Improved performance of solar cells using chemically synthesized SnSe nanosheets as light absorption layers

Chemically synthesized semiconductor nanocrystals have characteristics that differ from bulk materials, such as quantum effects and large specific surface areas, and are therefore being studied for application to a wide range of electronic and optoelectronic devices. The main material systems include CdSe, CdS, CdTe, PbS, PbSe, but recently, the requirement to eliminate toxic metals such as Cd and Pb is driving the development of new materials [1‐7]. SnSe is one of the most promising materials [8‐11], and has been shown to have the potential to be applied to various devices [12‐15]. Bandgap energies of bulk SnSe have been reported, including about 0.8 eV for indirect transitions [16, 17] and about 0.9 eV for direct transitions [18]. The properties of devices using nanocrystals are highly dependent on their shape and size, and SnSe is prone to quantum confinement effects because the Bohr radius is clearly larger than the size of the nanocrystals [19]. SnSe nanosheets (NSs) are promising materials for solar cells because their bandgap energy overlaps well with solar energy.

SnSe has a layered GeS-type crystal structure, which promotes two-dimensional anisotropic growth [5]. We have successfully synthesized SnSe NSs, nanorods, and nanodots by adjusting the nanocrystal synthesis conditions using the hot injection method [11]. We reported that the bandgap energy can be tuned from 0.82 up to 1.74 eV in response to these shape changes. It has been reported that an even higher bandgap energy of 2.9 eV can be achieved in SnSe NSs prepared using the one-pot method [9]. Such wide-gap semiconductors have been developed in the electronics field as high-voltage power device materials, and in the optoelectronics field, they are expected to be used as materials that can absorb and emit short-wavelength light [20‐23]. It can be applied as a top layer material in tandem solar cells that can achieve high photoelectric conversion efficiency (PCE) by stacking multiple light absorption layers with different absorption wavelengths. In order to use it as the top layer of a tandem solar cell, it is required not only to have a wide bandgap but also to have a band diagram that allows carriers to flow smoothly to both electrodes. Ga₂O₃, a wide-gap semiconductor that has attracted attention in recent years, can be produced by chemical synthesis [24], but its conduction band is far away from the vacuum level [25], making it difficult to use as the top layer of tandem solar cells. On the other hand, SnSe nanocrystals meet the demand for a band diagram as a top layer because their conduction band is at a good distance from the vacuum level [26, 27]. SnSe solar cells with good PCE have been limited to those deposited on substrates by vapor deposition, including those with the highest 6.44% [12, 28‐36]. Chemically synthesized SnSe nanocrystals are attractive because solar cells can be easily and inexpensively fabricated by coating [12, 31]. In this study, we investigated methods for improving the performance of solar cells using light absorption layers coated with chemically synthesized SnSe NSs. Two methods for synthesizing SnSe NSs were investigated and their suitability for use as light absorption layers was compared. A conductive polymer was mixed to support conductivity between NSs. Solar cells were fabricated using SnSe NSs films formed on Si substrates as light absorption layers. After investigating the relationship between the solar cell performance and the mixing ratio of NSs and conductive polymer, we achieved very high PCE compared to previously reported solar cells using chemically synthesized SnSe nanocrystals.

SnSe NSs were chemically synthesized using two known methods: hot injection method and one-pot method. In the hot injection method [37], 0.26942 g SnO, 5.0 mL oleic acid, and 4.0 mL octadecene were first stirred at 230 °C for 1 h in a nitrogen atmosphere in a four-necked flask. Separately, 0.15792 g Se and 2.0 mL trioctylphosphine were stirred at 60 °C for 1 h to be fully dissolved, then injected into the above solution using a syringe, and heated at 230 °C for 150 s in a nitrogen atmosphere. The obtained SnSe NSs were washed with ethanol and finally dispersed in toluene. In the one-pot method [38], 0.01 g SnCl₂ and 20 mL oleylamine were mixed in a four-necked flask. Separately, 0.15792 g Se and 2.0 mL trioctylphosphine were mixed, and then 1 mL hexamethyldisilazane (HMDS) was mixed. The mixture was added to the above solution, stirred for 15 min, and further heated at 240 °C for 30 min in a nitrogen atmosphere. The obtained SnSe NSs were similarly washed with ethanol and finally dispersed in toluene.

A light absorption layer containing SnSe NSs was deposited on a (001) n-Si substrate by the following procedure. In the case of a film consisting only of SnSe NSs, NSs dispersed in toluene was dropped and dried, and then a conductive polymer was spin coated (4000 rpm 2 min) on top of it. When forming a film by mixing SnSe NSs and conductive polymer, the mixed solution was spin coated on a Si substrate at 1000 rpm for 2 min. To prepare the mixed solution, a completely dried SnSe NSs solution was added to a conductive polymer and stirred well. As conductive polymers, poly[(3,4-ethylenedioxythiophene)-(styrenesulfoonate) (PEDOT: PSS) or poly[2-methoxy-5-(3′,7′-dimethyloctyloxy)-1,4-phenyleneninylene] (PPV) was used.

SnSe NSs solar cells were fabricated as follows. Si substrates were semiconductor grade and their thickness was 270 μm. After forming films containing SnSe NSs on substrates using the method described above, annealing was performed at 130 °C for 15 min in a nitrogen atmosphere to remove the toluene solvent. Thereafter, electrodes were formed by vapor deposition. Au fishbone-shaped electrodes were used for the anodes, and Al electrodes were used for the cathodes. SnSe NSs layer side was the anode, and Si substrate side was the cathode. That is, the SnSe NSs layers were exposed through the gap between the fishbone-shaped electrodes on the anode side. The thickness of both electrodes was approximately 100 nm. The size of the solar cell was 15 mm x 15 mm.

The microstructure of the samples was observed using a transmission electron microscope (TEM) and a scanning electron microscope (SEM). The crystal structure of the sample was confirmed by X-ray diffraction (XRD) measurement using Cu-kα radiation as a light source. The bandgap energy was evaluated by optical absorption spectrometry. In the evaluation of solar cells, a halogen lamp was used as the irradiation light source.

The shapes of the NSs obtained by the two synthetic methods were very different. Figure 1 shows electron microscopy images of SnSe NSs synthesized by hot injection and one-pot methods. As shown in Fig. 1a, the hot injection method produced irregularly shaped NSs with one side on the order of micrometers. A curved pattern was observed on the surface of the NSs, suggesting that the NSs themselves were folded or intertwined. On the other hand, the one-pot method produced tiny square or triangular NSs with relatively uniform size, as shown in Fig. 1b. The maximum side length was about 500 nm, but there were many NSs much smaller than that. Although some of the NSs appeared to be in close contact with each other, almost all of them seemed to be separate and independent.

We investigated the fabrication of light absorption layers using SnSe NSs synthesized by hot injection method. Figure 2a and b are in-plane and cross-sectional SEM images of a film formed by dropping and drying NSs dispersed in toluene. NSs did not form flat films. Considering that each sheet was several microns in size, the folded structure seen here is thought to be due to deformation of the sheet itself. It seems that SnSe NSs of various sizes not only cause unevenness themselves, but also tend to become entangled with each other. Due to agglomeration, Due to aggregation, these SnSe NSs covered only part of the substrate in one drop drying, and in order to cover the entire substrate, it was necessary to repeat the drop drying until the film thickness increased to 10 μm or more. The difference in unevenness in the thickness direction was at most 10 μm when the average film thickness was approximately 14.8 μm. Since the film has such large irregularities, even if a conductive polymer film is applied on it, the metal electrode formed thereon will be cut by the cliffs and carrier flow will be obstructed.

Figure 2c and d show examples of films formed by mixing with conductive polymers. Figure 2c shows the case when 0.675 µL SnSe NSs and 150 µL PEDOT: PSS were mixed, and Fig. 2d shows the case when 0.675 µL SnSe NSs, 0.75 mg PPV, and 150 µL toluene were mixed. As in these cases, SnSe NSs aggregated even in mixed films, and a homogeneous film could not be obtained. The flat areas were probably films made only of conductive polymer. The large SnSe NSs aggregates protruded approximately 20 μm from the flat area. We thought that the high viscosity of the conductive polymers was the reason why SnSe NSs were not sufficiently dispersed by stirring, so we added toluene and acetone to these conductive polymers at various ratios other than those shown here to form a film. However, similar aggregation occurred and the unevenness of the film was not improved. It was concluded that SnSe NSs synthesized by the hot injection method is not suitable for use as a light absorption layer because it is difficult to form thin and uniform films due to large and irregular size.

We investigated the fabrication of light absorption layers using SnSe NSs synthesized by one-pot method. Figure 3a is an in-plane SEM image of a film formed by dropping and drying SnSe NSs dispersed in toluene. Compared to the SnSe NSs films produced by the hot injection method, the SnSe NSs films produced by the one-pot method had an very smooth surface, with a maximum height difference of about 1 μm. The size uniformity of SnSe NSs synthesized by the one-pot method may be responsible for producing relatively homogeneous films. In fact, unlike the SnSe NSs synthesized by the hot injection method, the thickness of the SnSe NSs synthesized by the one-pot method that completely covered the substrate surface was approximately 1.5 μm.

Plan-view SEM images of films spin-coated with a mixture of PEDOT: PSS and SnSe NSs synthesized by the one-pot method at varying volume fractions are shown in Fig. 3b–f. During mixing, 23 µL of dimethyl sulfoxide (DMSO) and 0.5 µL of tritonX-100 were added to 450 µL PEDOT: PSS. The volume fractions of SnSe NSs are 0.108, 1 0.121, 1.740, and 4.120 vol%, respectively. Although the entire substrates were completely covered with the mixed solution by spin coating, the NSs did not completely cover the films. As a result of evaluation while adjusting the SEM magnification (see Fig. 3f), the sheet coverage of SnSe NSs was found to be 2.2, 11.3, 1.1, and 1.7% in descending order of volume fraction, respectively. Note that these values are only a guide, as NSs may actually be present even though they appeared not to be present in the SEM field of view. The higher the volume fraction of the film, the more SnSe NSs aggregates were present. There were regions around the aggregates that appeared to have less SnSe NSs (inset in Fig. 3d). The formation of those aggregates is the reason why the sheet coverage did not increase as the volume fraction of SnSe NSs increased.

Cross-sectional SEM images of these samples are shown in Fig. 4. The SnSe mixed layer appears relatively white on the substrate surface in the figures. It is clearly shown that the lower the volume fraction of SnSe NSs, the flatter the surface and the fewer protrusions caused by aggregation. The height of most protrusions ranged from 200 nm to 1 μm, with some of the largest ones exceeding 3 μm. Based on their height and shape, these protrusions were determined to be either SnSe NSs not oriented parallel to the substrate surface or multiple SnSe NSs overlapping. The larger the volume fraction of SnSe NSs, the higher the number of large protrusions. This result suggests that the method of dispersing SnSe NSs into conductive polymers is still not sufficient.

The crystal structure and optical properties of SnSe NSs synthesized by both the one-pot method and the hot injection method were evaluated. XRD curve is shown in Fig. 5a. Characteristic diffraction peaks were observed at about 25, 30, 31, 38, and 49 degrees for both samples. Comparison with the pdf data sheet confirmed that these correspond to the (201), (111), (400), (311), and (511) planes of orthorhombic SnSe, respectively. The (400) peak was very strong in NSs synthesized by either method. This suggests that the sheets grew in the (001) direction. The very similar XRD curves indicated that the crystalline states of these two samples were almost the same. Figure 5b shows Tauc plots based on the measured optical absorption spectra of SnSe NSs assuming direct transition. Tauc plot analysis [39] uses the equation: \({\left(hv\alpha \right)}^{2}=k\left(hv-Eg\right)\), where h is a Plank’s constant, ν is a frequency, α is an absorption coefficient, k is a proportionality constant, and Eg is a direct bandgap energy. The absolute value of this equation depends on the proportionality constant k, which depends on the sample thickness, etc., but the bandgap energy is determined by drawing a tangent to the plot and finding the intersection with x-axis. As a result of analysis, the bandgap energy was found to be 1.25 eV for the sample synthesized by the one-pot method and 1.00 eV for that by the hot injection method. Both values are larger than the bulk literature value of 0.9 eV [18], suggesting that the bandgap of these NSs was determined by quantum size effects. The thickness of the synthesized SnSe NSs should be less than 23 nm, which is the Bohr radius of the excitons in SnSe [40]. The bandgap energy of the sample synthesized by the one-pot method was slightly higher than that by the hot injection method, suggesting that the film thickness of the former one was thinner. Figure 6 shows the band diagram of a solar cell using a sample synthesized by the one-pot method, estimated based on various literature values and this measurement result [26, 41, 42]. Since the literature value of the bandgap energy of SnSe NSs used as reference here is different from that obtained in this study, the diagram was created by distributing the energy difference equally between the conduction band and the valence band. This band diagram suggests that the absorbed carriers flow smoothly to the electrodes on both the conduction and valence band sides, unhindered by large barriers. As a result, this structure is a tandem solar cell that does not require tunnel junction layers [43]. The SnSe NSs synthesized in this study had a slightly wider bandgap energy than Si. Therefore, although SnSe NSs do not have the reported extremely wide bandgap energy of 2.9 eV [9], they act as a top layer relative to the bottom layer Si.

Solar cells were fabricated using SnSe NSs synthesized by the one-pot method. Figure 7 shows the current density–voltage (J–V) characteristics of the six types of solar cells fabricated. Here, ‘SnSe only’ refers to the cell that uses a film made by dropping and drying. The average thickness of the SnSe NSs mixed layer was 1.5 μm. For comparison, the characteristics of a solar cell without SnSe NSs layers are also shown (denoted as ‘no SnSe’. Cells were prototyped several times under each condition, and the best characteristics among them are shown here. Compared with the solar cell without SnSe NSs layer, the performance of the solar cells with SnSe NSs layer was all better. This suggests that the SnSe NSs layers functioned as the top layers of the tandem cell structure. Table 1 shows various characteristic parameters that can be read from these J–V curves. The case with a volume fraction of 0.121% recorded the highest PCE of 4.8%, and the fill factor (FF) was also good at 69%. The PCE is very high compared to that of previously reported SnSe solar cells using chemically synthesized SnSe nanocrystals. The table also shows the SnSe sheet coverage estimated from SEM observations. The coverage did not increase monotonically with increasing volume fraction. This is because when the volume fraction was high, SnSe NSs aggregated and did not spread uniformly in the conductive polymer film. The highest PCE was obtained for the cell with the highest SnSe sheet coverage at the volume fraction of 0.121%. When the volume fraction increased beyond this value, the series resistance (R_s) remained almost the same, but the parallel resistance (R_sh) decreased extremely and the sheet coverage also decreased. As is well known, R_sh is a resistance component due to leakage current of solar cells. Assuming that there is no significant difference in the insulation properties of the electrodes of the fabricated solar cells, it is reasonable to assume that the difference in R_sh was due to the difference in leakage current in the light absorption layer. The decrease in sheet coverage with increasing volume fraction indicates the formation of aggregates of SnSe NSs. The side reactions caused by the aggregates were probably the cause of the leakage current.

From the parameter changes in Table 1, we can infer what happened inside the solar cell. The theoretical formula for J–V characteristics based on the well-known single diode model [44] is where I is the output current, I₀ is the saturation current, q is the electron charge, V is the output voltage, n is the diode factor, k is the Boltzmann constant, T is the measurement temperature, and I_ph is the photocurrent. In Table 1, the short-circuit current density (J_SC) of the solar cell using the mixed layer increased as the volume fraction of SnSe NSs increased. As the volume fraction increased, the current generation ability itself should improve, so I_ph in Eq. (1) increases. In other words, if the light irradiation intensity is the same, I_ph should increase according to the volume fraction. For that reason, J_SC may have increased with the volume fraction in our experiments. During carrier extraction to the electrodes, most of the current flowed through the aggregate-free region, so there was no significant difference in R_s and the drop in open circuit voltage (V_OC) was small up to a volume fraction of 1.740%. However, if sheet coverage does not increase with volume fraction, I_ph will not increase in proportion to volume fraction. When the volume fraction was further increased to 4.120%, I_ph could no longer compensate for the increase in leakage current of the light absorption layer, and V_OC decreased accordingly. The increase in leakage current appeared as an extreme decrease in R_sh. Figure 8 shows how the SnSe sheet coverage and the volume fraction affected PCE and J_SC. Figure 8a clearly indicates that the PCE did not improve simply depending on the SnSe sheet coverage. Furthermore, it can be seen that PCE had a negative correlation with J_SC. The reason why this opposite phenomenon occurred is that V_OC and FF were high when J_SC was low. This is understood to be due to not only an increase in I_ph but also an increase in the leakage current. Figure 8b shows that J_SC improved almost monotonically as the volume fraction increased. This suggests that the the number of carriers generated changed depending on the amount of SnSe NSs mixed. However, PCE was negatively correlated with the volume fraction. This was directly due to the decrease in V_OC and FF and was probably related to the aggregation of SnSe NSs.

Table 1 also shows literature values for the characteristics of solar cells using SnSe NSs. Previous reports have shown that the PCE of solar cells using chemically synthesized SnSe NSs was less than 0.1%. Compared to them, the solar cells in this study showed very good characteristics. Our best PCE was slightly inferior to 6.44% for solar cells using thermally deposited SnSe films on Si substrates [32], but it is significant that almost equivalent value was obtained using films formed by simple coating methods.

Let us provide an overview of the techniques for achieving good properties in coated SnSe solar cells. This research has shown that it is possible to create SnSe solar cells using the coating method that are comparable to those produced using the vapor phase method. The SnSe NSs prepared by the one-pot method were small and uniform, making them suitable for mixing with conductive polymers. Even though the SnSe sheet coverage was only about 14%, relatively high PCE was obtained when mixed films with PEDOT: PSS were used. R_s of these solar cells was low. This suggests that PEDOT: PSS filled the gaps between SnSe NSs and smoothed carrier transfer. The higher the volume fraction of SnSe NSs in the mixed layer, the higher the J_SC, suggesting that the larger the amount of SnSe NSs, the more carriers were generated. What should be improved to achieve even higher PCE is the removal of obstacles for the generated carriers, i.e., the reduction of SnSe NSs aggregates. The ‘SnSe only’ solar cells had extremely high R_s. When using only SnSe NSs, each SnSe NSs is not necessarily in close contact with each other inside the films. It is believed that the large contact resistance between the SnSe NSs was the cause of the large R_s of the solar cell. As a result, even if the SnSe sheet coverage was 100%, only a low PCE was obtained. A similar phenomenon should occur in aggregates of SnSe NSs in mixed films with conductive polymers. The motion of the generated carriers will be hindered in the aggregation region, causing problems such as carrier accumulation, carrier scattering, and the occurrence of side reactions associated therewith. In this study, we obtained PCE that was nearly two orders of magnitude higher than that of previous SnSe-coated solar cells. The result clearly indicated the potential of chemically synthesized SnSe. However, the performance of the prototype cells was still not very stable. Characteristic variations were not large, but in rare cases there were some particularly bad ones. This may be due to the unstable mixing of SnSe NSs and conductive polymer. Solar cell properties will be further improved if high volume fraction and high sheet coverage of SnSe NSs could be achieved simultaneously by investigating additives to PEDOT: PSS and / or other conductive polymers as mixtures. In that case, stable production of cells will also be achieved. Additionally, it is important that SnSe NSs fabricated by the one-pot method can be made into wide-gap semiconductors depending on the synthesis conditions. Their use in the top layer of tandem solar cells will contribute to the realization of coated solar cells with very high PCE.

We investigated the method for realizing high-performance solar cells using coated layers of SnSe NSs obtained by chemical synthesis as light absorption layers. The SnSe NSs synthesized by the hot injection method were nonuniform in size, and the occurrence of their aggregation was inevitable. SnSe NSs synthesized by the one-pot method was small and uniform in size, and when mixed with conductive polymers, it was suitable for fabricating flat and homogeneous coatings with high SnSe sheet coverage. In SnSe solar cells in which the coating layers of the mixture were used as light absorption layers, very high PCE of 4.8% was obtained, which was nearly two orders of magnitude higher than that of previous solar cells using the similar chemically synthesized SnSe nanocrystals. The PCE was comparable to that of cells using thermally deposited SnSe films on Si substrates, which demonstrated the potential of chemically synthesized SnSe. Solar cell performance will be further improved by examining the mixing conditions of SnSe NSs with conductive polymers and increasing the volume fraction and sheet coverage. SnSe NSs, which do not contain toxic metals, are expected to be used not only as a single light absorption layer but also as a top layer of tandem solar cells, and has the potential to make an important contribution to the field of solar power generation in the future.