In this experiment we cut FTO glass into small pieces with size of 1× 2 cm². We put FTO glass soak in the detergent for 3 hours in order to eliminate contaminant on FTO more thoroughly. Then put the substrate into the detergent, deionized water and ethanol for 1 hour by ultrasonic cleaning respectively. Finally the FTO substrates were stored in a beaker with anhydrous ethanol.

A three-electrode system was used with FTO as working electrode, a platinum (Pt) gauze as counter electrode and Ag/AgCl as reference electrode. The electrolyte contained CuSO₄, ZnSO₄, SnSO₄ and with complexing agent-Triethanolamine (TEA) and sodium citrate. First, sodium citrate was infunded in a beaker with a stirring bar at room temperature, and the bar stirred vigorously. Second, CuSO₄ was dropped into the liquid tardily, a step which changed the liquid’ s color immediately, from transparent to deep blue in a few seconds. Then, ZnSO₄, SnSO₄ and TEA were added to the blue liquid in sequence. Finally, the deionized water was added and the color of the solution turned from deep blue to light blue. Electrodeposition was carried out in a potentiostatic mode at, -1.3 V vs Ag/AgCl respectively. As a result, the Cu-Zn-Sn precursors were successfully prepared on FTO substrates.

According to the area of the sample put 15 mg· cm^-2 sulfur powder into the vial, and put the precursor films on the glass slide. Then the precursor films and vial filled with sulfur powder were introduced into the quartz tube. The quartz tube was sealed and pumped to 0.15 Pa, then refilled with argon to 0.85 atm. The left temperature zone was heated up to 550 ℃ in half an hour, and held at 550 ℃ for 15 min. The right temperature zone was heated up to 360 ℃ in 20 min, and held at 360 ℃ for 45 min. Finally, two zones were cooled down to room temperature naturally.

The chemical composition and morphologies of the CZTS films were determined by scanning electron microscope (SEM) using a ZEISS EVO 18 with energy dispersive spectroscopy (EDS). The structural properties and crystallization of the films were carried out using an X-ray diffraction (XRD) (Rigaku, Cu-Kα radiation). The samples were scanned between 10° and 80° with a step size of 0.01° . The micro-Raman spectra of the CZTS films were recorded using a scanning probe microscope (Alpha SNOM, WITec, Germany) equipped with a spectrometer and CCD detector. An Nd YAG linearly polarized laser (Verdi, Coherent Inc. Santa Barbara, CA) with 532 nm wavelength and at 10~30 mW laser power to the coupling optic fiber was used for Raman excitation. Raman spectra were collected in a black scattering geometry with a grating of 1 800 lines· mm^-1. A 50× objective with 0.75 NA (Nikon) was used to focus the laser beam onto the specimen and collect Raman signals. The optical band gap energy of the sulfurized films was measured using UVevis spectrometer (UV-2100) in the 200~900 nm wavelength range. The J-V characteristics of CZTS PEC devices were measured using an Electrochemical Analyzer (CH Instruments, Austin TX). The most important part of the PEC device is the Polytetrafluoroethylene (PTFE) boxes in which a three-electrode configuration with a Pt-coil counter electrode and a saturated calomel reference electrode is placed and connected to an Electrochemical Analyzer. The inside test units were submerged into the electrolyte. The electrochemical cell was placed in front of a 150 W Newport lamp with an AM 1.5D filter with a shutter for switching the light.

The SEM image and EDS characterization of a sulfured CZTS thin film are shown in Figure 1 and Table 1, respectively. The surface of the sulfured film is uniform and possesses several long particles. In order to determine what kind of particles were formed on the surface of the CZTS films, EDS analysis was used to check different area in a scale of 1 mm². The EDS confirmed the atomic composition of each element. From the results shown in the table 1, Cu composition is a little low and the ratio of Cu/(Zn+Sn) is 0.69, where the Zn concentration is a little low with Zn/Sn ratio of 0.85. The ratio of S/metal is 1.01. The measure ratio of Cu:(Zn+Sn):S is 0.82:1.18:2, which is close to the stoichiometric CZTS formation represented as Cu₂ZnSnS₄.

Fig.1

	Figure Option ViewDownloadNew Window
	Fig.1 SEM plane-view images of Cu2ZnSnS4 structure annealed film at different magnifications

Table 1

Table 1

Table 1 Compositional analysis of the typical CZTS sample sulfured at 550 ℃ Element Cu/(at%) Zn/(at%) Sn/(at%) S/(at%) Cu/(Zn+Sn) S/(metal) -1.3 V, sulfurized 20.21 13.54 15.94 50.30 0.69 1.01

Table 1 Compositional analysis of the typical CZTS sample sulfured at 550 ℃

X-Ray diffraction (XRD) patterns of as-deposited and sulfurized CZTS films on FTO substrates are shown in Figure 2. The existence of elementary Cu, Zn, or Sn and several secondary phases were observed in the deposited films. The effect of sulfuring was distinctive in nature to obtain a better phase purity of electro deposited CZTS. The annealing was found to decrease all elementary phases of Cu, Zn and Sn as well as allotropes of sulfur. When the film was annealed at 550 ℃, the prominent (112), (220) and (312) peaks corresponding to crystalline phase of CZTS appeared. This film was purer, except for the coexistence of SnS₂. The Three prominent peaks are located at 2θ =28.5° , 47.5° and 56.0° , respectively.

The size of CZTS crystals can be calculated using well known Scherrer equation.

D=Kλβcosθ

Fig.2

	Figure Option ViewDownloadNew Window
	Fig.2 XRD patterns of as-deposited and sulfurized CZTS sample on FTO at T=550 ℃ during 0.5 h

Where D is the grain size, K=0.94, λ is the wavelength of Cu-Kα X-ray, β is the full width at half maximum of the selected peak, θ is the Bragg angle of the selected peak. The calculated crystalline size by Scherrer equation is between 35 and 45 nm.

However, in the XRD patterns, peaks corresponding to different phases may complicate and overlap the patterns. A complex peak appears as a broad one. Like these phases cannot be distinguished by XRD patterns. Because of this limitation, the phase identification in complex systems such as quaternary compound semiconductors may not be possible by XRD studies alone. Raman spectroscopy is often complementary to XRD techniques and has been included in the following subsection.

Raman spectra is commonly used to observe vibrational, rotational, and other low-frequency modes in a system to provide fingerprints of molecules. The main peak of CZTS sample located at 338 cm^-1 with strong intensity was compared to other two lower peaks located at 290 and 377 cm^-1 respectively, which is in agreement with other reported from kesterite CZTS. No Raman peaks for Cu_2-xS and SnS₂ with amode at 315 cm^-1 were found. It means the content of SnS₂ is very less. The two most intense peaks at 290 and 338 cm^-1 assigned as the dominant non-polar A-symmetry modes, is consistent with earlier reports^{[26, 27, 28]}. Raman peak at 377 cm^-1 is identified with B-symmetry mode, while the peak at 366 cm^-1 is identified with E-symmetry mode^[29]. In general, the Raman spectroscopy agreed very well with the XRD measurement: the prepared CZTS film shows a strong phase of single chalcopyrite structure without much second phases.

In order to study the band gap of CZTS thin films, the optical properties of CZTS on FTO were measured via UV-Vis absorption spectroscopy. Figure 4 (a) shows the absorption coefficient (α ) as a function of wavelength or photon energy can be determined. The energy gap can be obtained using transmittance data from the absorption spectra by the following equation: Lambert-Beer Law.

α=k(hν-Eg)hν

Fig.3

	Figure Option ViewDownloadNew Window
	Fig.3 Raman spectra of annealed CZTS sample in sulfur at T=550 ℃

Fig.4

	Figure Option ViewDownloadNew Window
	Fig.4 (a) NIR-Vis absorption spectrum of CZTS films; (b) Plot of (α hν )2vs. photon energy (hν ) of the as re-crystallized CZTS

Where α is absorption coefficient, E_g is the band gap and k is a constant. Extrapolation of a plot of (α hν ) vs. hν curve to zero, the calculated of E_g is about 1.45 eV, which is suitable for device fabrication.

Figure 5 shows PEC performance of CZTS films (front and rear illumination) was investigated by linear sweep voltammetry (LSV) range 0 to -0.3 V. The big up climbing of photocurrent occurred when the CZTS film was illuminated under front/back irradiation. Upon switching to the dark, the photocurrent rapidly dropped. Photoelectrons transferred from conduction band of CZTS to the redox species through the interface of semiconductor electrode/electrolyte can be interpreted. Figure 5 also shows rear illumination caused a similar photocurrent density (0.18 mA· cm^-2) compared with (0.16 mA· cm^-2) front illumination at a biased potential of -0.2 V. And the voltammogram upon rear illumination was found to have lower photocurrent drop than front illumination after photocurrent reached the maximum height after illuminated. This result implicates that degradation might take place on the surface of CZTS film upon illumination. In general, light current is produced by FTO/CZTS under front/back irradiation, and the photocurrent flows in the same direction in both cases. This is a very promising result for bifacial solar cells.

Fig.5

	Figure Option ViewDownloadNew Window
	Fig.5 PEC of CZTS on FTO conductive glass glass