A non-toxic, solution-processed, earth abundant absorbing layer for thin-film solar cells

Abstract

Copper zinc tin sulfide (Cu₂ZnSnS₄, CZTS) has attracted significant attention in the past few years as a next generation absorber material for the production of thin film solar cells on large scales due to the high natural abundance of all constituents, tunable direct band gap energy ranging from 1.0 to 1.5 eV, and large absorption coefficient. In addition, to address the issue of expensive vacuum-based processes, non-vacuum solution-based approaches are being developed for CZTS absorber layer deposition. Here, we demonstrate the fabrication of a high quality CZTS absorber layer with a thickness of 2.8–3.0 μm and micrometre-scaled grains (1–2.5 μm) using air-stable non-toxic solvent-based inks. Our approach for the fabrication of CZTS absorber, reported here, will be the first step in achieving low-cost and large area solar cells with high efficiency.

---

Broader context

Solution processing for chalcogenide absorber materials in thin film solar cells is an attractive area of research because these materials have advantages including suitability for use in large-area substrates, high throughput and efficient materials usage. In this communication, we present a facile route to fabricate a Cu₂ZnSnS₄ (CZTS) absorber layer using non-toxic solvent-based ink in which commercially available precursor particles such as Cu₂S, Zn, Sn, and S are dispersed. With our first cells exhibiting an efficiency of 5.14% under AM 1.5 illumination, the use of the non-toxic precursor ink in a scalable coating process provides convenient access to fabricate high quality CZTS absorber layers at low cost and contributes to the large-scale deployment of thin film solar cells.

Copper zinc tin sulfide (Cu₂ZnSnS₄, CZTS) is a very promising material for use as a low cost absorber alternative to other chalcopyrite-type semiconductors based on Ga or In, because it is only composed of abundant and economical elements.^1–5 In addition, CZTS has a direct band gap energy of 1.0–1.5 eV and a large absorption coefficient of over 10⁴ cm⁻¹, properties similar to those of Cu(In,Ga)Se₂ (CIGS), which is regarded as one of the best absorber materials for sustainable and highly efficient solar cells.^6–8 Typically, metal chalcogenide films such as CIGS and CZTS are deposited by evaporation or sputtering techniques that rely on vacuum environments.^9–11 However, this vacuum deposition process suffers from relatively low throughput, low material utilization, and difficulties associated with large-scale production.^12,13 In this regard, solution-based deposition methods are being developed because they have advantages including suitability for use in large-area substrates, high throughput, and efficient materials usage.^14–16 Various solution-based approaches for the fabrication of CZTS thin films have been reported including sol–gel^17,18 and nanocrystal dispersion processes,^4,19 but they face some limitations. The sol–gel method is vulnerable to contamination by carbon, oxygen, and other impurities from precursors or starting solutions, which inevitably leads to the formation of a porous structure with small grain size due to significant shrinkage. The nanocrystal dispersion method requires the complex synthesis of nanocrystals, and it is difficult to achieve dense organic residue-free thick films from the dispersion of nanocrystals capped with stabilizing molecules. Recently, Todorov et al. reported the fabrication of CZTS thin film solar cells with 9.6% power conversion efficiency (PCE) using a hydrazine-based hybrid slurry approach.^20,21 However, hydrazine is a highly toxic and very unstable compound that requires extreme caution during handling and storage. Furthermore, due to the reactive nature of this solvent,²² all processing for slurry and film preparations must be performed under inert atmospheric conditions, and thus hydrazine would not be easily adaptable for large-scale solar cell fabrication. With these considerations, it is highly desirable to develop a robust, easily scalable and relatively safe solution-based process for the fabrication of a high quality CZTS absorber layer.

Here, we devise for the first time a non-toxic solvent-based process for the fabrication of a dense CZTS absorber layer. The slurry (or ink) employed for CZTS deposition is a commercially available powder mixture of Cu₂S, Zn, Sn, and S dispersed in ethanol that is safe and easy to use. As an environmentally benign solvent, ethanol was selected because it can evaporate quickly and thus may minimize residual carbon- or oxygen-containing impurities in the film. The slurry composition was controlled to have the atomic ratio of Cu : Zn : Sn : S = 2 : 1 : 1 : 4. Our simple slurry approach may encounter phase segregation and the presence of unreacted species or other unwanted intermediate compounds in the final film because the precursor particles of Cu₂S, Zn, and Sn are insoluble, unlike hydrazine. In other words, to achieve our goal, the precursor powders should be well-dispersed in the solvent and must be reactive enough to be converted into CZTS granular films during thermal treatment. We employed a milling process to grind precursor powders to nanosize particles and to obtain homogeneous well-dispersed slurry. The large surface areas of finely milled precursor particles can induce material transfer and interparticle densification. In addition, some of the precursor particles retain the low melting points of Zn (420 °C) and Sn (231 °C), which will bring about reactive liquid-phase sintering between constituent particles and/or intermediate compounds at temperatures above 500 °C, even in the presence of Cu₂S with its high melting point (1130 °C).

The thermal behavior of the CZTS precursor ink-containing powder mixture (Cu₂S, Zn, Sn, and S) was analyzed by thermo-gravimetry coupled with differential scanning calorimetry (TG-DSC) under a nitrogen atmosphere (150 cm³ min⁻¹) as shown in Fig. 1a. A weight loss of ∼2.7% accompanying the exothermic peak at 200 °C is ascribed to the partial sublimation of sulfur, while the endothermic peak at 480 °C likely results from the crystallization of CZTS. The TG-DSC data indicate that the precursor inks might be converted into the CZTS phase by annealing at temperatures around 500 °C, lower than the glass transition temperature (T_g) of the soda lime glass (SLG) that is typically used to fabricate the CZTS thin film solar cells. The phase development of the precursor films during annealing is presented in Fig. 1b. Sharp peaks at 2θ = 28.45°, 47.3° and 56.2° can be attributed to the diffraction of the (112), (220) and (312) planes of kesterite structure CZTS (JCPDS no. 26-0575), respectively, suggesting the formation of a CZTS crystalline phase at temperatures ranging from 500 to 530 °C. Fig. 1c presents SEM images showing the microstructural evolution of the precursor films annealed under N₂ + H₂S (5%) atmosphere in a tubular furnace at temperatures ranging from 400 to 530 °C for 30 min. The particle size in the as-prepared granular film was smaller than ∼150 nm (see ESI, Fig. S1†). As the annealing temperature increased, the films were gradually densified, while the grain size increased. When annealed at over 530 °C, a relatively dense structure with large grains (1–2.5 μm) and occasional voids developed. Microstructural observations support the use of particle mixture-based ink for the production of the solution-processed absorbing layer.

Fig. 1 (a) TG-DSC analysis of the ethanol-based CZTS precursor ink. This analysis was performed under a nitrogen atmosphere. (b) XRD analysis of the CZTS film as a function of the annealing temperatures. Enlarged graphs in the 2θ-angle range from 28° to 33° are displayed to show low-intensity peaks. (c) Microstructure evolution of the CZTS film as a function of the annealing temperatures ranging from 400 to 530 °C. The precursor films were annealed under N₂ + H₂S (5%) atmosphere in a tubular furnace.

It should be noted that three XRD diffraction peaks at around 2θ = 28.6°, 47.5° and 56.3° overlap with those of Cu₂S and ZnS, so the crystallization of CZTS cannot be confirmed solely by XRD analysis. Therefore, Raman spectroscopy was utilized to obtain further insight into the phase identification, and the results of the CZTS films as a function of the annealing temperatures are shown in Fig. 2. The as-prepared precursor films exhibited a strong peak at 473 cm⁻¹ as well as small peaks at around 260 cm⁻¹, which correspond to precursor components such as Cu₂S and ZnS. These undesirable phases disappear completely when annealed at 530 °C, which is in good agreement with the XRD results. For the sample annealed at 530 °C, peaks were observed at 251, 287, 338, and 368 cm⁻¹; all of these peaks can be assigned to kesterite CZTS.^23,24 Raman analysis indicates that the large surface area of the finely milled precursor particles and the low melting points of Zn (420 °C) and Sn (231 °C) promoted the crystallization of CZTS at the temperature of 530 °C and allowed for the formation of a dense absorbing layer by a reactive liquid phase sintering.

Fig. 2 Raman spectra of the CZTS films as a function of annealing temperatures. Detailed graphs indicate traces of unreacted phases.

A cross-sectional image of a film annealed at 530 °C is shown in Fig. 3a, in which a uniform dense structure without significant large pores and/or cracks can be observed. It should be noted that such a relatively thick (∼2.9 μm) film can be achieved only by three consecutive spin-coatings. The surface composition of the CZTS film was determined by electron probe microanalysis (EPMA) as shown in Fig. 3b and Table 1. The surface composition of the film was relatively uniform and the average composition was close to the starting precursor composition (25.0 at.% Cu, 12.5 at.% Zn, 12.5 at.% Sn, and 50.0 at.% S). We also confirmed that the impurity levels of carbon and oxygen in the film prepared under atmospheric conditions were about 3%, which is lower than that (>5%) of the chalcogenide films fabricated by other solution-based approaches.^25,26 In addition, considering that the oxygen of the SLG substrate could be detected by EPMA, these impurity levels may be regarded as negligible. Fig. 3c shows the compositional depth profile of the CZTS films annealed at 530 °C for 30 min. No significant compositional variation can be observed across the film. The average composition across the films was also similar to the starting precursor composition. This means that the precursor particles are homogeneously well-dispersed in the ink, resulting in CZTS phase formation with uniform composition after heat treatment even though four individual precursor particles are involved.

Fig. 3 (a) Cross-sectional image of the film annealed at 530 °C. (b) Composition mapping at the surface of the CZTS films annealed at 530 °C by EPMA. (c) Component depth profile by Auger electron spectroscopy. (d) Band gap energy of the CZTS films annealed at 500 and 530 °C.

Table 1 Composition ratios at the surface of the CZTS films annealed at 530 °C for 30 min by EPMA

Composition ratio (before heat treatment)			Composition ratio (after heat treatment at 530 °C under N2 + 5% H2S)
Cu/(Zn + Sn)	Zn/Sn	S/metal	Cu/(Zn + Sn)	Zn/Sn	S/metal	Atomic ratio % (Cu/Zn/Sn/S/O/C)
1.06	0.84	1.0	0.99	0.93	0.98	23.5 : 11.5 : 12.3 : 46.2 : 3.5 : 2.9

---

The optical absorption coefficient (α) is obtained from the measured spectral transmittance (T_λ) and reflectance (R_λ) data using the following formula:

α = 1/t ln[(1 − R_λ)²/T_λ] (1)

where t is the thickness of the film. The nature of the optical transitions and the optical band gap (E_g) of the film are obtained from eqn (2):

α = A(hν − E_g)ⁿ/hν (2)

where A is a constant. The exponent n can take values of 2, 1/2, 3 or 3/2 for indirect-allowed, direct-allowed, indirect-forbidden or direct-forbidden transitions, respectively. The values of α are found to obey eqn (2) for n = ½, indicating that the optical transitions are direct-allowed in nature. Therefore, the E_g of CZTS with direct transitions can be determined by applying eqn (3):

(αhν)² = A(hν − E_g) (3)

The optical band gap is determined by extrapolating the linear region of the plot (αhν)²versus hν and taking the intercept on the hν-axis. Fig. 3d presents (αhν)²versus hν plots of the CZTS films annealed at temperatures of 500 and 530 °C. The direct optical band gap energy is found to be 1.44 eV for the CZTS film annealed at 530 °C. This E_g value measured from our CZTS films is similar to the band gap of bulk CZTS reported by others. In contrast, for the sample that was heat treated at 500 °C, the band gap values of the CZTS films increased to 1.66 eV. This increase might be due to the presence of Cu₂S, which has the direct optical band gap in the range 1.7–2.16 eV.^27,28

We fabricated thin film solar cells using non-toxic solvent-based ink. The CZTS precursor layer of ∼2.9 μm in thickness was formed by three spin coatings. The precursor film was dried at 80 °C followed by annealing under N₂ + H₂S (5%) atmosphere at 530 °C for 30 min. Chen et al. reported that Cu-poor and Zn-rich conditions improve the efficiency of the CZTS solar cells because a Cu-poor composition enhances the formation of Cu vacancies, which gives rise to shallow acceptors in the CZTS, while a Zn-rich condition suppresses the substitution of Cu at Zn sites, which results in relatively deep acceptors.²⁹ Therefore, our film composition for the cell performance measurement was selected to include Cu-poor and Zn-rich compositions (approximately Cu/(Zn + Sn) = 0.8 and Zn/Sn = 1.2) (see ESI, Fig. S2†). The sintered CZTS absorber films were processed into photovoltaic devices following standard procedures, including the chemical bath deposition of CdS (∼50 nm), DC sputtering of i-ZnO (∼50 nm), RF sputtering of ITO (∼250 nm), and thermal evaporation of a patterned Ni/Al grid as the top electrode, as shown in Fig. 4a. Finally, the samples (2 × 2.5 cm²) were mechanically scribed into the cells with a total area of 0.25 cm². The current–voltage (I–V) characteristics for our best performing CZTS solar cell measured in the dark and under AM 1.5 illumination are shown in Fig. 4b. All device performance parameters were reported based on the cell area, excluding the shaded areas (∼11% of the total device area) by the Ni/Al finger electrode. The as-fabricated device exhibited a total area efficiency of 5.14% [open-circuit voltage (V_oc) = 0.517 V, short-circuit current density (J_sc) = 18.86 mA cm⁻², fill factor (FF) = 52.8%]. Fig. 4c shows the external quantum efficiency (EQE) of the corresponding solar cell as a function of photon wavelength. The maximum quantum efficiency of 74% is obtained for a photon wavelength of 540 nm. The band gap of the absorber layer is determined to be 1.51 eV by fitting a plot of [E ln(1 − EQE)]²vs. E near the band edge, as shown in the inset of Fig. 4c. The band gap energy of a CZTS film with the Cu-poor and Zn-rich composition is larger than the stoichiometric film (1.44 eV). The observed value is reasonable since the band gap energy of CZTS shifts to higher energies as Cu/(Zn + Sn) decreases.³⁰

Fig. 4 (a) Cross-sectional image of the CZTS thin film solar cell. Inset shows a photograph of the as-fabricated solar cell. The samples (2 × 2.5 cm²) were mechanically scribed into the cells with a total area of 0.25 cm². Note that the stoichiometry of the CZTS film was selected to yield Cu-poor and Zn-rich compositions (approximately Cu/(Zn + Sn) = 0.8 and Zn/Sn = 1.2). (b) Current–voltage (I–V) characteristics of the CZTS solar cell annealed at 530 °C for 30 min. The efficiency of the cell is 5.14% under standard AM 1.5 illumination. (c) External quantum efficiency (EQE) curve of the corresponding cell. The band gap of the absorber layer is determined to be 1.51 eV by a plot of [E ln (1 − EQE)]²vs. E, as shown in the inset.

Although the initial efficiency of our cell is lower than those of the other solution-processed cells such as those produced by the hybrid slurry method (9.6%) and nanocrystal dispersion (7.2%),^4,20 improvements in device performance are expected with further studies to resolve several issues. A possible reason for the low efficiency is the relatively thick CZTS absorber layer. Although the devices made with thicker CZTS layers absorb more light, the high intrinsic resistivity of the p-type CZTS absorber layer itself and the charge carrier trap density inherent in the thick layer can contribute to the increase in high series resistance and low short circuit current that lead to losses in efficiency,^4,31 suggesting the fabrication process must be optimized for the thinner film. In addition, fine-tuning the band gap of the CZTS film through the replacement of S by Se and MgF₂ antireflection coating on top of the device are currently underway to improve the cell efficiency. We believe that resolving these issues will allow us to produce large and low cost CZTS solar cells with much higher efficiencies, which is highly desirable for photovoltaic applications.

Conclusions

In summary, our simple solution-based deposition approach employs a non-toxic solvent (ethanol)-based ink composed of commercially available precursor particles. Our readily achievable air-stable precursor ink, without the involvement of complex particle synthesis, high toxic solvents, or organic additives, facilitates a convenient method to fabricate a high quality CZTS absorber layer with uniform composition at the surface and across the thin depth. Well-dispersed ink containing four different finely milled precursor particles of low melting points allows for the CZTS crystallization when annealed at 530 °C and forms dense films with large grains (1–2.5 μm), possibly by a reactive liquid-phase sintering between the constituent particles. The preliminary conversion efficiency and fill factor for the non-toxic ink based solar cells are 5.14% and 52.8%, respectively, although the processing details are not yet optimized. Our simple and safe approach reported here represents the first step toward realizing low-cost, large-area, high efficiency solar cells.

Acknowledgements

This research was financially supported by the Basic Research Laboratory (BRL) Program through an NRF grant funded by the MEST (No. 2011-8-2048). It was also partially supported by the Second Stage of the Brain Korea 21 Project.

References

1. J. Paier, R. Asahi, A. Nagoya and G. Kresse, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 79, 115126.
2. J. Scragg, P. Dale, L. Peter, G. Zoppi and I. Forbes, Phys. Status Solidi B, 2008, 245, 1772.
3. H. Katagiri, K. Jimbo, W. Maw, K. Oishi, M. Yamazaki, H. Araki and A. Takeuchi, Thin Solid Films, 2009, 517, 2455.
4. Q. Guo, G. M. Ford, W. Yang, B. C. Walker, E. A. Stach, H. W. Hillhouse and R. Agrawal, J. Am. Chem. Soc., 2010, 132, 17384.
5. C. Steinhagen, M. G. Panthani, V. Akhavan, B. Goodfellow, B. Koo and B. A. Korgel, J. Am. Chem. Soc., 2009, 131, 12554.
6. K. Ito and T. Nakazawa, Jpn. J. Appl. Phys., 1988, 27, 2094.
7. T. Tanaka, T. Nagatomo, D. Kawasaki, M. Nishio, Q. Guo, A. Wakahara, A. Yoshida and H. Ogawa, J. Phys. Chem. Solids, 2005, 66, 1978.
8. S. C. Riha, B. A. Parkinson and A. L. Prieto, J. Am. Chem. Soc., 2009, 131, 12054.
9. F. Liu, K. Zhang, Y. Lai, J. Li, Z. Zhang and Y. Liu, Electrochem. Solid-State Lett., 2010, 13, H379.
10. H. Katagiri, K. Jimbo, S. Yamada, T. Kamimura, W. S. Maw, T. Fukano, T. Ito and T. Motohiro, Appl. Phys. Express, 2008, 1, 041201.
11. K. Jimbo, R. Kimura, T. Kamimura, S. Yamada, W. Maw, H. Araki, K. Oishi and H. Katagiri, Thin Solid Films, 2007, 515, 5997.
12. A. Weber, R. Mainz and H. W. Schock, J. Appl. Phys., 2010, 107, 013516.
13. S. E. Habas, H. A. S. Platt, M. F. A. M. van Hest and D. S. Ginley, Chem. Rev., 2010, 110, 6571.
14. T. K. Todorov and D. B. Mitzi, Eur. J. Inorg. Chem., 2010, 2010, 17.
15. F. C. Krebs, Sol. Energy Mater. Sol. Cells, 2009, 93, 394.
16. Z. Zhou, Y. Wang, D. Xu and Y. Zhang, Sol. Energy Mater. Sol. Cells, 2010, 94, 2042.
17. M. Yeh, C. Lee and D. Wuu, J. Sol-Gel Sci. Technol., 2009, 52, 65.
18. K. Tanaka, N. Moritake and H. Uchiki, Sol. Energy Mater. Sol. Cells, 2007, 91, 1199.
19. T. Kameyama, T. Osaki, K. Okazaki, T. Shibayama, A. Kudo, S. Kuwabata and T. Torimoto, J. Mater. Chem., 2010, 20, 5319.
20. T. K. Todorov, K. B. Reuter and D. B. Mitzi, Adv. Mater., 2010, 22, E156.
21. O. Gunawan, T. K. Todorov and D. B. Mitzi, Appl. Phys. Lett., 2010, 97, 233506.
22. L. F. Audrieth and B. Ackerson Ogg, The Chemistry of Hydrazine, John Wiley & Sons, New York, 1951.
23. H. Yoo and J. Kim, Sol. Energy Mater. Sol. Cells, 2011, 95, 239.
24. P. A. Fernandes, P. M. P. Salomé and A. F. da Cunha, Thin Solid Films, 2009, 517, 2519.
25. M. Kemell, M. Ritala and M. Leskela, J. Mater. Chem., 2001, 11, 668.
26. M. Krunks, O. Kijatkina, H. Rebane, I. Oja, V. Mikli and A. Mere, Thin Solid Films, 2002, 403–404, 71.
27. Y. B. Kishore Kumar, P. Uday Bhaskar, G. Suresh Babu and V. Sundara Raja, Phys. Status Solidi A, 2010, 207, 149.
28. A. C. Rastogi and S. Salkalachen, Thin Solid Films, 1982, 97, 191.
29. S. Chen, X. G. Gong, A. Walsh and S. Wei, Appl. Phys. Lett., 2010, 96, 021902.
30. G. S. Babu, Y. B. K. Kumar, P. U. Bhaskar and V. S. Raja, Sol. Energy Mater. Sol. Cells, 2010, 94, 221.
31. V. A. Akhavan, B. W. Goodfellow, M. G. Panthani, D. K. Reid, D. J. Hellebusch, T. Adachi and B. A. Korgel, Energy Environ. Sci., 2010, 3, 1600.

---

Footnotes

† Electronic supplementary information (ESI) available: A detailed description of the experimental methods, surface SEM image of as-prepared CZTS film, component depth profile of the CZTS film with the Cu-poor and Zn-rich composition for cell fabrication. See DOI: 10.1039/c1ee02314d

‡ The paper was presented in part at the International Chemical Congress of Pacific Basin Societies (Pacifichem 2010), in Honolulu, Hawaii, USA, December 15–20, 2010.

---