell applications, where mass production of thin films is needed at low cost. Solar cells consisting of SnO2 and a narrower-bandgap material Cu2O have been reported.13 Our group also reported the fabrication of SnO2/SnS and SnO2/Co3O4 solar cells.14,15 In addition, a few groups have reported the fabrication of SnO2/NiO UV photodetectors.16,17

In a previous work, we reported a new chemical deposition technique, drop–dry deposition (DDD).18 DDD is a simple chemical method for thin film fabrication: a thin film can be deposited by dropping the deposition solution on the substrate with subsequent drying, as shown in Fig. 1. During the drying process, a metal hydroxide or oxide with low solubility is deposited on the substrate. DDD has been applied for fabrication of several oxides and hydroxides (Mg(OH)2, NiO, Co3O4).18,19,20 NiO and Co3O4 are formed by annealing the hydroxide films obtained by DDD in a furnace. DDD is considered to have some advantages over other chemical techniques for semiconductor thin film deposition. For example, for spray pyrolysis, an elaborate apparatus (atomizer and heater) is needed, and heat treatment is needed for the sol–gel method. Chemical bath deposition (CBD) is a simple and low-temperature process, but the material efficiency is low. DDD requires only a very simple apparatus, and the material efficiency is higher than that with CBD. Therefore, DDD has good potential for production of thin-film solar cells. Moreover, DDD is a low-temperature process and thus will also be useful for transparent electronics applications, where transparent electronic devices need to be fabricated on glass or plastic sheets.

Fig. 1
figure 1

Schematic of the drop–dry deposition (DDD) process.

Full size image

For the fabrication of solar cells, deposition of both p-type and n-type semiconductor thin films is needed. As noted above, we succeeded in DDD of p-type semiconductors (Co3O4 and NiO). ZnO may be the most popular n-type semiconductor among oxides, but ZnO cannot be deposited by DDD because of the high solubility of ZnO in an alkaline solution. Thus, we attempted DDD of another popular n-type semiconductor, SnO2.15 In that previous work, we used SnSO4 as the tin source. However, an aqueous solution containing tin(II) (Sn2+) ions tends to be unstable: Sn2+ ions undergo a hydrolysis reaction, and tin hydroxides are spontaneously formed and precipitate. The hydrolysis reaction is formally expressed as21

$$ {\text{Sn}}^{{{2} + }} + {\text{ 2H}}_{{2}} {\text{O}} = {\text{Sn}}\left( {{\text{OH}}} \right)_{{2}} + {\text{ 2H}}^{ + } $$
(1)

although more complex ions and molecules of tin are formed in the actual reaction. The spontaneous hydrolysis can be suppressed by using a proper complexing agent. In the previous work, we used sodium thiosulfate (S2O32− ions) as the complexing agent and fabricated SnO2 thin films by DDD.15 In addition, SnO2/Co3O4 solar cells were fabricated with Co3O4 layers also fabricated by DDD. However, the deposition solution was still not very stable: it became hazy mainly because of sulfur colloids. Moreover, a significant amount of sulfur was contained in the deposited films.

In this work, we avoid the hydrolysis reaction by using sodium stannate Na2SnO3, where tin is in the state of tin(IV) (Sn4+). Tin(IV) does not cause the hydrolysis reaction (1), and thus the solution is expected to be stable. In fact, Na2SnO3 is one of the common tin sources used in tin metal plating. Na2SnO3 has also been used to synthesize SnO2 nanoparticles, which were primarily applied for catalysts and electrodes.22,23,24 As shown below, SnO2 films can be deposited by DDD using solutions containing Na2SnO3.

In addition, in this work we fabricate SnO2/NiO transparent solar cells, combining the SnO2 film with the NiO film, which is also deposited by DDD. NiO is a wide-bandgap p-type semiconductor. Compared with conventional thin film solar cells including perovskite,25 organic,26 and dye-sensitized solar cells,27 visible-light transparent solar cells generally have low energy-conversion efficiency because in transparent cells, only the ultraviolet (UV) range of the solar spectrum can be used for power generation. However, the transparent cells can be used as window panels and also for designed objects,28 and thus can have wide applicability as a ubiquitous power source. There are a number of papers on NiO-based transparent solar cells.28,29,30,31,32,33,34 Our group also fabricated NiO-based diodes and solar cells with NiO layer deposited by electrochemical deposition35 and DDD.19 In those previous works, ZnO was used as the n-type layer in the pn heterojunction. ZnO cannot be deposited by DDD as noted above, and thus in this work, we fabricate solar cells by DDD with SnO2 adopted for the n-type layer. Although SnO2 is another popular n-type wide-bandgap material, there are only a few reports of fabrication of SnO2/NiO heterojunction photodetectors as noted above.16,17 We fabricate SnO2/NiO solar cells using DDD; both the fabrication technique and combination of materials are unique in this work.

In the next section, details of the DDD of SnO2 using Na2SnO3 are given. Then, characterization results for the SnO2 films are given. Finally, characteristics of the SnO2/NiO solar cell are given and discussed.

Experimental

The DDD procedure is schematically shown in Fig. 1. The deposition solution for SnO2 was an aqueous solution containing Na2SnO3. A small amount of the solution was dropped on the substrate heated to 60°C on the heater plate. In the drying process, SnO2, having a low solubility, was first deposited, and then more soluble substances precipitated on it. After the dropped solution was completely dried up, the sample was rinsed in water to remove the soluble substances. The process was repeated until a sufficient film thickness was achieved. The substrate was indium-tin-oxide (ITO)-coated glass sheets. The formation of SnO2 from Na2SnO3 can be written as

$$ {\text{Na}}_{{2}} {\text{SnO}}_{{3}} + {\text{H}}_{{2}} {\text{O}} = {\text{SnO}}_{{2}} + {\text{ 2NaOH}} $$
(2)

In fact, Na2SnO3 is tri-hydrated, and the actual formula of the stannate ion will be [Sn(OH)6]2−. Then the formation reaction of SnO2 will actually be written as

$$ {\text{Na}}_{{2}} \left[ {{\text{Sn}}\left( {{\text{OH}}} \right)_{{6}} } \right] = {\text{SnO}}_{{2}} + {\text{ 2NaOH }} + {\text{ 2H}}_{{2}} {\text{O}} $$
(3)

The validity of this chemical equation will be examined on the basis of the deposition results below. The deposition conditions adopted are shown in Table I. The films were annealed in a conventional quartz-tube furnace in air.

For the fabrication of a pn heterostructure, a NiO film was also fabricated by DDD following the recipe reported in the previous work.19 The deposition solution contained 10 mM Ni(NO3)2 and 15 mM NaOH. The heater plate temperature was 60°C, and the drop–dry cycles were repeated twice. The as-deposited film was Ni(OH)2, and it was converted to NiO by annealing in air at 400°C.

Table I Deposition conditions for the DDD of SnO2
Full size table

The film thickness was evaluated using an Accretech Surfcom 1400G profilometer. A JEOL JAMP-9500F field-emission microprobe was used for the elemental composition measurement, and scanning electron microscopy (SEM) was conducted at a probe voltage of 10 keV. A JASCO V-570 optical spectrophotometer was used to measure transmittance, with the ITO substrate as reference: the transmittance of the film was obtained by dividing the data of the sample by those for the bare ITO substrate. X-ray diffraction (XRD) was measured using a Rigaku SmartLab SE x-ray diffractometer with a Cu Kα radiation source. Raman scattering was measured using a JASCO NRS-3300 Raman spectroscope. Photoelectrochemical (PEC) measurement was performed using a three-electrode electrochemical cell with an Ag/AgCl reference electrode. An ABET Technologies 10500 solar simulator was used as the light source. The electrolyte contained 100 mM Na2SO4, and 100 mW/cm2 light having the AM1.5 spectrum was irradiated intermittently at 5 s intervals. For the current–voltage (I–V) measurement of the heterostructure samples, 1 mm2 × 1 mm2 indium electrodes were fabricated on the sample surface by vacuum evaporation. The solar simulator mentioned above was used to observe photoconductive and photovoltaic properties.

Results and Discussion

SnO2 Films

A transparent film with a thickness of about 0.2 µm was deposited under the conditions shown in Table I. We examined many different deposition conditions, and found that with greater Na2SnO3 content in the solution or a larger amount of the solution dropped on the substrate, the film tended to be more hazy: the surface roughness tended to be enhanced. A transparent film with sufficient thickness for characterization was deposited with four drop–dry cycles under the conditions shown in the table. The surface morphology (SEM image) of the as-deposited film is shown in Fig. 2. Round grain structure can be seen on a compact film.

Fig. 2
figure 2

SEM image of the surface of the as-deposited SnO2 film.

Full size image

Figure 3 shows the AES spectra for the as-deposited film and the film annealed at 400°C. As shown in the figure, the O/Sn ratio did not change with the annealing. Using a purchased SnO2 chemical as the reference, the O/Sn ratio was evaluated. The ratio is about 2 for both the as-deposited and annealed films. This indicates that the deposit is SnO2 or Sn(OH)2. However, Sn in the source chemical (Na2SnO3) is tin(IV) (Sn4+ charge state), and thus Sn(OH)2, a tin(II) hydroxide, would not be formed. Therefore, as shown in the reaction (2) or (3), SnO2 was synthesized and deposited by DDD using Na2SnO3.

Fig. 3
figure 3

AES spectra for the as-deposited SnO2 film and 400°C-annealed film.

Full size image

In the previous work, we used SnSO4, a tin(II) compound, as the Sn source in DDD with a complexing agent Na2S2O3.15 In contrast to the results in Fig. 2, a significant amount of S was detected by AES for the as-deposited film, and the deposit was considered to be a mixture of Sn(OH)2 and SnS. Thus, by using Na2SnO3 as the source material, the formation of tin hydroxide and sulfide was suppressed, and the deposit became mainly SnO2.

Figure 4 shows the optical transmission spectra for the as-deposited and 400°C-annealed films. No absorption edge was observed for the as-deposited film, while an absorption edge seems to appear around 370 nm for the annealed film. The bandgap was roughly estimated to be in a range of 3.65–3.75 eV by the (αhν)2-hν plot, where α is the absorption coefficient and hν the photon energy; since a clear straight line part was not observed, the bandgap value was not accurately determined. The bandgap values of SnO2 were reported to be 3.5–4.6 eV, i.e., were scattered in a rather wide range.36,37,38,39,40 Thus, the apparent bandgap seems to depend on the fabrication method and conditions. The value estimated for the annealed film in this work is in the range reported for the SnO2 bandgap. On the other hand, no clear absorption was observed for the as-deposited film, and thus its bandgap will be considerably larger.

Fig. 4
figure 4

Optical transmission spectra for the as-deposited SnO2 film and 400°C-annealed film.

Full size image

To assess crystallinity, we measured XRD for the as-deposited and 400°C-annealed films, but no peaks attributed to SnO2 were observed, as shown in Fig. 5. The marked peaks are all due to ITO. In addition, we performed the Raman scattering measurement, but no Raman signal due to SnO2 was observed for the as-deposited and annealed films as shown in Fig. 6, where the observed broad peaks are due to ITO. Thus the fabricated films will be amorphous or nanocrystalline even after the annealing. This could lead to a larger apparent bandgap than that of the crystalline SnO2, especially for the as-deposited sample.

Fig. 5
figure 5

XRD spectra for the as-deposited and 400°C-annealed films. The marked peaks are due to ITO.

Full size image
Fig. 6
figure 6

Raman spectra for the as-deposited film and the film annealed at 400°C.

Full size image

Figure 7 shows the PEC measurement results: Fig. 7a for the as-deposited film, and Fig. 7b for the films annealed at 200°C and 400°C. In the PEC measurement, a current due to the minority carriers is significantly enhanced by photo-irradiation. Thus, a positive photocurrent is dominantly observed for an n-type sample, and a negative photocurrent for a p-type sample. As shown in the figure, a positive photocurrent was observed in the positive potential range, and therefore the deposited films are n-type. This is consistent with the fact that even before annealing, the deposit is SnO2, which is an n-type semiconductor. The magnitude of the photocurrent was much larger for the annealed films than for the as-deposited film. This could be mainly because light absorption is much lower for the as-deposited film, as indicated by the optical transmission data shown in Fig. 4. Differences in crystallinity may have also affected the photosensitivity; however, the effects of the crystallinity cannot be clearly understood because we failed to observe the XRD or Raman scattering for those samples and thus could not assess the crystallinity.

Fig. 7
figure 7

PEC results for (a) the as-deposited and (b) annealed films.

Full size image

SnO2/NiO Heterojunction

The Ni(OH)2 layer was first deposited by DDD on the ITO substrate, and then it was converted to NiO by annealing at 400°C. The properties of the NiO layer have been reported in the previous work.19 The thickness of the NiO layer was about 0.2 µm. Then the SnO2 layer was deposited on NiO. Figure 8 shows the optical transmission spectrum of the heterostructure thus fabricated. The annealing was not performed after the SnO2 deposition. As can be seen from the figure, the sample is highly transparent in the visible range.

Fig. 8
figure 8

Optical transmission spectrum of the SnO2/NiO heterostructure.

Full size image

The I–V characteristics of the SnO2/NiO heterostructure diode with as-deposited SnO2 are shown in Fig. 9. Clear rectification properties were observed as shown in Fig. 9a, although the leakage current started to increase at a rather small reverse voltage (about −0.9 V). The dashed line shows the current under the photo-irradiation (AM1.5, 100 mW/cm2). The photovoltaic properties were observed as shown in the enlarged figure Fig. 9b. The solar cell parameters deduced from the figure are as follows: open-circuit voltage VOC = 390 mV, short-circuit current JSC = 0.7 μA/cm2, fill factor FF = 0.3, and energy conversion efficiency η = 8.2 × 10−5%. VOC is fairly large despite the very small JSC, which indicates good rectification properties of the diode.

Fig. 9
figure 9

(a) I–V characteristics of the SnO2/NiO heterostructure diode. The photovoltaic output is enlarged and shown in (b).

Full size image

The heterostructure samples were annealed at 200°C and 400°C after the SnO2 deposition. According to the PEC results shown in Fig. 7, the photocurrent was significantly enhanced by the annealing. However, for the annealed heterostructures, no rectification properties or photovoltaic properties were observed. We also fabricated NiO/SnO2 heterostructures; the SnO2 layer was first deposited on the ITO substrate and then the Ni(OH)2 layer, and the sample was annealed at 400°C. However, the observed I–V characteristics were almost ohmic. Thus, good rectification properties were not obtained if the sample was annealed after the deposition of the second layer (fabrication of the heterostructure). The reason for the poor properties of the annealed heterostructure is unclear, but the volume change due to the annealing could be a possible main cause of the deterioration. When Ni(OH)2 is converted to NiO by annealing, the volume is decreased to about half. For SnO2, the volume change is thus not significant, but some small volume changes may be induced due to the escape of water included in the as-deposited film. Therefore, when the sample is annealed after the second layer deposition, the interface may undergo stress due to the different amounts of volume change between the first and second layers. Then the interface bonding could be partly destroyed, and the pn junction properties could be degraded.

As noted in the introduction, there are only a few reports on SnO2/NiO photosensitive diodes.16,17 In those previous works, the devices were fabricated by electrospinning16 and electron beam evaporation.17 In this work, as shown above, a transparent oxide solar cell was fabricated by DDD. DDD is a very simple, low-cost technique, and thus would be advantageous for the fabrication of low-cost solar cells. In future work, we will attempt to improve the performance of the SnO2/NiO solar cell and also fabricate SnO2-based heterostructures with another low-cost p-type material, such as Co3O4 and SnS, to establish a low-cost technique for solar cell production. In addition, we should attempt to apply the DDD-SnO2 films to perovskite and dye-sensitized cells and also to devices for transparent electronics such as thin-film transistors.

Conclusion

SnO2 thin films were deposited on ITO substrates by DDD, which is a simple, low-cost technique for thin film deposition. The deposition solution contained Na2SnO3 as the Sn source. The solution was dropped on the substrate heated to 60°C on a heater plate. After drying, the sample was rinsed in water, and the drop–dry cycles were repeated to obtain the necessary thickness. AES confirmed that the O/Sn ratio was about 2. The film is transparent in the visible range and shows n-type response in the PEC measurement. The pn heterostructure was fabricated by depositing SnO2 on NiO. The NiO film was also fabricated by DDD. Ni(OH)2 was deposited using a solution containing Ni(NO3)2 and NaOH, and was then converted to NiO by annealing at 400°C. The SnO2/NiO heterostructure is transparent in the visible range and shows clear rectification properties and photovoltaic effects. Thus, a transparent solar cell has been fabricated by DDD for the first time. However, the efficiency is still low: the output current is less than 1 µA/cm2. Annealing would improve the bulk properties of the SnO2 film but deteriorate the properties of the heterointerface. To improve the solar cell performance, the deterioration of the interface should be suppressed by preventing the interdiffusion and strain generation at the heterointerface.