Introduction

The III–VI semiconductors have been the subject of many investigations due to their peculiar electrical and optical properties, and their potential applications in electronic and optoelectronic devices [1–4]. Among these semiconductors, γ-In2Se3 has attracted attention because it is suitable for use in photovoltaic applications [5]. In the recent years, many researchers have been interested in the synthesis of the nanoscale materials due to their unique properties and novel applications in optoelectronic and electronic devices [6–8]. Although some progress has been achieved regarding the growth and characterization of γ-In2Se3 epilayers, the γ-In2Se3 nanostructures have not been grown and investigated yet. The γ-In2Se3 nanostructures may show potential applications in optoelectronic device such as lasers, light emitting diodes (LEDs), and solar cells, due to their high surface-to-volume ratio.

When excess energy is supplied to a carrier by optical excitation or an applied electric field, the energetic carrier becomes hot. The hot carriers then relax toward less energetic state by two competing processes, namely scatterings with other carriers and emission of phonons [9]. The understanding of this energy relaxation process constitutes a direct probe of a very fundamental interaction in condensed matter physics, namely, the electron–phonon and electron–electron interactions. Also, the subject is of obvious technological significance since many devices work mostly in high-field conditions. High electric fields may lead to carrier heating and, consequently, transport effects related to the hot carrier distribution function. A knowledge of hot carrier relaxation mechanisms is thus essential not only for understanding the fundamental process in semiconductor materials but also for evaluating optical device performance.

In this study, the single phase γ-In2Se3nanorods on silicon (111) substrates were grown by metal-organic chemical vapor deposition (MOCVD). The excitation power dependence of photoluminescence (PL) in γ-In2Se3nanorods was studied. The high-energy tails of the low-temperature PL were characterized by effective electron temperatures which increase with increasing excitation intensity. It is found the main path of energy relaxation of the hot electrons in the γ-In2Se3nanorods is the LO-phonon emission.

Experiment

The γ-In2Se3nanorods were grown on Si (111) substrates by using an MOCVD system at atmospheric pressure with a vertical reactor. The liquid MO, a TMIn compound, and gaseous H2Se were employed as the reactant source materials for In and Se, respectively. The gaseous N2was used as the carrier gas in the process. The substrates used in this experiment were cut from a 6-inchp-type vicinal (111)-oriented Si wafer. Before the growth, Si substrates were baked at 1100 °C for 10 min in gaseous HCl and H2to remove the native oxide. After the thermal etching process, the reactor was cooled down to 425 °C and the γ-In2Se3started to grow. The gaseous flow rate of TMIn was kept at 3 μmol/min and that of H2Se was controlled at 40 μmol/min. The gaseous H2Se was mixed with 85% hydrogen and 15% H2Se. The γ-In2Se3nanorods were grown at 425 °C during a total growth time of 50 min. The structure of the γ-In2Se3nanorods was examined by the X-ray diffraction (XRD) in a θ–2θ geometry. The XRD measurements were performed by using the CuKα-radiation (λ = 1.541 Å) to test the phases of samples. PL was made using the Ar-ion laser operating at a wavelength of 514.5 nm. The room-temperature PL measurements were performed using a confocal microscopy. The collected luminescence was dispersed by a 0.75 m spectrometer and detected with a photo-multiplier tube (PMT).

Results and Discussion

The morphology of the grown γ-In2Se3 nanorods was investigated by the scanning electron microscopy (SEM). The cross-sectional image of SEM for the γ-In2Se3 nanorods is shown in Fig. 1, indicating a high density and narrow size distribution. The crystallographic face of each nanorod is shown in the inset of Fig. 1, revealing the hexagonal top end of the γ-In2Se3 nanorods. The inset of Fig. 2 shows the XRD pattern of γ-In2Se3 nanorods. A high intensity of the XRD pattern from the Si (111) plane was clearly observed at 2θ = 28.44°. Furthermore, the XRD reflection from the plane of γ-In2Se3 was also observed at 2θ = 27.59°, confirming the hexagonal single phase for the γ-In2Se3 nanorods [10]. The 300-K PL spectrum of the γ-In2Se3 nanorods is shown in Fig. 2. A clear PL peak was observed with the peak position of 1.95 eV, corresponding to the near band gap edge emission [11]. Observation of the room-temperature luminescence of the γ-In2Se3 nanorods indicates the good quality of our sample.

Figure 1
figure 1

The γ-In2Se3nanorods morphology obtained by the cross-section SEM image. The inset shows the top-view SEM image

Full size image
Figure 2
figure 2

Room-temperature PL of the γ-In2Se3nanorods. The inset shows the XRD pattern of the γ-In2Se3nanorods

Full size image

In the process of the hot PL, the photoexcitation creates energetic electrons in the conduction band, which relax toward less energetic state by transferring energy to the lattice (via the electron–phonon scattering) and other electrons (via the electron–electron scattering). If the electron–electron collision rate is larger than the phonon emission rate, then the non-equilibrium electron population in the electron gas relaxes toward a Maxwell distribution and can be characterized by anT e(T e) which is higher than the lattice temperature (T l) [12]. Figure 3(a–d) shows the high-energy tail of the 15-K PL in γ-In2Se3 nanorods with different excitation power densities. The spectra show that the high-energy tail of each PL decreases exponentially with photon energy, revealing that the PL is related to the hot carrier recombination. The high-energy tail of each PL in Fig. 3 can be analyzed by the function [6]:

(1)

where E 0 is the specific energy. With low excitation power, E 0 reflects the sample quality at low temperatures [6]. Under higher photoexcitation, E 0 can reflect the kinetic energy of the thermalized electrons and a well-defined T e can be extracted. We have fitted the high-energy tail of PL using Eq. 1, as shown by the solid lines in Fig. 3.

Figure 3
figure 3

Measured (open squares) and fitted (solid line) of the high-energy tail of the PL for different excitation power: (a) 353 W/cm2, (b) 530 W/cm2, (c) 707 W/cm2, (d) 1414 W/cm2

Full size image

The inverseT eversus the excitation power is plotted as the open squares in Fig. 4. The slope of the inverseT e, displayed as the solid line, corresponds to a value of 19 meV. To find out whether this energy is related to the phonon energy in γ-In2Se3nanorods, we performed the Raman scattering measurements. Figure 5is the Raman spectrum of γ-In2Se3nanorods, displaying a clear peak located at 152 cm−1, whose energy corresponds to ~ 19 meV. Thus, the energy extracted from the slope of the inverseT eis in good agreement with the phonon energy measured from the Raman scattering. This indicates the phonon scattering is very efficient in transferring energy from electrons to the lattice. In other words, the phonon emission is the dominant energy loss mechanism in the energy relaxation processes of hot electrons in γ-In2Se3nanorods.

Figure 4
figure 4

The temperature dependence of PL spectra in the γ-In2Se3nanorods. The inset shows the temperature dependence of peak position in PL. The solid line in the inset shows the fit according to Eq. 2

Full size image
Figure 5
figure 5

The temperature dependence of PL intensity in γ-In2Se3nanorods. The theoretical fit according to Eq.3is displayed as the dashed line

Full size image

To obtain the energy loss rate per electron from experiments, the power balance equations were used. As the steady-state electron population increases by increasing the excitation density, enhanced electron–electron scattering results in a larger fraction of the available energy being shared with the electron gas. Thus, the T e is determined by balancing the rate of generation for the energetic electrons with the rate of energy loss from the electrons to the lattice. For the photoexcitation, the pump power per electron P e given to the electron is [12]

(2)

where I is the laser power absorbed per unit area, d is the absorption length at laser energy, n is the carrier concentration, and W is the part of the photon excess energy obtained by electron. The carrier concentration n was obtained from the room-temperature Hall-effect measurements. The open square in Fig. 4 displays the T e as a function of the power input per electron (P e). If we assume the dominant process for this relaxation is through LO-phonon emission and T e is much larger than T l,then the energy loss rate per electron due to the LO-phonon scattering can be given by [13].

(3)

where τph is the effective phonon lifetime, E LO is the LO-phonon energy, , and K 0 is the modified Bessel function of the order of zero. In the steady state, the power input per electron P e is equal to the power loss to the lattice through phonon scattering. Taking values of 19 meV, 1.12×1016 cm−3, 2.12, 2.41 eV, 4.8 × 10−6 cm for E LO n W h ν0 d, respectively, the solid line in Fig. 4 displays the fitted T e with the power loss per electron. Good agreement between experiments and calculations indicates that the model based on the carrier scattering by LO-phonon is able to explain the measuredT e variation with excitation power. It demonstrates again that the LO-phonon emission is the dominant energy loss mechanism in the energy relaxation processes of hot electrons in γ-In2Se3 nanorods.

Summary

In summary, the γ-In2Se3nanorods were successfully grown on Si (111) substrates by using MOCVD. A clear room-temperature PL with the peak position of 1.95 eV was observed, corresponding to the near band edge emission. The high-energy tail of PL can be characterized by an effectiveT ewhich increases with increasing excitation intensity. The relationship between theT eand the electron energy loss rate can be explained by a model based on the carrier scattering by the LO-phonons.