*3.2. E*ff*ect of Selenization Time on Properties of CMZTSSe Films*

After a series of analyses and characterizations, it was proved that the best selenization temperature is 530 ◦C. As we all know, selenization time is also one of the major parameters influencing the performance of the absorber layer. In our study, after optimizing the annealing temperature, the influence of annealing time on the properties of CMZTSSe films has been studied. The CMZTSSe films were annealed for 10, 15, and 20 min at 530 ◦C under atmosphere of Se, afterward referred as B1, B2, and B3, respectively.

Figure 7a–c shows the XRD patterns of CMZTSSe films annealed for different times from 10 to 20 min (samples B1, B2, and B3). For the samples B1, B2, and B3, five diffraction peaks located at 28.53◦, 47.33◦, 56.17◦, 69.27◦, and 76.44◦ were observed, conforming to the (112), (220), (312), (008), and (332) planes of kesterite CZTS respectively [25,26]. The characteristic diffraction peaks of other impurity phases were not observed in Figure 7. We can explore the influence of annealing time on the structural performance of CMZTSSe films by observing the peak intensity and peak shift. It is observed from Figure 7a,b that the intensity of the (112) peak is increased, which indicates that the crystal quality is enhanced by increasing the annealing time from 10 to 15 min. Furthermore, the characteristic peak intensity is almost unchanged with the increase of annealing time from 15 to 20 min, as displayed in Figure 7c. In addition, the position of the (112) peaks are shifted to lower 2θ angle with selenization time increasing, as shown in the illustration of Figure 7. Since the concentration of Se in the CMZTSSe matrix increases with the increase of selenization time, there is enlargement of unit cell size, causing the change of the lattice distance in the films. According to the analysis results of XRD, it was found that the crystal structure of CMZTSSe was not changed with the increase of annealing time, and we speculate that the crystal growth of CMZTSSe was completed when the annealing time reached 15 min.

The full width at half-maximum (FWHM), grain size, a-axis lattice constant (a), and c-axis lattice constant (c) of the films annealed for different times from 10 to 20 min (samples B1, B2, and B3) are displayed in Table 4. It was found that by increasing the annealing time from 10 to 20 min, the grain size first increases and then decreases, while the FWHM value first decreases and then increases. When the annealing time is 15 min, the grain size has a maximum value, while the FWHM has a minimum value. The a-axis lattice constant gradually increases from 5.665 to 5.684 Å as the annealing time increases from 10 to 20 min. Meanwhile, the c-axis lattice constant also increases from 11.334 to 11.453 Å. It was concluded that the optimal annealing time is 15 min.

**Figure 7.** XRD spectra of CMZTSSe films annealed at 530 ◦C for different time (**a**) 10 min, (**b**) 15 min, (**c**) 20 min. Inset: Enlarged view of the corresponding (112) diffraction peaks of the CMZTSSe films annealed for different times.

**Table 4.** The full width at half-maximum (FWHM), grain size, a-axis lattice constant (a), and c-axis lattice constant (c) for CMZTSSe films annealed for different times.


As we all know, the diffraction peaks of Cu3SnS4 with a tetragonal structure and ZnS with a cubic structure are close to the diffraction peaks of CZTSSe with a kesterite structure [38,39]. Thus, it is very difficult to detect possible secondary phases by XRD only. In order to further detect the phase compositions of CMZTSSe films, Raman scattering spectra were measured for the films annealed for different times from 10 to 20 min (samples B1, B2, and B3), as shown in Figure 8. The Raman spectrum of sample B1 was fitted using the Gaussian fitting method, and three Raman peaks located at ~193 (A1 mode), 173 (A2 mode), and 234 cm−<sup>1</sup> (E mode) were observed, which conform to the Raman characteristic peaks of CZTSSe [29]. The characteristic peaks of some possible impurity phases were not detected. It is suggested that the CMZTSSe films are composed of a single phase of kesterite CZTSSe. In addition, these characteristic peaks slightly shift to lower values as the selenization time increases, owing to the incorporation of Se in the CMZTSSe compound. It should be noted that the Raman spectra are consistent with the XRD patterns, and some possible secondary phases (Cu2SnSe3, SnSe2, SnSe, ZnSe, and MgSe) were observed. Furthermore, the phase change of the CMZTSSe films did not occur with increasing selenization time from 10 to 20 min. It was concluded that the pure-phase CMZTSSe films were successfully synthesized.

Figure 9a–d depicts the SEM surface images of CZTSSe annealed at 530 ◦C for 15 min (sample A) and CMZTSSe films annealed at 530 ◦C for different times from 10 to 20 min (samples B1, B2, and B3), respectively. As shown in Figure 9a, the surface of the sample A displays pin-hole free morphology and small grain size between 500–800 nm, with a rough surface. Figure 9b shows the surface morphology of sample B1, and it was clearly observed that the CMZTSSe film still consists of nanograins, where the grain size is between 500 and 1000 nm with a large number of holes on the surface. As the annealing time increased to 15 min, obvious morphological change is observed (Figure 9c), with the grains size of sample B2 increasing sharply to the micron level (1.0–2.5 μm) and the surface becoming dense and flat. By further extending the annealing time to 20 min, the surface morphology of sample B3 becomes rough but still dense, and the grains size slightly decreases to 0.8–1.3 μm. It was concluded

that the optimal selenization time is 15 min, and the CMZTSSe films obtained at this time have the best crystallinity, the largest crystal grain size, and their surfaces are dense and flat.

**Figure 8.** Raman spectra of the CMZTSSe films annealed for different times. Inset: The main Raman peaks of A1 mode for CMZTSSe films annealed for different times.

**Figure 9.** SEM images of CZTSSe annealed at 530 ◦C for (**a**) 15 min and CMZTSSe films annealed at 530 ◦C for (**b**) 10 min; (**c**) 15 min; and (**d**) 20 min.

The EDS results of the films annealed for different times from 10 to 20 min (samples B1, B2, and B3) are displayed in Table 5. It was found that by increasing the annealing time from 10 to 20 min, the atomic percentage of S evidently decreases from 10.23% to 2.26%, the atomic percentage of Se increases from 36.24% to 46.44%, and the ratio of Se/(Se + S) significant increases from 77.99% to 95.36%. The ratios of Cu/(Zn + Mg + Sn) in all films were significantly larger than those in the precursor, and the ratios of Mg/(Mg + Zn) in all films significantly decreased from 0.19 to 0.12 with increasing annealing time from 10 to 20 min. These changes were ascribed the decrease of Mg content with the increase of the annealing time, as shown in Table 5. Figure 10 represents the elemental composition analysis of the films annealed for different times from 10 to 20 min. It can be clearly seen that the atomic percentages of Se increase, while the atomic percentages of S decrease with the increase of annealing time. Compared with the increase of Se content and the decrease of S content, the atomic percentages of other elements (Cu, Zn, and Sn) changed only slightly. Furthermore, the changes were irregular and almost negligible, and the changes had less effect on the crystal quality of CMZTSSe films. According to the analysis of elemental composition, the change of selenization time mainly affects the atomic percentages of Se and S elements, and has little effect on other elements.


**Table 5.** EDS results of the CMZTSSe films annealed for different times from 10 to 20 min.

**Figure 10.** EDS composition analyses of CMZTSSe films annealed for different times.

Microstructure, composition, and grain size have great influences on the crystal quality of CMZTSSe, and the bandgaps are also crucial. According to the previous analysis, the change in the ratio of Se/(Se + S) with the increase of selenization time significantly influences the crystal quality, composition, and grain size. The effect of annealing time on the optical bandgaps of the CMZTSSe films has been evaluated by a UV-vis-NIR spectrophotometer. The theoretical basis of bandgap calculation is consistent with Formula 1. As shown in Figure 11, the bandgaps of the CMZTSSe films show a declining trend (1.06–0.95 eV) with increasing annealing time from 10 to 20 min. The illustration of Figure 11 displays the dependence of bandgaps on selenization time for CMZTSSe films. We can clearly see that the values of bandgap are 1.06, 1.02, and 0.95 for sample B1, sample B2, and sample B3, respectively. Combined with the analysis results of XRD, Raman, and EDS, the decline in bandgaps with the increase of the selenization time is ascribed to the increase of elemental Se.

As shown in Table 6, the impacts of selenization time on the conductivity, carrier concentration, and mobility of CMZTSSe (samples B1, B2, and B3) were investigated by Hall measurements at room temperature. It was observed that the p-type conductivity of CMZTSSe was not changed with the increase of annealing time from 10 to 20 min. As shown in Table 6, when the annealing time increased from 10 to 15 min, the hole concentration of the CMZTSSe films increased obviously from 9.22 <sup>×</sup> <sup>10</sup><sup>17</sup> to 6.47 <sup>×</sup> 10<sup>18</sup> cm−3, the resistivity decreased from 6.18 <sup>×</sup> 10<sup>0</sup> to 2.85 <sup>×</sup> 10−<sup>1</sup> <sup>Ω</sup>·cm, and the mobility decreased from 1.09 <sup>×</sup> 10<sup>0</sup> to 3.31 <sup>×</sup> 10−<sup>1</sup> cm2V−1s−1. Combined with the analysis results of SEM, by increasing annealing time from 10 to 15 min, the grain size of CMZTSSe becomes bigger and the surface becomes smooth and hole-free, which leads to the improvement of electrical performance. When the annealing time increases from 15 to 20 min, the crystallinity of CMZTSSe films deteriorates, and hence the hole concentration and resistivity decrease to 5.54 <sup>×</sup> 1017 cm−<sup>3</sup> and 1.10 <sup>×</sup> <sup>10</sup><sup>2</sup> <sup>Ω</sup>·cm, respectively. It was found that when the selenization temperature and selenization time are 530 ◦C and 10 min, the best electrical properties of the films are obtained.

**Figure 11.** The plot of (α*h*υ) <sup>2</sup> vs. *h*υ for the absorption spectra. Inset: Bandgap variation as a function of the selenization time.

**Table 6.** Electrical properties of the CMZTSSe films annealed for different times from 10 to 20 min.


#### **4. Conclusions**

In summary, we have successfully fabricated pure-phase CMZTSSe films at different selenization temperatures and selenization times through the sol–gel method. It was found that the properties of CMZTSSe films are greatly affected by selenization temperature and selenization time. Combined with the results of XRD, Raman, XPS, and EDS, it is clear that single-phase CMZTSSe films have been synthesized at different selenization temperatures and times, and the content of Se increases gradually while the content of S decreases gradually with increasing selenization temperature and selenization time. The SEM results suggested that the crystal quality of CMZTSSe is the best at the optimal selenization condition of 530 ◦C for 15 min, where the grain size reaches 1.0–2.5 μm. In addition, the grain-boundary passivation due to the crystal quality improvement will result in the improvement of electrical performance. The CMZTSSe films with p-type conductivity and high hole concentration of 6.47 <sup>×</sup> 1018 cm−<sup>3</sup> were obtained by selenization at 530 ◦C for 15 min. The *E*<sup>g</sup> of CMZTSSe films is decreased from 1.04 to 0.93 eV with increasing selenization temperature from 500 to 560 ◦C. When selenization time is increased from 10 to 20 min, the *E*g of CMZTSSe can be adjusted from 1.06 to 0.96 eV. It is concluded that the structure, and optical and electrical properties of CMZTSSe will be optimal at an optimized selenization temperature and selenization time of 530 ◦C and 15 min, respectively, which will create an ideal absorber material for preparing higher efficiency kesterite solar cells.

**Author Contributions:** Conceptualization, Y.S. and Y.Z.; Writing-Original Draft Preparation, D.J.; Software, W.H.; Formal analysis, Y.S., Z.W. and L.Y.; Investigation, Y.S. and F.W.; Writing-Review & Editing, Y.S. and B.Y.

**Funding:** This research was funded by the National Natural Science Foundation of China under Grant No. 61,505,067, 61,605,059, 61,475,063, 61,775,081, 61,705,079.

**Conflicts of Interest:** The authors declare no conflict of interest.
