Investigation of efficient protocols for the construction of solution-processed antimony sulphide solid-state solar cells

Abstract

Antimony sulfide solar cells have been studied with the purpose to investigate an easy and efficient procedure of antimony sulfide deposition on mesoporous titania films. Two principal deposition methods have been studied based on different sulfur precursors (thiosulfate vs. thiourea) and chemical bath vs. spin-coating deposition. The two approaches led to the same practical effect thus highlighting spin-coating deposition using antimony chloride and thiourea as the simplest and most practical method of construction of antimony sulfide/titania photoanodes. Poly(3-hexylthiophene) was employed as a hole-transporter while PEDOT:PSS and Ag were used as the counter electrode. They were all deposited under ambient conditions. XPS measurements provided information about antimony sulfide stoichiometry while UPS measurements gave the ionization potential in each case. Both deposition methods led to sulfur-deficient antimony sulfide.

---

Introduction

Hybrid organic–inorganic solid state solar cells (HSCs) are usually made by successive deposition of a thin layer of an oxide semiconductor, a layer of a semiconductor sensitizer and a layer of a Hole Transporting Material (HTM) and are completed by depositing a counter electrode on the top. Such cells have recently enjoyed immense popularity thanks to the employment of organometal halide perovskites as sensitizers or as functional photoanodes.^1,2 This popularity is justified by the easiness of perovskite cell construction, the large spectral range of light absorption, the high current densities and large open circuit voltage they produce and the resulting high conversion efficiency presently mounting to 20%.¹ However, there is a serious disadvantage associated with perovskite solar cells, in particular their sensitivity to humidity and to several organic agents, in the presence of which organometal halide perovskites are easily decomposed.² For this reason, other more stable functional materials are also studied with a lot of interest. One such case is Sb₂S₃, which is the subject of the present study and which has been successfully employed as sensitizer of nanoparticulate titania^3–9 and more generally as component of solid-state solar cells.¹⁰ Solar cells based on Sb₂S₃/TiO₂ photoanodes absorb photons in a large spectral range up to 750 nm with high extinction coefficient (≈1.8 × 10⁵ cm⁻¹ at λ = 450 nm)⁷ and thus they can produce and do produce high current densities.^3,7,9 In addition, the energy levels of the components involved in the construction of antimony sulfide solar cells offer a perfect match justifying their combination, as can be seen in the diagram of Fig. 1. Compared to perovskites, antimony sulfide solar cells suffer of much lower open circuit voltage V_oc. Indeed, V_oc in the former case ranges up to 1.2 V (ref. 11) while in the latter case it is less than 0.6 V.⁹ V_oc is theoretically defined by the difference between the potential of the final electron injection level, i.e. the conduction band of titania, which is approximately located at −4.2 eV, and the final hole injection level, i.e. the HOMO level of the HTM (P3HT in Fig. 1), which is approximately located at −5.2 eV. Therefore the maximum expected V_oc value should be no more than 1 V. As already said, it is in fact much lower due to inevitable losses.^3,9

Fig. 1 Energy states corresponding to the components of a typical antimony sulfide solar cell.^3,6

A major concern in the study of antimony sulfide solar cells is the quality of the Sb₂S₃ synthesis product. First, if such cells can hold the ambition of making it to the market, antimony sulfide should be deposited by an easy procedure under ambient conditions. The standard method of Sb₂S₃ formation in mesoporous titania is aqueous chemical bath deposition at low temperature (≤10 °C) using SbCl₃ and sodium thiosulfate as precursors.^3,12–14 Formation of Sb₂S₃ in an aqueous environment is described by the following reactions:⁷

2SbCl₃ + 3Na₂S₂O₃ → Sb₂(S₂O₃)₃ + 6NaCl

Sb₂(S₂O₃)₃ + 6H₂O → Sb₂S₃ + 3HSO₄⁻ + 3H₃O⁺

Cooling is necessary in order to slow down these reactions, avoid precipitation and allow adsorption in the mesoporous oxide film. Other researchers have followed the opposite route by growing Sb₂S₃ in mixed water–alcohol environment at high temperature in an autoclave.⁴ A very interesting approach was recently adopted by Choi and Seok⁹ who deposited Sb₂S₃ on mesoporous titania films by a very simple room temperature method involving spin-coating of a concentrated solution of SbCl₃ and thiourea in dimethylformamide (DMF) followed by annealing in inert atmosphere. The results were impressive since the efficiency of the obtained solar cells was comparable to champion reports^9,15,16 at the same time offering simplicity and practicality of cell construction procedures. This room temperature synthesis of antimony sulfide proceeds through the formation of a complex with thiourea (TU)⁹

SbCl₃ + 2TU → [Sb(TU)₂]Cl₃

and the subsequent decomposition of the complex during annealing. The quality of the obtained antimony sulfide by any of the above methods is mainly characterized by the crystallinity of Sb₂S₃ nanoparticles and the excess or the deficit of sulfur. Too much sulfur decreases conductivity of the obtained sensitizer. Too little sulfur introduces traps, which become sites of electron–hole recombination.

In the present work, we have employed and compared the above two main procedures of Sb₂S₃ synthesis in mesoporous titania and have systematically characterized the obtained materials in order to appreciate under similar conditions the capacity of each procedure to produce efficient photoanodes.

Results and discussion

Characterization of the Sb₂S₃/TiO₂/FTO films

Photoanode electrodes made by depositing Sb₂S₃/TiO₂ films on FTO were first characterized by XRD. The results are shown in Fig. 2. The major peaks derive from FTO and TiO₂ but in the presence of antimony sulfide, its chief crystalline species, stibnite, was also clearly detected. Crystallinity seems more pronounced in the case of the chemical-bath-deposited species while in the case of the spin-coated species an area of amorphous phase systematically showed up between 10 and 20°. Fig. 2 then reveals that both procedures lead to, at least partly, the formation of stibnite, i.e. crystalline Sb₂S₃, best defined in the case of chemical bath deposition. It must be underlined at this point that the stibnite phase was the result of careful annealing under certified inert atmosphere. On the contrary annealing in air, as it has been found in our previous publication,³ led to the formation of antimony oxide and a secondary crystalline phase of senarmontite.

Fig. 2 XRD analysis of Sb₂S₃/TiO₂/FTO films. The small circles indicate the presence of stibnite, i.e. the chief crystalline form of antimony sulfide.

The composition of the antimony sulfide films in each case was studied by XPS. The survey spectra (not shown) of the samples showed the presence of the following elements: Sb, S, Ti, O and C. Ti2p_3/2, which was at a binding energy 459.0 eV was assigned to the Ti⁴⁺ chemical state. Fig. 3 shows the XP spectra of S2p for (1) spin coating and (2) chemical bath samples. The S2p spectra of both samples consist of one main peak at binding energy 161.8 eV assigned to sulphur in Sb₂S₃ bonds whereas a low intensity peak at ∼168 eV reveals the presence of small amount of S–O bonds. Fig. 4 shows the XP spectra of Sb3d_3/2. The peak is analyzed into two equal width components differing by 0.8 eV and assigned to Sb₂S₃ and to Sb₂O₃ chemical states. The data extracted from Fig. 3 and 4 are listed in Table 1. The atomic ratio of Sb : S was calculated with 10% experimental error. It has to be mentioned that the Sb3d_3/2 was chosen for analysis because of the overlapping of Sb3d_5/2 with O1s peak.

Fig. 3 Core level XPS-S2p spectra of (1) spin-coated and (2) chemical-bath-deposited sample.

Fig. 4 Deconvolution of XPS-Sb3d_3/2 core level spectra of (1) spin-coated and (2) chemical-bath-deposited sample.

Table 1 Composition of Sb₂S₃/TiO₂ films studied by XPS and values of their ionization potential obtained by UPS

Method of film deposition	Elements detected	Chemical state	Atomic ratio Sb : S	Ionization potential (eV)
Chemical bath	Ti, O, C, Sb, S	Ti^4+(TiO2), Sb^5+, S–Sb, S–O	1 : 1.10	5.8
Spin-coating	Ti, O, C, Sb, S	Ti^4+(TiO2), Sb^5+, S–Sb, S–O	1 : 0.78	5.6

---

From these results it is concluded that both processes have led to sulfur-deficient samples since the Sb₂S₃ stoichiometry would necessitate 1.5 atoms of sulfur for each atom of antimony. The quantity then of sulfur was lower than expected, in particular, in the case of the spin-coated sample. Table 1 also lists the calculated ionization potentials obtained by UPS measurements. The obtained values were a bit higher (4–7%) than those presented in previous publications^3,6 and they are close to each other within experimental error. Therefore, the diagram of Fig. 1 is justified on the basis of these data.

The light harvesting capacity of the Sb₂S₃/TiO₂/FTO photoanodes can be appreciated by the diffuse reflectance UV-vis spectra of Fig. 5. The absorbance in the two cases cannot be safely compared when the data are obtained by diffuse reflectance, however, it may be argued that the spin-coating deposition method demonstrated higher capacity of photon harvesting. The important conclusion to keep is that in both cases samples were deeply colored and they were strong light absorbers. Also both spectra had practically the same absorption threshold located around 740 nm.

Fig. 5 Diffuse reflectance absorption spectra Sb₂S₃/TiO₂/FTO photoanodes made by (1) chemical bath and (2) spin-coating deposition.

Action spectra have been recorded by using devices made by employing the two types of photoanodes. The obtained results are shown in Fig. 6. The two spectra were very similar. Absorption threshold perfectly matched action spectra threshold, indicating that in both cases Sb₂S₃ is the functional light harvesting species. The slightly higher maximum obtained with the spin-coated sample may be associated with the higher absorbance in that case. Fig. 6 suggests the greater importance of longer wavelength photons in current conversion. However, curve 3 of Fig. 6 gave a different form of IPCE spectrum. Curve 3 corresponds to a cell based on a photoanode made by the spin-coating method but annealing was carried out in air and not in an inert atmosphere as in the case of curves 1 and 2. This result came as a surprise and in order to verify its validity, 5 different cells were constructed and tested. It was then verified that photoanodes annealed in inert atmosphere led to greater contribution from longer wavelength absorbing antimony sulfide species. Annealing in air has been studied in our previous publication.³ In that case, antimony oxide formation was detected and further loss of sulfur was recorded, i.e. the samples were then even more sulfur deficient.

Fig. 6 Action spectra of solar cell devices made with Sb₂S₃/TiO₂/FTO photoanodes based on (1) chemical bath and (2) spin-coating deposition. Curve (3) corresponds to a photoanode made by spin-coating but calcined in air. The maximum values of the action spectra are adapted to the corresponding values of current density given in Table 2.

The above data support the fact that antimony sulfide deposited on titania films by the two methods may both be used as sensitizers of titania in HSCs. The current–voltage data obtained by characterization of the constructed devices are presented in Fig. 7 and Table 2. The two methods of antimony sulfide deposition produced the same practical effect. Spin-coating yielded a bit higher current and voltage but suffered of a lower fill factor. Over all, chemical bath deposition led to cells of slightly higher efficiency but the difference is judged as being within experimental error. Therefore, both methods are applicable and since spin coating is much simpler and handy, it is suggested as the method of choice. The results of Fig. 7 and Table 2 report improved data compared to data reported in our previous publication.³ In any case, differences are within experimental error.

Fig. 7 Current density–voltage curves recorded with solar cell devices comprising Sb₂S₃/TiO₂/FTO photoanodes based on (1) chemical bath and (2) spin-coating deposition.

Table 2 Current density–voltage data extracted from Fig. 7

Method of film deposition	Jsc (mA cm^−2)	Voc (volts)	FF	η%
(1) Chemical bath	18.9	0.50	0.40	3.8
(2) Spin coated	19.8	0.53	0.33	3.5

---

We have also examined the possibility of hole transporting material substitution and its effect on the behavior of antimony sulfide solar cells. The results are shown in Fig. 8 and Table 3. It is obvious that the best combination is P3HT/PEDOT:PSS/Ag. Absence of PEDOT:PSS had a detrimental effect on cell efficiency while the employment of spiro-OMETAD, which has been very successful in the case of perovskite solar cells is of no use in the case of antimony sulfide solar cells. This of course, is not a particular problem since both P3HT and PEDOT:PSS are more easy and less expensive to synthesize and this is reflected on their commercial prices. The HOMO level of spiro-OMETAD is located at about −5.22 eV,¹⁷ i.e. very close to the presently used P3HT (−5.2 eV), therefore, there exist no potential constraints for its acting as hole scavenger. Apparently, the low efficiency obtained with spiro-OMETAD is a matter of poor interface forming between antimony sulfide and this small-molecule hole transporter. Obviously, this should be a subject of further research.

Fig. 8 Current density–voltage curves recorded with solar cell devices comprising a Sb₂S₃/TiO₂/FTO photoanode and various hole transporting or counter electrode materials: (1) P3HT/PEDOT:PSS/Ag; (2) P3HT/Au and (3) spiro-OMETAD/Au.

Table 3 Current density–voltage data extracted from Fig. 8

Hole-transporter/counter electrode	Jsc (mA cm^−2)	Voc (volts)	FF	η%
(1) P3HT/PEDOT:PSS/Ag	18.9	0.50	0.40	3.8
(2) P3HT/Au	15.8	0.35	0.30	1.7
(3) Spiro-OMETAD/Au	4.5	0.34	0.40	0.7

---

Experimental

Materials

Unless otherwise indicated, reagents were obtained from Aldrich and were used as received. Commercial nanocrystalline titania Degussa P25 was used in all cell constructions and Millipore water was used in all experiments. SnO₂:F transparent conductive electrodes (FTO, resistance 8 Ω sq.⁻¹) were purchased from Pilkington, PEDOT:PSS conductive polymer (Baytron P) from H. C. Starck, regioregular P3HT from Rieke Metals and Silver ink (5007E) from DuPont.

Construction of photoanode electrodes

FTO-coated glass substrates were cut in the dimensions of 1 cm × 3 cm. One third of the conductive layer was removed using zinc power and hydrochloric acid. Then they were washed with mild detergent, rinsed several times with distilled water and subsequently with ethanol in an ultrasonic bath, finally dried under air stream. The deposition of the functional film was made on an area of 1 cm × 2 cm (including the etched area) while the rest 1 cm × 1 cm was left clear for electric conduct. A compact thin layer of TiO₂ was first deposited by aerosol spray pyrolysis using a solution of 0.2 M diisopropoxytitanium bis(acetylacetonate) in EtOH. During spraying, the substrate was heated at about 70 °C using a hot plate. After spraying, the samples were heated for 1 h at 500 °C. Subsequently, a mesoporous TiO₂ layer composed of dilute titania paste made of P25 nanoparticles was spin coated at 3000 rpm for 30 s followed by calcination for 15 min at 500 °C. Finally, a layer of TiCl₄, was deposited on the top of the TiO₂ layer by dipping the patterned electrodes into a solution made of 0.04 M TiCl₄ in H₂O at 70 °C for 30 min, then copiously rinsing and finally annealing at 500 °C. The titania film thickness was approximately 500 nm.

Sb₂S₃ was deposited on a freshly prepared titania film either by chemical bath deposition or by spin-coating a precursor solution. In the first case, the bath was prepared by dissolving 650 mg of SbCl₃ in 2.5 ml of acetone and subsequently adding 25 ml of 1 M Na₂S₂O₃ aqueous solution at 10 °C and sufficient water to make a final volume of 75 ml³. After many tests, we concluded that the optimum time of deposition is 2 h at 10 °C. The as-deposited orange films of amorphous Sb₂S₃ were annealed at 300 °C for 10 min in Ar atmosphere. After annealing, the color of the samples became dark-brown. In the second case, a precursor solution was prepared by dissolving 1 M of SbCl₃ in DMF, stirring for 30 min and then adding 2 M of thiourea and stirring for an additional 30 min.⁹ This solution is stable if stored in the dark. The next step was to spin-coat the precursor solution on the titania film at 2000 rpm for 1 min followed by calcination in Ar atmosphere at 300 °C for 10 min.

Device construction

As soon as the final Sb₂S₃/TiO₂/FTO films were taken out of the oven, the HTM, poly(3-hexylthiophene) (P3HT) in the present case, was successively deposited using a solution of 20 mg ml⁻¹ in chlorobenzene by spin coating at 3000 rpm for 1 min. The sample was dried for 15 min at 100 °C and then we deposited a layer of the as provided PEDOT:PSS solution by spin-coating at 1000 rpm. Then the assembly was dried again at 150 °C. Finally, silver counter electrodes were painted on the top of this cell by means of a thin brush and a mask using silver paste. Final further drying was done at 130 °C for 1 min. The size of the Ag electrodes was in all cases 0.2 cm × 0.5 cm. PEDOT:PSS and Ag may also be deposited by screen printing.³ Electric characteristics of the cells were measured by illumination through a mask with 0.2 cm × 0.5 cm aperture dimensions, i.e. the aperture corresponded to the size of the counter electrodes. Illumination of the samples was made with a PECCELL PEC-L01 Solar Simulator set at 100 mW cm⁻².

Measurements

J–V characteristic curves were recorded under ambient conditions with a Keithley 2601 source meter that was controlled by Keithley computer software (LabTracer). IPCE values were obtained with an Oriel IQE 200 system. UV-vis absorption spectra were recorded using a Shimadzu model 2600 spectrophotometer equipped with an integration sphere. XRD measurements were made with a Bruker D8 advance XRD, with the following characteristics: CuKα, λ = 1.5406 A, 40 kV, 40 mA and 2θ ranging 5–90°.

XPS experiments were carried out in an ultra high vacuum system (UHV) which consists of a fast entry specimen assembly, a sample preparation and an analysis chamber. The base pressure in both chambers was 1 × 10⁻⁹ mbar. Unmonochromatized MgKα line at 1253.6 eV and an analyzer pass energy of 97 eV, giving a full width at half maximum (FWHM) of 1.6 eV for the Au4f_7/2 peak, were used in all XPS measurements. The binding energy (BE) scale was calibrated by assigning the main C1s peak at 285.0 eV. UPS spectra were obtained using HeI irradiation with hν = 21.22 eV produced by a UV source (model UVS 10/35). During UPS measurements the analyzer was working at the Constant Retarding Ratio (CRR) mode, with CRR = 10. A bias of −12.29 V was applied to the sample in order to avoid interference of the spectrometer threshold in the UPS spectra. The high and low binding energy and highest occupied cutoff positions were assigned by fitting straight lines on the high and low energy cutoffs of the spectra and determining their intersections with the binding energy axis.

Conclusions

The above data show in a very clear manner that the two methods of antimony sulphide deposition have the same practical effect, therefore, spin-coating, which is simpler and easier⁹ is suggested as the preferred method of deposition. Antimony sulphide solar cells are easy to make and can be fabricated with relatively low cost materials. The best hole-transporter/counter electrode combination found in the present work is P3HT/PEDOT:PSS/Ag. All components, i.e. photoanode, hole-transporter and counter electrode can be deposited by easy procedures under ambient conditions. Antimony sulphide solar cells yielded high current densities but relatively low open-circuit voltage compared to perovskite solar cells. Structure amelioration is also necessary in order to obtain higher fill factors.

Acknowledgements

Authors are grateful to George Paterakis (FORTH-ICEHT) for his help with annealing of antimony sulphide films. Esmaiel Nouri acknowledges a fellowship granted by the Ministry of Science, Research and Technology of the Islamic Republic of Iran that allowed his visit to the University of Patras, Greece.

Notes and references

1. S. Ameen, M. A. Rub, S. A. Kosa, K. A. Alamry, M. S. Akhtar, H.-S. Shin, H.-K. Seo, A. M. Asiri and M. K. Nazeeruddin, ChemSusChem, 2016, 9, 10.
2. M. Shahbazi and H. Wang, Sol. Energy, 2016, 123, 74.
3. A. Nikolakopoulou, D. Raptis, V. Drakopoulos, L. Sygellou, K. S. Andrikopoulos and P. Lianos, J. Power Sources, 2015, 278, 404.
4. F. Yang, J. Xi, L.-Y. Gan, Y. Wang, S. Lu, W. Ma, F. Cai, Y. Zhang, C. Cheng and Y. Zhao, J. Colloid Interface Sci., 2016, 464, 1.
5. J. A. Christians, D. T. Leighton and P. V. Kamat, Energy Environ. Sci., 2014, 7, 1148.
6. J. Zhong, X. Zhang, Y. Zheng, M. Zheng, M. Wen, S. Wu, J. Gao, X. Gao, J.-M. Liu and H. Zhao, ACS Appl. Mater. Interfaces, 2013, 5, 8345.
7. K. C. Godel, Y. C. Choi, B. Roose, A. Sadhanala, H. J. Snaith, S. I. Seok, U. Steiner and S. K. Pathak, Chem. Commun., 2015, 51, 8640.
8. N. Maiti, S. H. Im, C.-S. Lim and S. I. Seon, Dalton Trans., 2012, 41, 11569.
9. Y. C. Choi and S. I. Seok, Adv. Funct. Mater., 2015, 25, 2892.
10. N. Balis, D. Konios, E. Stratakis and E. Kymakis, ChemNanoMat, 2015, 5, 346.
11. S. D. Stranks and H. J. Snaith, Nat. Nanotechnol., 2015, 10, 391.
12. I. Cordanov, Semicond. Sci. Technol., 1994, 9, 1234.
13. S. Messina, M. T. S. Nair and P. K. Nair, Thin Solid Films, 2007, 515, 5777.
14. J. H. Rhee, C.-C. Chung and E. W.-G. Diau, NPG Asia Mater., 2013, 5, 68.
15. S. H. Im, C.-S. Lim, J. A. Chang, Y. H. Lee, N. Maiti, H.-J. Kim, M. K. Nazeeruddin, M. Gratzel and S. I. Seok, Nano Lett., 2011, 11, 4789.
16. Y. C. Choi, D. U. Lee, J. H. Noh, E. K. Kim and S. I. Seok, Adv. Funct. Mater., 2014, 24, 3587.
17. J. Seo, N. J. Jeon, W. S. Yang, H. W. Shin, T. K. Ahn, J. Lee, J. H. Noh and S. I. Seok, Adv. Energy Mater., 2015, 5, 1501320.

---

Footnote

† Permanent address: Department of Materials Science and Engineering, Sharif University of Technology, Azadi Str., Tehran, Iran.

---