Metal sulfide thin films are bulk semiconductors that have attracted much attention recently because of their excellent properties. Examples include CuS,^[1] Cu₂S,^[2] ZnS,^[3] CdS,^[4] SnS,^[5] FeS₂,^[6] Sb₂S₃,^[7] CuAlS₂,^[8] Cu₂BaS₂,^[9] CuInS₂,^[10] and Cu₂ZnSnS₄.^[11] The synthesis process can be tuned and easily controlled,^[12,13] and their performance is better than nanocrystal films. This is influenced by the surface ligands of the nanocrystal.^[14] Therefore, metal sulfide thin films are promising candidates in optoelectronic devices such as photovoltaic and photoconductive devices.

CuInS₂ is a representative metal sulfide with many attractive attributes, such as optimal bandgaps,^[15] a large absorption coefficient,^[16] and low toxicity.^[17] To develop optoelectronic devices based on CuInS₂ thin films, it is necessary to study the charge carrier generation, transportation, and recombination mechanisms. However, the understanding of the carrier dynamics in the CuInS₂ thin films remains limited. These materials have an amorphous structure,^[12,13] and the corresponding photo-physical properties are different from nanocrystals. In addition, the carrier behavior in the inorganic semiconductor film is affected by many factors. Due to these complexities, a single test method is insufficient to analyze this process.

Here, we have simultaneously applied femtosecond transient absorption and nanosecond transient photocurrent to probe the carrier behavior in the CuInS₂ thin films. The results provide a better understanding of carrier behavior in the CuInS₂ film and aid in determining the intrinsic physical characteristics that are independent of the external condition.

CuInS₂ thin films were synthesized as described in the previous report.^[13] 2 mmol S powder, 2 mL 1-butylamine, 0.6 mL thioglycolic acid, 1 mmol Cu(Ac)₂, and 1 mmol In(Ac)₃ were added into 5 mL ethanol under room temperature to form stable CuInS₂ solution. The films were fabricated by spin-coating the solution on indium tin oxide (ITO) and quartz glass substrates followed by 350 °C sintering for 3 min, which was repeated several times. The device structure was designed by following a planar ITO/CuInS₂ film/Al structure as previously reported.^[18] The nominal device area was 4 mm², which was defined by the overlap between the anode and cathode. An ohmic contact was formed between ITO anode and CuInS₂ film,^[19] whereas the contact between CuInS₂ film and the Al cathode was a Schottky contact.^[20] Steady-state absorption measurements were performed with an ultraviolet–visible (UV–Vis) absorption spectrometer (Purkinje, TU-1810 PC). The crystallographic information on the samples was determined by a Bruker D8 advance x-ray diffractometer (Cu Kα, λ = 1.5418 Å, 40 kV, and 30 mA). The femtosecond transient absorption (TA) technique has been described in a prior report.^[21]

The nanosecond transient photocurrent was composed of a nanosecond laser (Quantel, QS450-2W3W) and oscilloscope (Tektronix TDS 3032C). To ensure uniform distribution of the photo-generated carriers, we used an optical parametric oscillator (OPO) module (MagicPRISM VIS) to obtain the excitation light pulse with a wavelength of 680 nm, where the absorption value of the sample was approximately 0.33.

Figure 1(a) shows the absorption spectra of the CuInS₂ thin film with a band-edge located near 830 nm. The spectra show no obvious features in the frequency domain.^[22] The cross-sectional scanning electron microscope (SEM) image in Fig. 1(a) shows a film thickness of 300 nm. To verify the phase purity, x-ray diffraction (XRD) is performed and a wurtzite structure^[23] is obtained for the CuInS₂ thin film, as shown in Fig. 1(b).

Fig. 1.

	Figure Option ViewDownloadNew Window
	Fig. 1. (a) The absorption spectrum of the CuInS2 film. (b) The XRD patterns of the CuInS2 thin films. Inset: SEM of the CuInS2 film.

Figure 2(a) shows a map of transient absorption data of the CuInS₂ thin film excited at 400 nm. Comparing with the steady state absorption, there is a positive signal from 550 nm to 750 nm and a minor negative signal above 550 nm, corresponding to the excited state absorption (ESA) and the ground bleaching, respectively. The ESA peak is gradually redshifted, and its intensity decreases with time. This redshift is attributed to the change in carrier energy, whereas the intensity decrease is assigned to population relaxation, diffusion, and recombination of photo-generated carriers.^[24,25]

Fig. 2.

	Figure Option ViewDownloadNew Window
	Fig. 2. (a) Map of transient absorption data. (b) Time-dependent carrier energy and population.

Figure 2(a) shows that the final energy of the photo-generated carrier is about 1.7 eV, which suggests that the energy of the photo-generated carrier varies about 0.2 eV during relaxation. To further analyze the carrier properties, the time-dependent variance of carrier energy (E–T curve) and intensity (I–T curve) of ESA are generated, as shown in Fig. 2(b). The E–T dynamics decay rapidly versus I–T. We have employed a multi-exponential function to analyze the curves and determine their average decay lifetime. The fitted results from complex and non-exponential decay behaviors are shown in Fig. 2(b), where the lifetimes of E–T and I–T are determined to be 2.00 ps and 1.49 ps, respectively.

Next, the nanosecond transient photocurrent^[26] method is employed to further investigate the dynamic behavior of the carriers. A bias electric field applied to the CuInS₂ film is used to extract the carriers. The illumination intensity-dependent photocurrent transients are shown in Fig. 3(a). The amplitude of the photocurrent increases almost linearly with the light intensity, as evidenced from the inset in Fig. 3(a). A saturation feature is obtained when the light is stronger than . This suggests that the loss of the carrier originating from the light intensity could be ignored with weak illumination.

Fig. 3.

Figure Option ViewDownloadNew Window

Fig. 3. (a) The illumination intensity-dependent and (b) applied electric field-dependent transient photocurrent curves. Insets: corresponding maximum (a) illumination intensity-dependent and (b) applied electric field-dependent transient photocurrent. (c) The concentration of extracted charge carriers and (d) α versus applied electric field measured under different illumination intensities ( and ), respectively. The illumination intensities are 8.25, 5.31, 2.99, 1.61, 0.96, 0.56, 0.32, 0.17, 0.062, 0.035, 0.019, and 0.0083 (in units of ) in (a) and in (b). The applied electric fields are in (a) and 6.33×107, 5.70×107, 5.06×107, 4.43×107, 3.80×107, 3.16×107, 2.53×107, 1.90×107, 1.27×107, and 0.63 ×107 (in units of ) in (b).

Figure 3(b) shows the electric field-dependent current density j₀ when the illumination is . j₀ is detected by the oscilloscope, even though there is no bias electric field E on the CuInS₂ film. This suggests that j₀ mainly originates from temperature-induced carrier diffusion.^[27] Under an applied E, j₀ is enhanced due to mobility of the photo-generated charge carriers, as shown in Fig. 3(b). Notably, both decay curves of j₀ in Figs. 3(a) and 3(b) show a non-exponential behavior because the diffusion, migration, and recombination of carriers participate in the relaxation behavior of j₀. In addition, the concentration of extracted charge carriers N also increases with j₀, as seen in the time integrated plots in Figs. 3(a) and 3(b). The evolution of N under different E values is given in Fig. 3(c) at the light intensity of and , respectively. Both the corresponding values of N gradually increase with E and reach saturation at each irradiation intensity. The number of photo-generated carriers in the film is constant under a fixed illumination intensity, and can be fully extracted when E is sufficiently high. The photo-generated carriers recombine under stronger light. However, the E-dependent evolution of N at an illumination intensity of is similar to that of , suggesting that it is insensitive to the number of charge carriers in the film.

To further discuss the role of E in the nanosecond transient photocurrent test, we define as the fraction of charge carriers driven by E, as shown in Fig. 3(d) as a function of E. The α gradually decreases with E and is almost zero when E increases to , which implies that the photo-generated charge carriers in the CuInS₂ film may be extracted more effectively under a higher electric field. However, α only undergoes a minor change with illumination intensity, which is likely due to photo-generated carrier recombination.

Figure 4(a) further shows the evolution of the photo-generated charge carriers versus different absorbed photon numbers (APNs) under an E of and , respectively. The N linearly increases with APN when the APN is lower than 1.79 ×10¹⁴ mm⁻³. At higher values, the variance of N induced by APN becomes smaller and eventually saturates as APN increases further. This is likely to be due to incomplete extraction of photo-generated charge carriers in the CuInS₂ thin film, as some of them may be lost due to recombination or collision annihilation.

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. APN-dependent (a) N and (b) APCE and PCRE under different applied electric fields ( and ). (c) The extracted velocity of charge carrier v and (d) γ versus N under an applied electric field of and , respectively.

The evolution of N at different APN values is similar as E increases from to , suggesting that this phenomenon is insensitive to E. APN-dependent absorbed photon-to-electron conversion efficiency (APCE) is shown in Fig. 4(b). At a fixed E of , the initial APCE is close to 81% and does not vary with APN. The APCE rapidly decreases after the APN increases to ( . This can be attributed to the loss caused by carrier recombination.

We also studied the photo-generated carrier recombination efficiency (PCRE) given by (81%-APCE)/81%; the results are shown in Fig. 4(b) with E of . The PCRE gradually increases with APN values over 9.92×10¹³ mm⁻³, indicating that the loss of carriers is due to the carrier–carrier collision at high carrier concentrations. The APCE is under 10% when E decreases to . The APCE is strongly dependent on E because the value of E determines the number of extracted free charge carriers. The PCRE at is similar to the value at , as shown in Fig. 4(b). This suggests that the recombination process in the film is affected by the APN but not by E.

The lifetime of j₀ corresponds to the time when j₀ decreases to the 1/e of its amplitude at t = 0. We also define to estimate the extraction velocity of charge carriers, where N_e is the number of extracted charge carriers during the period of . The v gradually increases with E and N, as shown in Fig. 4(c). However, the evolution of v at is similar to that at . To ascertain the reason behind the similarity of v at different E values, we define , as shown in Fig. 4(d). The γ is almost independent of N below 1.79 ×10¹³ mm⁻³, which implies that the carrier mobility during that process is constant. When N is higher than 1.79 ×10¹³ mm⁻³, γ begins to decrease with N at fixed E, suggesting that the variance of v induced by N gradually decreases. In this situation, migration plays a negative role in the extraction process of charge carriers because the carriers in the CuInS₂ film are congested, which decelerates directional movement of carriers and leads to carrier–carrier collisions. The evolution of γ at is similar to that at , suggesting that the evolution of γ at different N values is almost insensitive to E, even though it determines the amplitude of γ.

In summary, we have investigated the carrier kinetic behavior in CuInS₂ films based on femtosecond transient absorption and nanosecond transient photocurrent methods. Femtosecond transient absorption confirms that the carrier relaxation is complex and accompanied by a change of carrier energy. The nanosecond transient photocurrent analyzes the role of bias electric field and illumination intensity in the extraction process. Carrier recombination occurs when the illumination is higher than a certain threshold. The corresponding loss further increases with the illumination intensity. The recombination percentage is almost independent of the bias electric field. The carriers are extracted by the bias electric field, which improves the efficiency of photon conversion to charge carriers and enhances the velocity of the extracted charge carriers. These results provide insights into the dynamic behavior of charge carriers in inorganic thin films, which is essential for the development of optoelectronic devices.