10× faster synthesis of chalcogenide solid solutions with tunable S : Se ratio by NaBH₄-activated S + Se precursors

Abstract

Reducing the reaction time of solvothermal synthesis has significance not just for industrial but for academic applications. In this paper, we develop a 10× faster solvothermal method to synthesize metal chalcogenide solid solutions with tunable S : Se ratio by activated multi chalcogen precursors.

---

Low-cost and eco-friendly metal chalcogenide solid solutions M_y(S_1−xSe_x)_z, inheriting the properties from metal sulfides and metal selenides, have been widely applied in solar cells, photodetectors, biosensors, nanoscale electric switches, transistors, laser diodes, and optical memories.^1–6 Additionally, the tunability of S : Se ratios can modulate the properties of the solid solutions, such as band gap energy, electrical conductivity, and life time of the charge carriers.^3,7,8 M_y(S_1−xSe_x)_z solid solutions synthesized by pulsed laser deposition,⁹ thermal evaporation¹⁰ and reactive liquid-phase sintering^11,12 technologies have confirmed the application value of tunable S : Se ratios. However, stringent operating conditions and expensive instruments limit the application of these methods.

The solvothermal method, which is facile, cost-efficient and easily scalable, also can be used to synthesize metal chalcogenide solid solutions with tunable S : Se ratios, but it always takes a long time to obtain target products, for example, Cao et al. have prepared Cu₂ZnSn(S,Se)₄ (CZTSSe) with a tunable S : Se ratio, but it took as long as 24 h.¹³ Since the long reaction time lowers the efficiency in both industrial and academic applications, reducing the reaction time is of much significance.

Typically, the average reaction time for synthesizing metal selenides by solvothermal methods is over 24 h,^14–19 while that for the synthesis of metal sulfides is approximately 12 h.^20–23 The difference in the reaction time is due to the different activities between Se precursors and S precursors. Commonly used S precursors, including elemental S powder, thiourea²⁴ and dodecanethiol,²⁵ with a high solubility in commonly used amine solutions, provide a homogeneous reaction environment and highly reactive S_n²⁻ precursors,^26,27 which leads to shorter reaction time. While Se powder, which is always used as the Se precursor due to the high production cost for other Se precursors like sodium selenide and selenourea, is characterized by low solubility and little ionization degree in amine solution, leading to a heterogeneous reaction environment. The heterogeneous reaction environment and low reactivity of precursors limit the diffusion and reaction rate between the reagents. Furthermore, the difference between the reactivities of S and Se sources adds difficulty to the solvothermal synthesis of M_y(S_1−xSe_x)_z solid solutions.

Therefore, accelerating the solvothermal process to prepare chalcogenides solid solutions with continuously tunable S : Se ratios needs to develop multi-chalcogen precursors which can guarantee a high and balanced reactivity and a homogenous reaction environment. One effective way is to activate chalcogen precursors by reductants. Hsu et al. reported a solvothermal method to synthesize CuIn(S,Se)₂ by using sodium thioglycolate (TGA) as chelating agent and reducing agent for chalcogen precursors and reduced the reaction time to 3 h.²⁸ However, the S : Se ratio cannot be easily tuned due to excessive TGA required in the reaction solution, which is simultaneously used as sulfur source and reductant. Sodium borohydride (NaBH₄), a strong reductant which has been applied in the synthesis of selenides,²⁹ will not affect the control over the S : Se ratio in products by varying the proportion of starting materials.

Herein we report a time-saving and facile solvothermal method to synthesize nanocrystals of metal chalcogenides solid solutions with tunable S : Se ratios by using NaBH₄ to reduce elemental sulfur and selenium to generate highly activated species, such as HSe⁻, HS⁻, Se_n²⁻ and S_n²⁻.^29,30 As an illustration, Cu-based chalcogenides was synthesized to demonstrate the effectiveness of this method.

As depicted in Scheme 1, this time-saving solvothermal method consists of two steps: (i) preparing highly reactive multi-chalcogen precursor by mixing NaBH₄, S powder, Se powder with ethylenediamine (en) and reacting at 200 °C for 2 h; and (ii) synthesizing nanocrystals of metal chalcogenides solid solutions by solvothermal reactions at 200 °C for 2 h. By using this method, nanocrystals of Cu₂ZnSn(S_1−xSe_x)₄, Cu_2−y(S_1−xSe_x), CuIn(S_1−xSe_x)₂ with tunable S : Se ratios were prepared by tuning the ratio of S and Se powder in reaction precursors. The possible chemical reactions involved in the formation of Cu₂ZnSn(S_1−xSe_x)₄ are described by eqn (S1)–(S6) (ESI†).^30–33 After reactions, NaBH₄ may form the by-products of sodium salts and multipolymers of BH₃, which can be removed from the final products by purification. The crystal structure, morphology, composition and absorption spectrum of as-synthesized M_y(S_1−xSe_x)_z solid solutions were characterized by X-ray diffraction (XRD), scanning electron microscope (SEM), energy dispersive X-ray spectrum (EDS), Raman spectroscopy, high-resolution transmission electron microscopy (HRTEM) and UV-vis-NIR spectrophotometer. The experiment details can be found in the ESI.†

Scheme 1 Process flow chart of the presented time-saving solvothermal synthesis of chalcogenide solid solutions.

Fig. 1a and b present the XRD patterns of solvothermal synthesized Cu₂ZnSn(S_0.5Se_0.5)₄ nanocrystals at 200 °C with the NaBH₄-activated chalcogen precursors in 2 h and with the unactivated chalcogen precursors in 12 h, respectively. The diffraction peaks in Fig. 1a, which are well indexed to planes (112), (220/204), (312) respectively, can be assigned to kesterite Cu₂ZnSnS₄ (JCPDS no. 26-0575) and Cu₂ZnSnSe₄ (JCPDS no. 52-0868) and no split peaks are found. Diffraction peaks in Fig. 1b are split, illustrating an incomplete reaction of synthesizing Cu₂ZnSn(S,Se)₄. Because the characteristic diffraction patterns of Cu₂ZnSn(S,Se)₄ overlap with those of Zn(S,Se) and Cu₂Sn(S,Se)₃,³⁴ Raman spectroscopy and HRTEM were further utilized for phase identification. Fig. 1c shows a Raman spectrum of Cu₂ZnSn(S_0.5Se_0.5)₄ nanocrystals derived from the NaBH₄-activated chalcogen precursors. Two peaks located at 216 cm⁻¹ and 330 cm⁻¹ are observed, which correspond to the main vibrational A₁ symmetry modes from Cu₂ZnSnSe₄ and Cu₂ZnSnS₄, respectively.¹² The position and broadening feature of the observed peaks are consistent with that of Cu₂ZnSn(S,Se)₄ in previous reports.^12,35,36 Characteristics peaks from the other impurities such as binary and ternary chalcogenides are not observed. A typical HRTEM image of Cu₂ZnSn(S_0.5Se_0.5)₄ obtained from highly reactive chalcogen precursors is shown in Fig. 1d. Well-resolved lattice fringes are clearly observed. The fringe spacings of 0.314 nm, which corresponded to the (112) planes of kesterite Cu₂ZnSn(S_0.5Se_0.5)₄, is between that of Cu₂ZnSnS₄ (0.310 ± 0.004 nm) and Cu₂ZnSnSe₄ (0.327 ± 0.004 nm).^37,38 The above phase composition analyses demonstrates that the presented time-saving solvothermal synthesis approach exclusively produces single-phase Cu₂ZnSn(S_0.5Se_0.5)₄ solid solution. The difference in the XRD patterns demonstrates that the use of NaBH₄-activated chalcogen precursors can significantly promote the formation kinetics of metal chalcogenides solid solutions and reduce the reaction time to as short as 2 h, which is 10 times faster than the typical solvothermal methods (see Table S1, ESI†).^13–19 The promoted formation kinetics of metal chalcogenides solid solutions can be attributed to that the reduction of elemental S and Se by NaBH₄ to highly reactive species, such as HSe⁻, HS⁻, Se_n²⁻ and S_n²⁻, leads to higher reactivities and solubilities of chalcogen precursors.^29–33 The highly reactive chalcogen precursors and the homogeneous reaction environment may also contribute to the morphological uniformity of products. As shown in Fig. 1e, the Cu₂ZnSn(S_0.5Se_0.5)₄ nanocrystals prepared with NaBH₄-activated multi-chalcogen precursors has a uniform morphology with an average diameter of 50 nm. Slight clustering is observed, which may be caused by the solvent evaporation during testing samples preparation.

Fig. 1 XRD patterns of solvothermal synthesized Cu₂ZnSn(S_0.5Se_0.5)₄ nanocrystals prepared by NaBH₄-activated chalcogen precursors (a) and normal chalcogen precursors (b); Raman spectrum (c), HRTEM (d) and SEM (e) images of Cu₂ZnSn(S_0.5Se_0.5)₄ nanocrystals prepared by NaBH₄-activated chalcogen precursors.

In the presented method, the S : Se ratio could be tuned by varying the ratio of S to Se powders in starting materials. Fig. 2a shows the XRD patterns of Cu₂ZnSn(S_1−xSe_x)₄ with different S : Se ratios. All of the patterns are consistent with kesterite Cu₂ZnSnS₄ (JCPDS no. 26-0575) and Cu₂ZnSnSe₄ (JCPDS no. 52-0868). Meanwhile, no impurity diffraction peak is found. A slight peak shift is observed as S : Se ratios change. As shown in Fig. 2b, the amplified diffraction peaks for (112) crystal plane show a decrease in 2θ values with increasing Se content. This is associated with the expansion of the lattice induced by increasing the content of Se (1.98 Å), which has a bigger atomic size than S (1.84 Å).¹¹ The Raman spectra of Cu₂ZnSn(S_1−xSe_x)₄ with different S : Se ratios can be seen in Fig. S1.† The Raman peaks exhibit a trend of red shift as x value increases in Cu₂ZnSn(S_1−xSe_x)₄. This red shift in Raman spectra is partially due to an increase of bond length caused by substitution of Se(II) ion for smaller S(II) ion, resulting in weaker bonding,^12,39 and partially due to the replacement of Se (M_Se = 78.96) for lighter S (M_S = 32.066) and the mass effect described by Keating's model.^12,40,41 The EDS measurements on CZTSSe nanocrystals with different S : Se ratios are summarized in Table 1. It can be seen that the element compositions (Cu, Zn, Sn, S, Se) in all samples are very close to the stoichiometric chemical compositions. The SEM images of Cu₂ZnSn(S_1−xSe_x)₄ with different S : Se ratios is shown in Fig. S2.† The synthesized particles possess an uniform morphology with the size ranging from 50 to 100 nm. No significant difference among the products with different x value in terms of morphology are observed.

Fig. 2 (a) Powder XRD patterns of Cu₂ZnSn(S_1−xSe_x)₄ with different x values that synthesized with the activated chalcogen precursor at 200 °C for 2 h. (b) Magnification of (112) crystal plane.

Table 1 Measured composition by EDS^a and the proportional relation of as-prepared Cu₂ZnSn(S_1−xSe_x)₄ samples with different x values

Sample	Target product	Cu (at%)	Zn (at%)	Sn (at%)	S (at%)	Se (at%)	Cu/(Zn + Sn)	Zn/Sn	Se/(S + Se)
a EDS measurements: six measurements from two individual samples for each x value were collected and averaged.
1	Cu2ZnSn(S0.75Se0.25)4	27.76	14.54	12.06	32.90	12.75	1.05	1.23	0.28
2	Cu2ZnSn(S0.67Se0.33)4	30.51	15.51	12.76	24.37	16.87	1.10	1.21	0.41
3	Cu2ZnSn(S0.5Se0.5)4	27.68	13.27	12.98	22.23	23.84	1.06	1.04	0.52
4	Cu2ZnSn(S0.33Se0.67)4	28.86	12.67	12.53	14.87	31.07	1.15	1.07	0.68
5	Cu2ZnSn(S0.25Se0.75)4	27.46	14.39	12.49	11.88	33.80	1.02	1.15	0.74

---

Fig. 3 compares the value of Se/(S + Se) in CZTSSe products and precursors. The values of Se/(S + Se) in products vary from 0.28 to 0.74 and are consistent with the S : Se ratios in precursors, indicating that the S : Se ratios can be simply tuned by changing the ratio of S and Se powder in precursors, which demonstrates that the reactivity of S and Se powders be not only improved but also balanced by NaBH₄-activating. Additionally, the dispersion degree of the Se/(S + Se) in products determined by EDS is low, as exhibited by the box chart in Fig. 3. Furthermore, the yield of the presented synthesizing method is higher than 82% (see Table S2, ESI†). Theses results confirm that the presented time-saving solvothermal method is effective, repeatable, and economical.

Fig. 3 Comparison of Se/(S + Se) between Cu₂ZnSn(S_1−xSe_x)₄ products and precursors; box chart exhibits the dispersion degree of Se/(S + Se) in products.

The optical properties of CZTSSe nanocrystals can be tuned by varying their S : Se ratios. The band gap energy (E_g) of CZTSSe nanocrystals were estimated based on the UV-vis absorption spectra measurements.^2,42 As shown in Fig. 4a, the E_g of Cu₂ZnSn(S_1−xSe_x)₄ with different S : Se ratio (x = 0.25, 0.33, 0.50, 0.67, 0.75) are determined to be 1.02 eV, 1.05 eV, 1.12 eV, 1.17 eV, 1.27 eV, respectively. It can be noticed that the E_g of CZTSSe increases monotonically with the decrease of the selenium content, due to the valence band edge moves to higher binding energy with decreasing Se.⁴³ The E_g of CZTSSe are plotted as a function of the Se/S + Se ratio in Fig. 4b, which is consistent with the extended Vegard's law,^13,44 demonstrating the formation of homogeneous solid solution of Cu₂ZnSn(S_1−xSe_x)₄ in the presented experiments. The E_g of the synthesized CZTSSe is attractive for the single junction photovoltaic devices.⁴⁵ Furthermore, the ability to tune the E_g of CZTSSe presented here is helpful for developing advanced multi-junction solar cells.

Fig. 4 (a) The band gap energy of Cu₂ZnSn(S_1−xSe_x)₄ with different S : Se ratios; (b) band gaps of CZTSSe as a function of Se/S + Se.

In order to confirm the universality of the presented time-saving solvothermal method, Cu_2−y(S_1−xSe_x), CuIn(S_1−xSe_x)₂, and Cu₂ZnSnSe₄ were synthesized using NaBH₄-activated chalcogen precursors. Their crystal structure, composition, and morphology were characterized by XRD (Fig. S3–S5, ESI†), Raman spectroscopy (Fig. S3, ESI†), EDS (Tables S3 and S4, Fig. S6 and S7, ESI†), and SEM (Fig. S8 and S9, ESI†). Target products were successfully synthesized in 2 h. Their S : Se ratios and the band gap were simply tuned by varying the ratio of S and Se powder in precursors (Fig. S10 and S11, ESI†).

In conclusion, the reactivity of multi-chalcogen precursors for the solvothermal synthesis of chalcogenides solid solutions are simultaneously improved and balanced by NaBH₄ activating. The nanocrystals of Cu₂ZnSn(S_1−xSe_x)₄, CuIn(S_1−xSe_x)₂, and Cu_2−y(S_1−xSe_x) with tunable S : Se ratios was synthesized in 2 h using the activated multi-chalcogen precursors, which is 10× faster than typical solvothermal methods with normal chalcogen precursors. We believe that the presented time-saving synthetic strategy can be readily extend to other chalcogenides solid solutions and open up a more efficient way for both industrial applications and academic researches.

Acknowledgements

This research was supported by the National Science Funds of China (Contract No. 51102038 and 51472044) and Program for New Century Excellent Talents in University (Contract No. NCET-12-0098).

Notes and references

1. W. Wang, M. T. Winkler, O. Gunawan, T. Gokmen, T. K. Todorov, Y. Zhu and D. B. Mitzi, Adv. Energy Mater., 2014, 4, 1301465.
2. J.-J. Wang, D.-J. Xue, Y.-G. Guo, J.-S. Hu and L.-J. Wan, J. Am. Chem. Soc., 2011, 133, 18558–18561.
3. F. Y. Siddiqui, S. U. Shaikh, D. J. Desale, D. S. Upadhye, S. V. Mahajan, A. V. Ghule, P. Varshney, S.-H. Han and R. Sharma, Mater. Sci. Semicond. Process., 2014, 27, 404–411.
4. J. Varley and V. Lordi, J. Appl. Phys., 2014, 116, 063505.
5. W. Ma, J. M. Luther, H. Zheng, Y. Wu and A. P. Alivisatos, Nano Lett., 2009, 9, 1699–1703.
6. Y. Cao, M. S. Denny Jr, J. V. Caspar, W. E. Farneth, Q. Guo, A. S. Ionkin, L. K. Johnson, M. Lu, I. Malajovich and D. Radu, J. Am. Chem. Soc., 2012, 134, 15644–15647.
7. A. Redinger, M. Mousel, M. H. Wolter, N. Valle and S. Siebentritt, Thin Solid Films, 2013, 535, 291–295.
8. S. Levcenco, D. Dumcenco, Y. Wang, Y. Huang, C.-H. Ho, E. Arushanov, V. Tezlevan and K. Tiong, Opt. Mater., 2012, 34, 1362–1365.
9. J. He, L. Sun, N. Ding, H. Kong, S. Zuo, S. Chen, Y. Chen, P. Yang and J. Chu, J. Alloys Compd., 2012, 529, 34–37.
10. B. Shin, K. Wang, O. Gunawan, K. Reuter, S. Chey, N. Bojarczuk, T. Todorov, D. Mitzi and S. Guha, Proceedings of the 37th Photovoltaic Specialists Conference (PVSC), Washington, 2011.
11. M. Tsega and D.-H. Kuo, J. Alloys Compd., 2013, 557, 142–146.
12. J. He, L. Sun, S. Chen, Y. Chen, P. Yang and J. Chu, J. Alloys Compd., 2012, 511, 129–132.
13. Y. Cao, C. Wang, B. Li, K. Zhang, X. Xu, J. Hu and X. Chen, Jpn. J. Appl. Phys., 2011, 50, 5001.
14. S.-I. Gu, S.-H. Hong, H.-S. Shin, Y.-W. Hong, D.-H. Yeo, J.-H. Kim and S. Nahm, Ceram. Int., 2012, 38, S521–S523.
15. B. Feng, J. Cao, J. Yang, S. Yang and D. Han, Mater. Res. Bull., 2014, 60, 794–801.
16. Y.-F. Du, W.-H. Zhou, Y.-L. Zhou, P.-W. Li, J.-Q. Fan, J.-J. He and S.-X. Wu, Mater. Sci. Semicond. Process., 2012, 15, 214–217.
17. M.-H. Chiang, Y.-S. Fu, T.-F. Guo, H.-L. Liu and W.-T. Lin, Mater. Lett., 2012, 83, 192–194.
18. M.-H. Chiang, Y.-S. Fu, C.-H. Shih, C.-C. Kuo, T.-F. Guo and W.-T. Lin, Thin Solid Films, 2013, 544, 291–295.
19. P. Luo, P. Yu, R. Zuo, J. Jin, Y. Ding, J. Song and Y. Chen, Phys. B, 2010, 405, 3294–3298.
20. M. Shakouri-Arani and M. Salavati-Niasari, New J. Chem., 2014, 38, 1179–1185.
21. M. Shakouri-Arani and M. Salavati-Niasari, J. Ind. Eng. Chem., 2014, 20, 3179–3185.
22. M. Pal, N. Mathews, R. S. Gonzalez and X. Mathew, Thin Solid Films, 2013, 535, 78–82.
23. X. Chen, X. Wang, C. An, J. Liu and Y. Qian, J. Cryst. Growth, 2003, 256, 368–376.
24. T. R. Knutson, P. J. Hanson, E. S. Aydil and R. L. Penn, Chem. Commun., 2014, 50, 5902–5904.
25. C. A. Cattley, C. Cheng, S. M. Fairclough, L. M. Droessler, N. P. Young, J. H. Warner, J. M. Smith, H. E. Assender and A. A. Watt, Chem. Commun., 2013, 49, 3745–3747.
26. Y. Wei, J. Yang, A. W. Lin and J. Y. Ying, Chem. Mater., 2010, 22, 5672–5677.
27. T. Jiang, G. A. Ozin and R. L. Bedard, Adv. Mater., 1994, 6, 860–865.
28. W.-H. Hsu, H.-I. Hsiang and S. Yu, Ceram. Int., 2015, 41, 3208–3213.
29. J. Zhang, S. Zhang, H. Zhang, Y. Zhang, Z. Zheng and Y. Xiang, Mater. Lett., 2014, 122, 306–308.
30. O. S. Oluwafemi and S. P. Songca, Products and Applications of Biopolymers, ed. J. Verbeek, InTech, Rijeka, 2012, ch. 8, pp. 163–184.
31. J. L. Mi, X. B. Zhao, T. J. Zhu, J. P. Tu and G. S. Cao, J. Alloys Compd., 2005, 399, 260–263.
32. Y. Xie, J. Lu, P. Yan, X. Jiang and Y. Qian, J. Solid State Chem., 2005, 155, 42–45.
33. J. Lu, D. B. Dreisinger and W. C. Copper, Hydrometallurgy, 1997, 45, 305–322.
34. S. Ahn, S. Jung, J. Gwak, A. Cho, K. Shin, K. Yoon, D. Park, H. Cheong and J. H. Yun, Appl. Phys. Lett., 2010, 97, 021905.
35. K. Ou, J. Fan, J. Chen, C. Huang, L. Chen, J. Ho and J. Chang, J. Mater. Chem., 2012, 22, 14667–14673.
36. A. Singh, H. Geaney, F. Laffir and K. M. Ryan, J. Am. Chem. Soc., 2012, 134, 2910–2913.
37. S. Riha, B. A. Parkinson and A. L. Prieto, J. Am. Chem. Soc., 2011, 133, 15272–15275.
38. M. Pal, N. R. Mathews, F. Paraguay-Delgado and X. Mathew, Mater. Chem. Phys., 2015, 166, 201–206.
39. X. Peng, S. Zhang and Y. Xiang, J. Alloys Compd., 2015, 640, 75–81.
40. D. Papadimitriou, N. Esser and C. Xue, Phys. Status Solidi B, 2005, 242, 2633–2643.
41. H. Neumann, Helv. Phys. Acta, 1985, 58, 337–346.
42. A. Y. Anderson, Y. Bouhadana, H.-N. Barad, B. Kupfer, E. Rosh-Hodesh, H. Aviv, Y. R. Tischler, S. Rühle and A. Zaban, ACS Comb. Sci., 2014, 16, 53–65.
43. R. Haight, A. Barkhouse, O. Gunawan, B. Shin, M. Copel, M. Hopstaken and D. B. Mitzi, Appl. Phys. Lett., 2011, 98, 253502.
44. H. Wei, Y. Su, S. Chen, Y. Lin, Z. Yang, X. Chen and Y. Zhang, J. Mater. Chem., 2011, 21, 12605–12608.
45. T. Unold and H. W. Schock, Annu. Rev. Mater. Res., 2011, 41, 297–321.

---

Footnote

† Electronic supplementary information (ESI) available: Synthesis and characterization of Cu₂ZnSn(S_1−xSe_x)₄, Cu_2−y(S_1−xSe_x), and CuIn(S_1−xSe_x)₂, and experimental details. See DOI: 10.1039/c6ra15331c

---