Background

Recently, many inorganic metal chalcogenides based on earth-abundant elements such as copper zinc tin selenide (CZTS), lead sulfide (PbS), copper (I) sulfide (Cu2S), tin sulfide (SnS), and antimony sulfide (Sb2S3) have been investigated as absorber materials in low-cost thin film solar cells in order to replace the mainstream solution-processible absorbers such as copper indium gallium selenide (CIGS) and cadmium telluride (CdTe) [1]. However, the use of CZTS and PbS in the industry has severe drawbacks, because CZTS uses the toxic and harmful hydrazine (N2H4) and requires the complex control of multi-compound [2] and PbS contains Pb, which is also toxic and hazardous. Other potential materials such as Cu2S and SnS have relatively low efficiencies compared to those of CIGS and CdTe. Sb2S3, however, has attracted attention as a candidate material due to its suitable band gap (~ 1.65 eV) and high absorption coefficient (> 105 cm−1) for efficient light absorption, high dielectric constant for exciton dissociation, and good band alignment with various hole transport layers (HTLs) for efficient charge carrier transfer, in addition to its cost effectiveness, low toxicity, and excellent air stability [3,4,5,6].

There are two types of Sb2S3 solar cells based on the device structures: sensitized solar cell or planar-type solar cell. Sensitized solar cells originated from dye-sensitized solar cells (DSSCs) and have a F-doped tin oxide (FTO)/compact TiO2 (c-TiO2)/mesoporous TiO2 (m-TiO2)/Sb2S3/HTL/Au structure, while planar-type solar cells have a FTO/c-TiO2/Sb2S3/HTL/Au structure [7].

In terms of device efficiency, sensitized Sb2S3 solar cells have a higher value than planar types due to their enhanced light-absorbing interfacial area owing to the m-TiO2 structure. The factor that decides the performance of sensitized solar cells is their interface quality inside the device where charge carrier separation and transfer occur. Therefore, significant effort has been devoted to the optimization of the interfacial properties, including those of the m-TiO2/Sb2S3 interface, Sb2S3/HTL interface, and HTL material itself [8]. Various kinds of HTL materials, such as 2,2′,7,7′-tetrakis[N,N-di(4-methoxyphenyl)amine]-9,9′-spirobifluorene (Spiro-OMeTAD) [9]; CuSCN, an inorganic p-type material [10]; poly(3-hexylthiophene) (P3HT), a conducting polymer [11]; and poly(2,6-(4,4-bis-(2-ethylhexyl)-4H-cyclopenta [2,1-b,3,4-b′]dithiophene)-alt-4,7(2,1,3-benzothiadiazole)) (PCPDTBT), a newly developed conjugated polymer [12], have been applied to adjust the Sb2S3/HTL interface and hole transport properties leading to a high fill factor (FF) and increased short-circuit current density (JSC).

Several studies that focus on the improvement of the m-TiO2/Sb2S3 interface properties have been also reported. Tsujimoto et al. modified the m-TiO2 surface using Mg2+, Ba2+, and Al3+, which effectively increase the power conversion efficiency (PCE) of all inorganic Sb2S3 solar cells that have the FTO/c-TiO2/m-TiO2/Sb2S3/CuSCN/Au structure [13]. Lan et al. used Li-doped m-TiO2 to enhance the electron transport properties and change the Fermi energy level [14]. Fukumoto et al. reported the surface treatment of the Sb2S3/HTL interface using 1-decylphosphonic acid (DPA), which can be attached to both an uncovered m-TiO2 surface and Sb2S3 surface to reduce recombination and increase the open-circuit voltage (VOC) and FF [15].

In planar-type solar cells, in contrast to sensitized ones, charge carrier transport depends on the carrier mobility and diffusion length within the Sb2S3 layer, which are strongly correlated with the morphology, grain size, and crystallinity of the layer. Hence, most research on planar-type solar cells has been focused on improving the Sb2S3 thin film quality to achieve a large grain size and a high crystallinity by using various deposition techniques. For example, conventional chemical bath deposition (CBD) [16], thermal evaporation (TE) [17], rapid thermal evaporation (RTE) [18, 19], atomic layer deposition (ALD) [20], and nanoparticle ink coating [21] have been applied to fabricate Sb2S3 thin films. Recently, Wang et al. reported a fast chemical approach (FCA) that can be used to generate very large grain sizes via a one-step spin-coating process and subsequent annealing process using a butyldithiocarbamic acid (BDCA)-based metal-organic precursor solution [22]. Many types of metal oxides or hydroxides can be dissolved in BDCA, which is relatively nontoxic, inexpensive, and thermally degradable, and can be easily synthesized via the reaction of 1-butylamine (CH3(CH2)3NH2) and carbon disulfide (CS2) [23].

Although the sensitized solar cells have a higher PCE (3–7.5%) than planar-type ones (2.5–5.8%), their device structure and fabrication process are complicated. Moreover, they contain a high degree of interface defects. A planar-type Sb2S3 device would have more potential for use in industrial-scale solar cells with a high efficiency and low cost, because it is conceptually simpler and easier to scale up and it is highly reproducible [24, 25].

Here, we report the surface treatment of a c-TiO2 layer using Cs2CO3 solution to enhance the performance of planar-type Sb2S3 solar cells. The Sb2S3 layer was deposited via a simple FCA spin-coating process to realize a large grain size, which was previously reported by Wang et al.

Cs2CO3 has been widely studied for application in organic photovoltaics (OPV) [26,27,28], organic light-emitting devices (OLEDs) [29], and perovskite solar cells (PSCs) [30, 31] to improve electron transport due to its low-work function property. Although Cs2CO3 is usually decomposed at 550–600 °C, Liao et al. reported that Cs2CO3 can be decomposed into low-work function cesium oxide via a low-temperature (150–170 °C) thermal annealing process [26]. However, to the best of our knowledge, there is no study on the application of Cs2CO3 to Sb2S3 solar cells.

Surface treatment using Cs2CO3 can not only reduce the energy barrier by changing the work function of c-TiO2, but also reduce the series resistance of the device by reducing the surface roughness of c-TiO2. The treatment resulted in improved device parameters such as the VOC, JSC, and FF, and the PCE increased from 2.83 to 3.97%. We believe that this surface treatment of c-TiO2 using Cs2CO3 solution can provide a simple and effective way of improving device performance in planar-type inorganic metal chalcogenide solar cells.

Methods/Experimental

Materials Used and Synthesis of Sb Complex

Antimony (III) oxide (Sb2O3, 99.99%), CS2 (> 99.9%), n-butylamine (CH3(CH2)3NH2, n-BA, 99.5%), cesium carbonate (Cs2CO3, 99.9%), 2-methoxyethanol (CH3OCH2CH2OH, 99.8%), titanium (IV) isopropoxide (Ti(OCH(CH3)2)4, TTIP, 97%), poly(3-hexylthiophene) (P3HT, Mw 50–70K, regioregularity 91–94%, Rieke Metals), 1,2-dichlorobenzene (o-DCB, 99%), and ethanol (CH3CH2OH, anhydrous) were purchased from Sigma-Aldrich Co. and were used as received without further purification.

The Sb complex was synthesized according to a reported method [22]. Sb2O3 (1.0 mmol) was mixed with a solution of ethanol (2.0 mL) and CS2 (1.5 mL) with magnetic stirring at room temperature. Then, n-butylamine (2.0 mL) was added to the solution slowly under continued stirring for at least 30 min to obtain a homogenous solution of antimony butyldithiocarbamates (Sb(S2CNHC4H9)3). Afterwards, 2 mL of this solution was diluted with 1 mL ethanol to form the Sb complex.

Device Fabrication

The planar-type Sb2S3 solar cells in this study have a typical structure of FTO/c-TiO2/Sb2S3/P3HT/Au, where P3HT is employed as the HTL. The c-TiO2 layer was deposited onto a cleaned FTO surface by spin-coating a mixed solution of 2 mL TTIP, 60 mL ethanol, 0.225 mL distilled water, and 0.03 mL HNO3 at 3000 rpm for 30 s, followed by annealing at 500 °C for 60 min in air.

For surface modification using Cs2CO3, Cs2CO3 dissolved in a CH3OCH2CH2OH solution with certain concentrations (1, 3, 5, and 10 mg/mL) was spin-coated on a 10-min UV-ozone treated c-TiO2 layer at 6000 rpm for 45 s. The films were then heat-treated at 150 °C for 10 min before the Sb2S3 layer was spin-coated.

For the Sb2S3 thin films, the Sb complex solution was spin-coated at a speed of 6000 rpm for 30 s, after which the films were annealed on a N2-purged hot plate at 200 °C for 1 min and 350 °C for 2 min.

P3HT solution (10 mg in 1 mL o-DCB) was spin-coated on the Sb2S3/c-TiO2/FTO substrate at a speed of 3000 rpm for 60 s, which was then heated on a hot plate at 100 °C for 30 min in air. Finally, the Au counter electrode was deposited using a thermal evaporator under a pressure of 5.0 × 10−6 Torr. Each device had an active area of 0.16 cm2.

Measurement and Analysis

The surface and cross-sections of the Sb2S3 thin films were characterized using field-emission scanning electron microscopy (FE-SEM, S-4800, Hitachi). The surface morphology was studied using atomic force microscopy (AFM, Park NX10, Park Systems). The optical properties of c-TiO2 were determined using a UV-Vis (Lambda 750, Perkin Elmer). The current density–voltage (J–V) characteristics were determined using a specialized solar cell measurement system equipped with an electrometer (model 2400, Keithley) and solar simulator (91192, Newport) with a 1-kW Xenon arc lamp (Oriel). The light intensity was adjusted to one sun (100 mW/cm2) under AM 1.5G solar irradiation conditions using a radiant power energy meter (model 70260, Oriel). The series resistance (RS) and shunt resistance (RSH) were calculated from the slope of the corresponding J–V curves beyond VOC and JSC, respectively. The external quantum efficiency (EQE) was measured by a QuantX-300 quantum efficiency measurement system (Newport) equipped with a 100 W Xenon lamp. The structural information of FTO/c-TiO2(/Cs2CO3) sample was characterized by multi-purpose X-ray diffraction (XRD) system (Empyrean, PANalytical) with θ-2θ mode at a scan rate of 0.05°/sec. The electronic state and energy level were analyzed using X-ray photoelectron spectroscopy (XPS) and ultraviolet photoelectron spectroscopy (UPS) in an ultrahigh vacuum environment (ESCALAB 250Xi, Thermo Scientific). UPS and XPS spectra were obtained by using the He I line (hν = 21.2 eV) and the Al Kα radiation source (hν = 1486.6 eV), respectively. The XPS depth profiling was obtained using Ar+-cluster ion gun and etch rate of 1 Å/sec.

Results and Discussion

Figure 1a shows a schematic of the device structure. The bottom layer is composed of c-TiO2 layers on a glass/FTO substrate acting as electron transporting. Light is absorbed by the Sb2S3 layer, while holes are transported by the P3HT HTL and collected at the Au counter electrode.

Fig. 1
figure 1

a Schematic of the device structure of planar-type Sb2S3 solar cells. b Sb2S3 thin film fabrication process using FCA method

Full size image

The Sb2S3 absorbing layer was deposited via the FCA using the Sb complex precursor to realize very large grain sizes. The precursor was thermally decomposed to the amorphous state at 200 °C for 1 min and crystalline state at 350 °C for 2 min (Fig. 1b). The SEM image shown in Fig. 2 indicates a very large grain size, which is almost the same as the Sb2S3 thin film morphology reported by Wang et al. [22].

Fig. 2
figure 2

a Top view and b cross-sectional SEM images of Sb2S3 absorbing layer after annealing at 350 °C for 2 min

Full size image

The efficiency of the planar-type Sb2S3 solar cell was improved via surface treatment with Cs2CO3 of the c-TiO2 layer.

The device properties based on the concentration of Cs2CO3 solution were performed to determine the optimum Cs2CO3 concentration. Figure 3a and Table 1 show the J–V characteristics for the devices using different concentrations of Cs2CO3 solution under AM 1.5G illumination (100 mW/cm2). When the concentration is too low (1 mg/mL), there is a problem in whole coverage of the c-TiO2 surface with Cs2CO3. However, if it is too high (5 and 10 mg/mL), it acts as a dielectric material, resulting in an increase in the series resistance and decrease in the device efficiency. The optimum concentration of Cs2CO3 was found to be 3 mg/mL. (Hereafter, “with Cs2CO3 treatment” means treatment using 3 mg/mL concentration of Cs2CO3 unless otherwise noted.)

Fig. 3
figure 3

a Current density–voltage (J–V) characteristics and b EQE spectra of planar-type Sb2S3 solar cells with and without Cs2CO3 treatment of c-TiO2

Full size image
Table 1 Summary of device performances according to different Cs2CO3 concentrations under AM 1.5G condition
Full size table

As a result, the device had a PCE of 2.83%, VOC of 0.549 V, JSC of 10.71 mA/cm2, and FF of 48.14% before the treatment. However, after the treatment with 3 mg/mL solution, all these parameters increased significantly, i.e., to a VOC of 0.596 V, JSC of 11.71 mA/cm2, and FF of 56.89%, leading to a PCE of 3.97%. This treatment resulted in a ~ 40% improvement in the PCE. The higher EQE over full spectrum range as shown in Fig. 3b indicates that the light is more efficiently converted into current leading to increase in JSC by this Cs2CO3 treatment. From the EQE spectra, we can also see that the onset of EQE at 750 nm corresponds well to a band gap of 1.65 eV for Sb2S3 layer and a decrease in EQE from 500 to 650 nm is attributed to the absorption of P3HT HTL layer.

We measured the XRD patterns of the c-TiO2 on FTO glass substrates with and without Cs2CO3 treatment to investigate whether Cs2CO3 has effects on the crystallization of the c-TiO2 layer and/or the formation of new secondary phase by diffused Cs-related species. There was no change in the XRD peak after Cs2CO3 treatment as shown in Fig. 4. This indicates that the Cs2CO3 treatment has little effect on the crystal structure of c-TiO2 and also does not create a new phase. Furthermore, there was no evidence of a decomposed Cs-related phase (cesium oxide, cesium suboxide, or Cs element) after thermal treatment of Cs2CO3, which means that the thickness of the Cs2CO3 is very thin. As shown in Fig. 5d, the thickness of Cs-related species was about 2~3 nm, which was determined by XPS depth profile analysis for the sample of FTO/c-TiO2/Cs2CO3 (3 mg/mL). The measured thickness of Cs2CO3 (2~3 nm) is in good agreement with the AFM analysis, which shows improved surface roughness through Cs2CO3 treatment from 9.89 to 8.03 nm (see Fig. 6a).

Fig. 4
figure 4

XRD patterns of the c-TiO2 on FTO glass substrates with and without Cs2CO3 treatment

Full size image
Fig. 5
figure 5

XPS spectra of a survey scan and Cs 3d peak, b Ti 2p peak, c O 1 s peak for c-TiO2 surface with and without Cs2CO3 treatment, and d depth profile for Cs 3d peak for FTO/c-TiO2/Cs2CO3 sample to determine the thickness of Cs-related layer

Full size image
Fig. 6
figure 6

a AFM images (2 μm × 2 μm) of the surface morphology and b UV-Vis absorption and transmission spectra of c-TiO2 with and without Cs2CO3 treatment

Full size image

We studied the surface state of the c-TiO2 layer using XPS measurements. The XPS spectra in Fig. 5 show that both the survey scan and Cs 3d peak scan clearly indicate the existence of Cs on the c-TiO2 surface. The Ti 2p and O 1 s peaks were shifted to lower binding energies owing to the Cs2CO3 treatment, which indicates that the Cs2CO3 treatment affected the electronic structure of the c-TiO2 layer. The appearance of a slight shoulder at ~ 531 eV in the O 1 s spectrum could be attributed to the cesium oxide generated from Cs2CO3 decomposition via annealing at 150 °C, which has a low work function [26].

The AFM images in Fig. 6a reveal a difference in the surface morphology of the c-TiO2 layer before and after Cs2CO3 treatment. The surface became smoother and the root mean square roughness (Rg) decreased from 9.89 to 8.03 nm after treatment. This smooth surface was useful for increasing the physical contact between the c-TiO2(/Cs2CO3) layer and the Sb2S3 layer, leading to a decrease in the RS value from 11.14 Ω cm2 (without Cs2CO3) to 8.82 Ω cm2 (with Cs2CO3) (see Table 1). The decreased RS may have contributed to increasing the FF from 48.14 to 56.89% [5].

The UV-Vis transmittance spectra of the c-TiO2 films with and without Cs2CO3 are shown in Fig. 6b. The figure shows that there is little change in the optical transmittance between wavelengths of 300 and 800 nm, which confirms that Cs2CO3 treatment has a negligible effect on the intensity of light reaching the Sb2S3 layer.

UPS was used to determine the change in the work function of the c-TiO2 layer before and after Cs2CO3 treatment to investigate the effect of Cs2CO3 on VOC. The results are shown in Fig. 7a. The work function of c-TiO2 decreases by 0.3 eV after Cs2CO3 treatment. Cs2CO3 is widely used as an efficient electron transport material in many optoelectronic devices through thermal evaporation or solution process. However, the accurate analysis of electron transport mechanism and the type of decomposed Cs-related species that are responsible for electron transport property are still uncertain and controversial. Among previous reports on solution-processed Cs2CO3, Liao et al. showed that Cs2CO3 can be decomposed into low work function, doped semiconductor in the form of Cs2O doped with Cs2O2 after thermal annealing at 150 °C by using XPS analysis [26]. This form of doped cesium oxide can act as an n-type semiconductor with intrinsically low work function, which might contribute to work function reduction of c-TiO2 in our system. In addition, there was no change in the absorption onset as shown in Fig. 6b, indicating little change in the optical bandgap of the c-TiO2 after the treatment.

Fig. 7
figure 7

a UPS spectra of c-TiO2, b energy level diagram, and c proposed operating principle of planar-type Sb2S3 solar cells with and without Cs2CO3 treatment

Full size image

The energy band diagram in Fig. 7b shows that the conduction band energy level of c-TiO2 shifted toward a lower energy by 0.3 eV. This shift leads to not only an improved VOC due to an increase in the built-in potential (VBI) inside the devices, but also an increased JSC due to the alignment of the energy level between c-TiO2 and Sb2S3 to reduce the charge transport barrier at the interface. The proposed operating principle is illustrated in Fig. 7c. At open-circuit condition, the shifted conduction band of the c-TiO2 layer by Cs2CO3 treatment leads to the increased VBI, which contributes to the improved VOC. At the same time, the increased VBI results in the larger energy band bending of the Sb2S3 layer under short-circuit conditions, and thus the photogenerated electrons can move quickly toward the c-TiO2 layer. This fast electron transport is attributed to cause the enhanced JSC and FF. Thus, the Cs2CO3 treatment on c-TiO2 layer could increase both VOC and JSC simultaneously, leading to the enhanced PCE. Hence, Cs2CO3 is a promising material for c-TiO2 surface modification as it enhances device performance by changing the work function and improving the electron transport properties.

Conclusions

Cs2CO3 was found to be an effective surface modifier to enhance the charge transport ability of the c-TiO2 electron transport layer (ETL) for planar-type Sb2S3 solar cells. The UPS data show that Cs2CO3 treatment can shift the work function of c-TiO2 upward, possibly increasing the built-in potential of the device and reducing the energy barrier for charge transport. The c-TiO2 surface became smoother after Cs2CO3 treatment, resulting in increased physical contact with the Sb2S3 absorber. The solar cell performance was significantly improved in all parameters simultaneously including VOC, JSC, and FF. This resulted in an increase in the PCE from 2.83 to 3.97%, almost a 40% increase. This study shows that surface treatment using inorganic compounds such as Cs2CO3 will play an important role in the development of highly efficient planar-type Sb2S3 solar cells.