1 Introduction

Solar energy is one of the clean and renewable energy technologies available in the world. Compared with current main energy sources, i.e. fossil fuel such as crude oil, coal, and natural gas, it produces no harmful substances to the environment. Solar energy will never be depleted as long as the sun is available. Solar cells function by harvesting the sunlight, and converting it into electricity [1].

In 1954, the first generation of solar cell was made in Bell Laboratories by forming a diffused silicon p-n junction [2]. Silicon was chosen because of the availability and the popularity in semiconductor industry. However, these silicon-based solar cells are expensive to produce. The second generation of solar cells are known as thin film solar cells, which have similar working principle as the first generation solar cells. Thin film solar cells are aimed for mass production. However, deposition of thin film requires high temperature and high vacuum condition which cause high production cost of thin firm solar cells [3]. The combination of knowledge and experience acquired from the first and second generation of solar cells has become the foundation for the third generation solar cells to achieve lower production cost and high energy efficiency which can exceed the theoretical Shockley-Queisser limit of 33% [4, 5]. Solution-processed semiconductor solar cells such as dye-sensitized solar cells and quantum dot-sensitized solar cells (QDSSC) are catergorized as third generation solar cells [1].

Typically, QDSSC consists of photoelectrode, electrolyte, and counter electrode. The photoelectrode consists of conducting glass coated with wide band gap semiconductor, usually titanium dioxide (TiO₂) or zinc oxide (ZnO), with semiconductor quantum dots (QDs) attached on it. Commonly used material for counter electrode is platinum (Pt), but other materials like copper (I) sulfate (Cu₂S) is also used. QDs are nano-sized particles that are responsible for the light absorbance and generation of electron-hole pairs. The generated excited electrons will then be transfered to the wide band gap semiconductor and ultimately to the external circuit. QDs with small sizes can lead to quantum confinement effect, where the bandgap values of the QDs change accordingly. This effect is beneficial as the response to light can be tuned based on specific bandgap and QD’s size. Many semiconductor materials such as SnS, CdS, CdSe, and PbS have been identified for sensitizing solar cell application [6, 7].

Materials such as CdS, CdSe, and PbS have been commonly researched in the field of QDSSCs due to good optical properties. However, these materials show high toxicity that bring harms to the living creatures and environment, which are inconsistent with the main purpose of solar cells as a clean and green energy generator. On a bright side, few researches have demonstrated the application of SnS in QDSSCs. SnS is cheap and less toxic as compared with the materials mentioned above. SnS has been identified as a potential material for solar cells due to its low bulk bandgap of 1.09-1.30 eV and high conductivity and absorption coefficient. Besides, Sn and S are abundant on earth.

On the fabrication method of QDs, solution processed method like successive ionic layer adsorption and reaction (SILAR) is frequently employed. This method was first reported by Ristov et al. in 1985 [8]. SILAR method is basically an immersion of substrate in different cationic and anionic precursor solutions, and rinsing in deionized water for film growth purpose. The substrate must be rinsed after each dipping in precursor solution to washed away any unreacted ions. In general, SILAR method is an advanced version of chemical bath deposition (CBD) method. However, in CBD method, all the precursors are mixed together in one container, while in SILAR method, the precursors are separated. Thin film in SILAR method grows layer by layer, hence the film thickness can be controlled easily by varying the number of dipping cycle [9].

In this work, solution processed SnS QDSSCs were fabricated and investigated. Unlike previous reported studies, SnS was synthesized from SO₄^2-based salt as precursor, all in aqueous solution. The effect of concentration of precursor solution during the preparation of QDs was investigated. Besides, the number of dipping cycles on the efficiency QDSSC was also studied.

2 Materials and methods

3 Results and discussion

Three sets of QDSSC devices were fabricated from three different precursor concentrations (i.e. 0.05 M, 0.1 M and 0.5 M). These molarities were selected based on previous reported studies [11, 12]. As these studies did not have a thorough investigation on the precursor optimization, we selected a set of general acceptable molarities which started from low molarity. We confined the study to a few precursor concentrations and capped it at 0.5 M. In each set, the anionic and cationic precursor concentrations were maintained at the same concentration level. For example, devices fabricated from 0.05 M were prepared from 0.05 M of SnSO₄ and 0.05 M Na₂S acqueous solution. The fabricated SnS-sensitized TiO₂ electodes exhibited a dark brown color as shown in Figure 1. With higher precursor concentration, the sensitized layer became black. The same observation is also true for higher SILAR cycles. In this work, the SILAR cycles were limited to maximum 6 cycles as more dipping cycles would result in QDs overloading [13, 14].

Figure 1

SnS QD-sensitized electrode sample. All samples show good film adhesion to the substrate despite irregular surface area.

The prepared QD-sensitized TiO₂ electrodes were then characterized with SEM imaging. Figure 2 shows the surface morphology of the fabricated TiO₂ and TiO₂/SnS layer of the electrode samples. TiO₂ was successfully deposited on the substrate, as porous structure of TiO₂

Figure 2

SEM micrographs of photoanode surface at 40,000x magnification. (a) Bare TiO₂ layer; (b) TiO₂/SnS layer.

nanoparticles was observed. The TiO₂ particles were also “sensitized” (or adsorbed) with SnS layer, where reduced porosity of the TiO₂ layer was observed. The porosity of the TiO₂ layer was determined qualitatively via Image-J software. The presence of Sn and S elements was also verified by EDX analysis on the same surface analysis (see Table 1 for elemental analysis). In the preparation steps of the TiO₂ layer, it is crucial to ensure no cracks are visible as such cracks will cause high shunt resistance, R_SH and lower the efficiency of the device [15].

The performances of the fabricated SnS QDSSCs are shown in Figure 3. The J-V curves of the individual SnS QDSSCs prepared from 0.05 M, 0.1 M and 0.5 M precursor concentrations are shown in Figures 3a-c, respectively.

Figure 3

J-V curves for SnS QDSSC prepared from (a) 0.05 M precursor concentration; (b) 0.1 M precursor concentration; (c) 0.5 M precursor concentration with various dipping cycles. The efficiency trend is shown in (d).

Meanwhile, Figure 3d depicts the performance trends of the sample sets. On the other hand, the performance parameters of the SnS QDSSCs are tabulated in Table 2, with information pertaining to the concentrations of precursor solution and SILAR cycles.

From the result obtained, for samples prepared from 0.05 M precursor solutions, the optimal number of dipping cycles is 4 cycles where the device exhibits the highest open-circuit voltage (V_OC), short-circuit current density (J_SC), and efficiency at 0.43 V, 0.40 mA/cm², and 0.07%, respectively, within the same group. The overall performance is still low compared to other types of QDSSC like CdS and CdSe. This could be due to the intrinsic optical properties of SnS itself [10, 16, 17]. The low performance was also confirmed by other research groups although some were able to achieve better result due to optimization done on the photoanode layer [11, 12]. Further increasing the number of dipping cycles after 4 SILAR cycles does not show any significant improvement on the performance. In fact, the performance reduced dramatically with dipping cycles above 4 cycles. This could be attributed to the overloading of SnS QDs on the TiO₂ particles which subsequently increased the charge recombination rate at the TiO₂ interfaces [18]. At

this point, it can be assumed that saturation of QDs has been achieved. This observation is in-line with the work by Deepa et al. [19]. Also noted that the J-V curves have minor peak near the open circuit voltage, which indicates possible presence of pin hole in the sample.

Samples prepared from 0.1 M precursor concentration shows a similar pattern in which the best performing device is achieved when the sample was prepared with 4 SILAR cycles. Such sample has V_OC value of 0.34 V, J_SC value of 0.47 mA/cm², fill factor (FF) value of 34.3%, and efficiency of 0.055%. As for the samples prepared from 0.5 M precursor concentration, sample prepared with 3 SILAR cycles exhibit the best overall performance. It has V_OC of 0.31 V, J_SC of 0.25 mA/cm², and efficiency of 0.030%. All the samples exhibit a relatively low value of FF, which is less than 50%. This signifies a substantial charge recombination which needs to be tackled with. One of the remedies is to apply passivation layer [20].

On the other hand, samples prepared from higher precursor concentration have lower efficiency as higher precursor concentration leads to multiple nucleation sites for QD growth. Eventually, more QDs are formed until blocking the TiO₂ interface pathway. Thus, the thickening of SnS QD layers results in the impediment of electrons transfer across the interface. Meanwhile, V_OC was observed to increase with incremental of SILAR dipping cycles, up to optimum dipping cycle. The increased of V_OC is due to the increase of Fermi level of the photoelectrode with the presence of SnS QDs, allowing more injection of electrons from SnS QDs to the TiO₂ layer. Upon reaching the optimum parameter limits, J_SC and V_OC of the samples reduced which are due to the decrease in electrons recombination resistance in the devices led by overloading of SnS on the surface of TiO₂ layer. Overloading of SnS sensitizer also causes the reduction of V_OC and FF because it blocks the electrolyte penetration into the porous thin film [21]. One of the major contributing factors to the low performance of solar cell is the high charge recombination rate. Although impedance spectroscopy was not performed, judging from the performance parameters (as shown in Table 2), one can predict the low electrons recombination resistance within the interfaces of the samples. Such phenomenon can result in high charge transfer resistance R_ct with small capacitance Cμ, signifying slow charge transport at the interfaces. Ultimately, their electron lifetime values would be at the low side, demonstrating ineffective electrons transfer, which result in lower energy conversion efficiency [22, 23, 24, 25].

In the later stage, SnS QDSSC devices with the best parameter performances determined from I-V test were selected for UV-Vis absorption analysis in order to estimate the bandgap energy of the QD. Other samples were not tested for their absorption as the value from the best performing sample in each set is sufficient to provide the necessary information. Absorption spectra of TiO₂/SnS thin film with different concentration and SILAR cycles are shown in Figure 4a. The bandgap of samples are determined by extrapolating the Tauc plot to x-axis as shown in Figure 4b (for selected samples A4, B3 and C4).

Figure 4

(a) Absorption spectra of selected samples of TiO₂/SnS multilayer film; (b) Tauc plots of selected samples of TiO₂/SnS multilayer film.

Absorption spectra of TiO₂/SnS samples increased as the number of SILAR cycle increased. Also, the absorbance is higher for samples fabricated with higher precursor solution concentration. This is due to more SnS QDs formed on the TiO₂ layer which allow more light absorption. From Tauc plot analysis, samples with more SILAR cycles tend to have lower bandgap energy. The bandgap energy values for the selected samples are tabulated in Table 3. It is also observed that bandgap energy is also increased with the increasing deposition concentration except for samples prepared from 0.5 M precursor concentration.

The QD size was estimated from the absorption spectrum according to reported method [9]. The bulk bandgap energy for SnS is 1.3 eV, according to Deepa et al. [19]. The SnS QDs sizes increased with more SILAR cycles in each set of samples with the same concentration. For example, TiO₂/SnS samples prepared from 0.05 M precursor solution, the size of SnS QDs increased from 6.68 nm to 8.39 nm when the number of cycles increased from 4 to 6 cycles. This shows the quantum confinement effect as the dipping cycle and precursor concentration increased. By comparing among the different precursor concentration, there is an optimum QD size with its corresponding bandgap for the best device performance. The reduction of performance with higher QD size could be linked to the thickening

layer of the QDs which reduce the charge recombination resistance within the interface [26].

As the overall performance of the cells is still very low, plan is under way for optimization study. Although other semiconductor QD materials like CdS, CdSe and PbS have shown better performance, SnS being less toxic than the usual QD materials could have a better future application should more effort on optimization work is carried out. It must also be noted that the assembly cell samples in this work used generic setup in order to eliminate other positive contributing effect to the result, unlike other reported result in literature where optimized setup has been used. Some of the parameters that need to be focused on include dipping time and different precusor used. The deposition of blocking layer and passivation layer are also crucial for the better performance of the cells, which are not being applied in this preliminary study. Optimized electrolyte should also be looked into as different semiconductor QDs could perform well in different composition of polysulfide electrolyte [27, 28]. Lastly, suitable counter electrode materials should also be considered. Typically, Cu₂S counter electrode produces better result when used in QDSSC [29, 30, 31, 32].

4 Conclusion

Tin (II) sulfide QDSSCs have been successfully fabricated using SILAR method from aqueous solution. SEM and EDX analysis on TiO₂/SnS surface showed the presence of SnS nanoparticles attached to the TiO₂ particles. The devices performances were characterized using I-V tester and solar simulator. The overall efficiency of SnS QDSSCs increased when the concentration of precursor solution for QDs fabrication decreased from 0.5 M to 0.05 M. For a fixed precursor concentration, the efficiency of QDSSCs reduced once the optimal SILAR cycles were achieved. Further increasing the SILAR cycles did not increase the conversion efficiency of QDSSCs due to overloading of QDs where the aggregates would obstruct the electron transfer through the interface. The optimum SILAR cycles for 0.05 M and 0.1 M precursor concentration were 4 cycles, while for 0.5 M, optimum SILAR cycle was 3 cycles. The best performance was obtained from sample prepared from 0.05 M concentration, 4 SILAR cycles. Although the performance is relatively low, this work shows the feasibility of SnS QDs as alternative green material for QDSSC devices.