Enhanced Electrical Conductivity of Sb2S3 Thin Film via C60 Modification and Improvement in Solar Cell Efficiency.

Sb2S3 has attracted great research interest very recently as a promising absorber material for photoelectric and photovoltaic devices because of its unique optical and electrical properties and single, stable phase. However, the intrinsic high resistivity property of Sb2S3 material is one of the major factors restricting the further improvement of its application. In this work, the C60 modification of Sb2S3 thin films is investigated. The conductivity of Sb2S3 thin films increases from 4.71 × 10-9 S cm-1 for unmodified condition to 2.86 × 10-8 S cm-1 for modified thin films. Thin-film solar cells in the configuration of glass/(SnO2:F) FTO/TiO2/Sb2S3(C60)/Spiro-OMeTAD/Au are fabricated, and the conversion efficiency is increased from 1.10% to 1.74%.

Introduction

In recent years, a great deal of effort has been devoted to explore the application of novel materials such as Cu2ZnSn(S,Se)4,1, 2 SnS,3 CuSb(S,Se)2,4, 5 GeSe,6 Sb2(S,Se)3 7, 8, 9, 10, 11 for light absorption materials for solar energy conversion. Among them, antimony sulfide (Sb2S3) is a binary semiconductor compound with a single stable phase, which could avoid the formation of other secondary phases.12 In particular, Sb2S3 received considerable attention for light absorber material in solar cells due to its suitable bandgap of 1.7–1.8 eV with high absorption coefficient of 105 cm−1 in the visible range, abundant and environmental friendly compositional elements, and excellent air stability.13, 14, 15 This bandgap allows its application as absorber in single‐junction solar cell or in top cell of the multijunction tandem photovoltaic device.

The carrier diffusion length of Sb2S3 could reach hundreds of nanometers,14 which allowed its application in both sensitized mesoporous device structure and planar heterojunction solar cell configuration. Sb2S3 was employed as semiconductor sensitizer in mescoscopic solar cells in 2009, where the extremely thin Sb2S3 absorber (between several nanometers and several tens of nanometers) was made using chemical bath deposition (CBD), and a conversion efficiency of 3.37% was obtained.16 Choi et al. reported conversion efficiency as high as 7.5% in 2014, by post‐treatment for reducing the trap sites in Sb2S3.17

On the other hand, Sb2S3‐based planar heterojunction solar cells also gained great progress in device fabrication, device structure improvement, and so on. Compared to the several nanometers‐thick absorber in Sb2S3‐sensitized solar cells, the thickness of Sb2S3 absorber in planar heterojunction solar cell was in scale of hundreds of nanometers. Both the surface defect and bulk defect played critical roles in dictating solar cell device performance. The intrinsic high resistivity property of Sb2S3 material was one of the major factors restricting the further improvement of device performance.18 Introducing other elements as dopant to modify the host system or forming ternary alloyed compounds was an efficient route to tailor the electronic band dispersion, adjust the carrier concentration, or even vary the conductivity type of the Sb2S3‐based semiconductor. Very recently, Tang et al. reported controllable sulfur vacancy defect by introducing Zn ion into the films. The behavior of electron concentration in Sb2S3 layer was observed to increase with sulfur vacancy defects, which resulted in reduced series resistance and increased recombination resistance for the Sb2S3 thin film solar cells.19

However, the incorporation of metal or non‐metal element doping or formation of alloy with other material, usually presented in the final product in the form of ion, led to the distortion of lattice structure and induced parasitic effect to the Sb2S3. For example, the bismuth dopant decreased the surface state of Sb2S3 crystals and caused the variation of band dispersion and electronic structure of Sb2S3.20 Carbon was also employed as dopant in the Sb2S3 thin layers due to its very high solubility and very low diffusion coefficient. Cardenas et al. investigated the effect of surface modification of arc‐deposited carbon layer on Sb2S3 thin films. The resistivity of Sb2S3 thin films was reduced from 108 Ω cm for as‐prepared condition to 102 Ω cm for post‐treated thin films, while the bandgap for carbon‐doped sample was kept nearly values with the undoped crystalline Sb2S3 thin films.18 The investigation of the carbon doping on Sb2S3 was not enough.

In this work, we explore the C60‐modified Sb2S3 thin films by spin coating. The presence of C60 was checked by scanning electron microscopy (SEM) and Raman spectra. Both of the longitudinal and transverse conductivity of the Sb2S3 layer with and without C60 modification was analyzed. Planar solar cells in configuration of glass/SnO2:F (FTO)/TiO2/Sb2S3(C60)/Spiro:OMeTAD/Au were fabricated. The solar cells employing C60‐modified Sb2S3 absorber show better short‐circuit current, fill factor (FF), and conversion efficiency. Furthermore, the physical mechanism behind this improvement was discussed.

Results and Discussion

The surface morphologies of Sb2S3 thin films with and without C60 modification were characterized by SEM. As shown in Figure , the Sb2S3 layer was smooth and uniform without any pinhole or particle agglomeration that could be observed, which was in agreement with the previous reports. A large number of small bright dots were observed from the SEM image of the C60‐modified Sb2S3 thin films. In order to characterize the small bright spots on the film that were caused by C60 modification, we carried out the Raman spectra measurement in the wavenumber range of 150–600 and 1300–1700 cm−1, respectively. As show in Figure a, in the range between 150 and 600 cm−1, both of the Raman spectra exhibited similar behaviors, and four peaks at 188, 237, 280, and 303 cm−1 could be observed. The peaks centered at 188, 237, 280, and 303 cm−1 could correspond to the vibration of the Sb—Sb bond for Sb2S3 structural units, to the antisymmetric S—Sb—S vibration, the antisymmetric vibrations of Sb—S stretching in pyramidal symmetry, and the symmetric stretching of the Sb—S structural units, respectively.21 On the contrary, the Raman spectra in the range of 1300–1700 cm−1 displayed very different behaviors (Figure 2b). The Raman spectrum of bare Sb2S3 sample was smooth and no obvious peaks could be observed, while that of C60‐modified sample showed peaks at 1609, 1668, 1565, and 1434 cm−1, respectively. The peak centered at 1609 and 1668 cm−1 was corresponding to the vibration of Hg symmetry in the C60 unit.22, 23 The peaks centered at 1434 and 1565 cm−1 could be indexed to the Ag and Hg vibration mode of C60.24 This result suggested the presence of C60 in the Sb2S3 films. It should be noted that the X‐ray diffraction (XRD) patterns of bare and C60‐modified Sb2S3 thin films showed nearly the same behavior with each other. As shown in Figure 2c, both the XRD patterns of the bare and C60‐modified Sb2S3 films could be indexed to the orthorhombic structure of Sb2S3 (JCPDS: 06‐0474), and no other impurity phase or any shift could be observed. This behavior was much different from the case of metal or non‐metal element doping in the Sb2S3 thin films.19, 25 The lattice parameters a, b, and c of the bare Sb2S3 film were calculated to be 4.722, 5.695, and 15.681 Å, respectively. On the contrary, the parameters for the C60‐modified Sb2S3 film were 4.725, 5.691, and 15.673 Å, respectively. In addition, the transmittance spectra of the C60‐modified Sb2S3 film was basically the same as that of bare Sb2S3 film (Figure 2d). These results hinted that the presence of C60 did not induce any lattice distortions for Sb2S3 thin films. In addition, the optical bandgap of the unmodified Sb2S3 and C60‐modified Sb2S3 films was calculated to be about 1.76 eV, obtained from the optical transmittance spectra, which was very close to the values obtained by external quantum efficiency (EQE) curves (Figure b).

Figure 1

Plan‐view SEM images of the Sb2S3 thin films, a) bare Sb2S3 and b) C60‐modified Sb2S3 thin films.

Figure 2

Raman spectra of Sb2S3 for the wavenumber range of 150–600 and 1300–1700 cm−1. a) Raman spectra of Sb2S3 and C60‐modified Sb2S3 in the wavenumber range of 150–600 cm−1. b) Raman spectra of Sb2S3 and C60‐modified Sb2S3 in the wavenumber range of 1300–1700 cm−1. c) XRD patterns of the Sb2S3 thin films with and without C60 modification. d) Transmittance spectra of the Sb2S3 thin films with and without C60 modification.

Figure 3

a) J–V curves of Sb2S3 thin film solar cells with and without C60 modification. b) EQE spectra of Sb2S3 thin film solar cells with and without C60. c) The ratio of EQE (−0.2 V)/EQE (0 V).

We further characterized the effect of C60 modification on the photoconductivity of Sb2S3 thin film. Both the transverse and longitudinal optical flow responses of the bare and C60‐modified Sb2S3 thin films were measured in a certain time domain (Figure ). In the case of transverse photocurrent response measurement, a Sb2S3 film with or without C60 was spin‐coated directly onto the FTO glass substrate as the photon‐absorber layer, followed by the thermal evaporation of a 50 nm thick gold electrode on the Sb2S3 film. The device structure of glass/Sb2S3(C60)/Au is formed as shown in Figure 4a.26 Figure 4b exhibited the light and dark transverse photocurrent responses of the bare and C60‐modified Sb2S3 thin films as functions of time, where the photo current was generated with irradiation of a white LED with a power density of 5 mW cm−2. The photocurrent of the bare and C60‐modified Sb2S3 thin films exhibited good repeatability. The photocurrent and dark current of bare Sb2S3 thin film was 4.53 × 10−5 and 3.18 × 10−6 mA, respectively. The transverse electrical conductivity of the bare Sb2S3 thin film was low (4.71 × 10−9 and 3.43 × 10−10 S cm−1) and consistent with the values in previous references. In addition, the C60‐modified Sb2S3 thin film and bare Sb2S3 thin film were subjected to current density–voltage (J–V) test under light illumination, as shown in Figure 4c. However, the C60‐modified Sb2S3 thin film showed higher conductivity (2.86 × 10−8 and 2.24 × 10−8 S cm−1) both under light illumination and dark conditions. This was also confirmed by the J–V curves in Figure 4c.

Figure 4

a) Schematic diagram of the transverse conductivity measurement of the Sb2S3 thin films. b) Transverse photocurrent response of Sb2S3 thin films on glass substrate with and without C60 modification. c) Transverse I–V curve of the Sb2S3 thin films with and without C60 modification. d) Schematic diagram of the longitudinal structure of the Sb2S3 thin film solar cell. e) Longitudinal photocurrent response of Sb2S3 thin film solar cells with and without C60 modification.

For longitudinal photoconductivity measurement, we fabricated the Sb2S3‐based devices in configuration of glass/FTO/TiO2/Sb2S3(C60)/Spiro‐OMeTAD/Au. In the device shown in Figure 4d, the light passed through glass, FTO, and TiO2 layer, and was absorbed in the Sb2S3 layer, resulting in the generation of carriers. The photon‐generated carriers were extracted by TiO2 and Spiro‐OMeTAD layers and collected by FTO and gold contact. Figure 4e displays the longitudinal photocurrent responses of the device with bare or C60‐modified Sb2S3 films as a function of time. It was observed that both samples exhibited good repeatability and stability as the transverse photocurrent response spectra. Both dark currents were low but the photocurrent was much different. The photocurrent of the bare Sb2S3 sample was around 40 µA, while that of the C60‐modified Sb2S3 sample was higher than 70 µA under light condition. This result suggested that both the transverse and longitudinal photoconductivity of the Sb2S3 thin films could be improved by C60 modification.26, 27

The photovoltaic characteristics of the solar cells, in superstrate configuration, with bare Sb2S3 and C60‐modified Sb2S3 absorber layers, respectively, were measured under AM 1.5G illumination. The J–V curves of the devices are shown in Figure 3a. The solar cell with bare Sb2S3 absorber exhibited an open‐circuit voltage (V OC) of 0.465 V, short current density (J SC) of 7.03 mA cm−2, FF of 33.54%, and a power conversion of 1.10%, respectively. The cell with C60‐modified Sb2S3 absorber showed a power conversion of 1.75% with a V oc of 0.492 V, J sc of 8.44 mA cm−2, and FF of 42.15%. As shown in Figure 3b, the edge of EQE spectra of the Sb2S3 and C60‐modified Sb2S3 solar cells were about 750 nm, but the plateau region was different from each other. The EQE spectra of the Sb2S3 solar cells sharply decreased at about 400 nm, which could be ascribed to the incomplete collection of photogenerated carriers in the solar cell. In the contrast, the plateau region of EQE for C60‐modified Sb2S3 solar cells was beyond 400 nm, indicating that the presence of C60 can effectively enhance the collection length of photogenerated carriers in the device. Furthermore, the biased EQE measurement in Figure 3c exhibited higher EQE values under −0.2 V bias voltage, suggesting an enhanced carrier collection.28, 29

To further explore the charge transport characteristics of the device, we performed the electrochemical impedance spectroscopy (EIS) measurement. The impedance spectra of the devices were recorded at a potential of 0 V at frequencies ranging from 1 Hz to 0.1 MHz. The Nyquist plots are shown in Figure , and the inset is the equivalent electrical circuit model for the Sb2S3‐based thin film solar cell. The equivalent circuit diagram consisted of a series connection of a resistor (R1), a parallel combination of a constant phase element (CPE1), and a resistor (R2).30, 31, 32 In this experiment, the resistor R1 was related to the internal series resistance of the device, and R1 was the starting point of the real part in the Nyquist plots. R2 and CPE1 were associated with the interface of the absorber layer and the TiO2 layer. CPE1 can be represented by a capacitor (CPE1‐T) and a nonhomogeneity constant (CPE1‐P). The fitting results for these parameters are shown in Table . The value of R1 for the solar cell with bare Sb2S3 absorber was 33.23 Ω cm2, while that value for the device with C60‐modified Sb2S3 absorber decreased to 9.22 Ω cm2. Moreover, the values for the composite resistors R2 were raised from untreated 1358 to 23680 Ω cm2 after modification. The results indicated that the modification of C60 on Sb2S3 facilitated charge transport from absorber to the TiO2 layer with less recombination.33, 34, 35

Figure 5

EIS of the device based on Sb2S3 and Sb2S3‐C60 absorber layers. The solid lines are fitted results, and the inset is the equivalent circuit.

Table 1

Fitted parameters of the equivalent circuit from the EIS measurements

Absorber	R1 [Ω cm2]	CPE1‐T [10−8 F cm−2]	CPE1‐P [F cm−2]	R2 [Ω cm2]
Sb2S3	33.23	7.13	0.85	1358
Sb2S3‐C60	9.22	6.49	0.94	23680

Conclusion

In summary, the addition of C60 into the solution‐processed Sb2S3 thin films enhanced the transverse and longitudinal conductivity of the Sb2S3 films. The crystallinity and optical properties exhibited nearly the same behavior for the Sb2S3 films with and without C60 modification. The presence of C60 can effectively improve the thin film conductivity and facilitate the collection of photogenerated carriers in the absorber layer. Thus, the Sb2S3‐based planar heterojunction solar cells exhibited higher conversion efficiency after C60 modification. EIS measurement showed that the C60‐modified Sb2S3 thin film solar cell had a smaller series resistance and less interface recombination.

Experimental Section

The Sb2S3 thin film was deposited by spin coating on a commercial FTO glass substrate covered with a 50 nm thick TiO2 layer. The TiO2 layer was deposited by spin coating at 5000 rpm for 30 s and annealing in a muffle furnace at 500 °C for 60 min. The Sb2S3 precursor solution was prepared as follows: 1.0 mmol of Sb2O3 powder was dissolved in a mixed solution of 2.0 mL of anhydrous ethanol (CH3CH2OH, AR) and 1.5 mL of carbon disulfide (CS2), and then 2.0 mL of n‐butylamine (CH3(CH2)3NH2, GR 99.5%) was slowly added and stirred well. After that, every 2 mL of the above solution was diluted with 1 mL of anhydrous ethanol. Similarly, the C60 solution was prepared as follows: 5 mg of C60 powder was loaded into a 10 mL vial containing 2.0 mL of anhydrous ethanol (CH3CH2OH, AR) and 1.5 mL of carbon disulfide (CS2, GC), and then 2.0 mL of n‐butylamine (CH3(CH2)3NH2, GR 99.5%) was slowly added to the vial and stirred at room temperature. After stirring well, 2.16 mL of C60 solution was added to 1 mL of Sb2S3 precursor solution and stirred. Three hundred nanometers of Sb2S3 layers with and without C60 modification were deposited on commercial FTO glass, coated with a TiO2 layer, by spin coating at 3000 rpm for 60 s. Following the deposition, thin film was annealed on a hot plate in N2‐purge glove box at 100 °C for 1 min and 300 °C for 2 min. Sixty nanometers of Spiro‐OMeTAD was spin‐coated on Sb2S3 (C60)/TiO2/FTO substrate for 30 s at 3000 rpm as the hole transport layer. Solar cells were manufactured using the glass/FTO/TiO2/Sb2S3 (C60)/Spiro‐OMeTAD/Au structure. Gold contacts of 70 nm thickness and 0.09 cm2 area were deposited by thermal evaporation.

The surface morphology of bare and C60‐modified Sb2S3 thin films was characterized by SEM (FEI nova nano SEM450). XRD patterns of the films were characterized by XRD with Cu Kα1 (1.54056 Å) radiation (Bruker D8 Advance), and the optical properties of the films were measured using a spectrophotometer equipped with an integrating sphere (Perkin‐Elmer Lambda 950). In order to characterize the presence of C60 in the C60‐modified film, Raman measurements were performed using Raman spectrometer (Horiba Jobin Yvon, HR Evolution) equipped with the 532 nm line of the laser. Under standard test conditions (25 °C, AM 1.5, 100 mW cm−2), the J–V measurements were performed on bare and C60‐modified Sb2S3 thin film solar cells using an AM 1.5 solar simulator equipped with a 300 W xenon lamp (Model No. XES‐100S1, SAN‐EI, Japan). EQE of the bare and C60‐modified Sb2S3 film was obtained by Enlitech QER3011 system. Carrier transporting behavior of the device was characterized by EIS (PP211, ZAHNER, Germany) under appropriate open‐circuit voltage.

Conflict of Interest

The authors declare no conflict of interest.