In this chapter, novel optical properties of ZnSe0.8Te0.2/ZnSe multiple quantum wells (MQWs) structures were studied by excitation power dependent photoluminescence (PL) and temperature (T) dependence time resolve photoluminescence. A significant blueshift of PL energy and long exciton lifetimes were observed. The T dependence of radiative efficiency was found to be governed by the dissociation of spatially separated exciton at low T region. The thermal quenching of the PL intensity was found to depend crucially on the excitation power.
The 10-period ZnSe0.8Te0.2/ZnSe MQWs, were grown on the GaAs substrates by molecular beam epitaxy. A 150 nm-thick ZnSe buffer was grown before the deposition of the MQW. The cell temperatures of Zn, Se and Te were set at 300 oC, 170 oC and 290 oC, respectively. The substrate surface temperature was kept at 310 oC throughout the growth of the MQWs. The Te concentration was determined by an energy dispersive X-ray diffraction measurement using a ZnSe0.8Te0.2 epilayer grown under the same conditions as the MQW structures. The thickness of the barrier layer (ZnSe) was kept at 20nm, and the well-widths (Lw) for the four 10-period MQWs are 1 nm, 2 nm, 3nm and 5 nm, respectively. A pulsed GaN diode laser (396 nm) with pulse
duration of 50 ps and repetition rate of 2.5 MHz was used as an excitation source for the time-resolved measurement. The peak power of the pulse was estimated to be below 0.1 mW. The transient PL spectra were analyzed by a high-speed photomultiplier tube, followed by a personal computer plug-in time-correlated counting card. The overall time resolution of the detection system was about 300 ps.
Bright luminescent energies from yellow to green spectral region were detected from the samples at low T. Figure 1(a) shows the low-T PL spectra of the MQWs, pumped under an excitation power below 0.1 mW. A significant blueshift of PL peak energy with decreasing Lw is attributed to the quantum confinement effect. The large PL linewidth of the MQWs was attributed to the radiative recombination between the holes trapped in the short-range potentials of the isoelectronic impurities in ZnSe0.8Te0.2 and the electrons bound in ZnSe by the Coulomb interaction.
In Fig. 1(c) shows the evolution of PL peak intensity as a function of time. The decay profiles could be fitted by a single time constant using
IPL( )t =I0exp( /−t τPL) (1) where ( )IPL t and I are the PL intensity at time t and 0, respectively. 0 τPL is the
PL decay time. The spectral distributions of PL decay time (τPL) were shown in Fig.
1(b). The τPL values of the ZnSe0.8Te0.2/ZnSe MQWs are monotonically decreasing
only determinant of the recombination linewidth, since inhomogeneous broadening is also relevant [15]. The different aggregates of Te-clusters is responsible for inhomogeneous broadening. The τPL decreases with the increasing Lw. The increase in Lw reduces the wavefunction overlap between the spatially separated electrons and holes [43], therefore the decay rate decreases. A decay time of 30 ns has been reported for a bulk ZnSe1-xTex with x = 0.12 [15]. The long-lived exciton in bulk ZnSe1-xTex is due to the suppression of hole wavefunction by Te-clusters [44]. The further enhancement about an order of magnitude in the τPL’s of the QW structures is ascribed to a more spatial separation of the carriers that were confined in different layers. The observed lifetimes are longer than that was reported for type-II ZnSe/BeTe superlattice structures [45]. This indicates that Te doping can be a factor for the further reduction of decay rate by modifying the hole wavefunction in the ZnSeTe quantum structures.
We further confirm the type-II nature of the emissions in the MWQs. Figure 2 plots the evolution of PL peak energy as a function of excitation power for the MQW (Lw = 3 nm). The PL bands blueshift markedly as the excitation power increases. This phenomenon, which is usually interpreted as being caused by the carrier-induced band-bending effect at the heterointerfaces [5,6]. As the excitation density increases, the increase in population of spatially confined electron-hole pairs strengthens the
band-bending effect at the heterointerfaces. The quantization energy was found to increase in proportion to the cube root of the excitation power [46]. This is in good agreement with the experimental data (inset of Fig. 2). Hence, the PL of the ZnSe0.8Te0.2/ZnSe MQWs is attributable to the radiative recombination of holes localized in the ZnSe0.8Te0.2 layers and electrons in the ZnSe barrier region.
The T-dependent PL efficiency was examined by the T-dependent PL intensity (IPL) and lifetime. Figure 3(a) shows the τPL (solid circles) measured at the PL peak energy as a function of T for the sample with Lw = 1 nm. The lifetime increases initially from low T to 100 K (from 160 ns to 184 ns), and decreases monotonically above 120 K. The radiative (τr) lifetime and nonradiative (τnr) lifetimes were estimated by a combined analysis of T-dependent IPL and τPL [47]. Considering only radiative and nonradiative recombination process in the system. The internal radiative yield, η τ= nr /(τnr +τr)=IPL( ) /T IPL(0), rules the variation of I versus T, where PL the ( )IPL T and IPL(0) are the PL intensity at temperature T and 10 K, respectively.
The radiative efficiency is assumed to be unity at low temperature (30 K) because the excitons are strongly localized by the Te-clusters, thus, the migration of excitons to the nonradiative centers is inhibited [15]. Since the measured τPL can be given by
1 1 1
PL r nr
τ− =τ− +τ− , we can obtain the radiative lifetime as τR =τPL/η. The decrease in
at low T, dominating the recombination at over 120 K. The increase in τr can be understood as a consequence of ionization of weakly bound electrons from the strongly localized holes. Therefore, the dependence of τr on T can be fitted by the formula [8]:
τr( )T =τr(0) /[1−Cexp(−ε/k TB )] (2) where τr(0) is the radiative time at the low T limit, C is a constant and ε is a characteristic energy that is of the order of the electron-hole binding energy. The obtained binding energy, indeed, depends sensitively on the T range chosen for fitting.
The best fit yields a characteristic energy of 12 ± 3 meV for the binding of indirect excitons, where the error bar covers the fitting results for the different range of T.
Such a low exciton binding energy (Eb) is expected for type-II MQW structures due to the spatial separation of electrons and holes. Figure 3(b) shows the Arrhenius plot of the normailized PL peak intensity for the same sample. The solid line is fitted by the equation：
IPL( )T =IPL(0) /[1+CAexp(−EA/k TB )] (3) where IPL(0) is the PL intensity at 0 K, CA is fitting constant, kB is the Boltzmann constant and EA is activation energy. We fitted the experimental valves for the low T range where the excitons are not totally dissociated (below 100 K). The obtained activation energy E (18 meV) is comparable to (E ) obtained above.
We also plot the T-dependences of IPL (Lw = 1 nm) at a pumping power of 40 mW and below 0.1 mW in Fig. 4. The data were fitted by Eq. (3). The value of EA is larger than the room temperature thermal energy and approximately twice that was obtained from the low-power excitation. It can be explained by the aforementioned band-bending effect, which strongly modifies the electron-hole wavefunction overlap.
In the high excitation case, the wavefunctions of the carriers are expected to be strongly confined in the bent zones, which results in a strongly bound electron-hole pairs. Indeed, a decrease of three orders of magnitude has been reported for the recombination lifetimes of ZnSe/BeTe superlattices under high excitation [45].
Notably, the PL efficiency is mostly kept in the high excitation case. A drop of only about an order of magnitude in IPL was found from low T to 300 K. In contrast, the PL efficiency decreases by at least two orders of magnitude in the weak excitation case and is hardly detected at room temperature. The indirect-exciton recombination rate enhances under a high excitation density, based on the band-bending model. The effect plays an important role in the observed high emission efficiency.
In conclusion, the tunability of the emission energy, oscillator strength and PL efficiency by varying the well thickness and excitation density was demonstrated in the strained ZnSe0.8Te0.2/ZnSe MQWs. The type-II nature of the recombination was
the binding energy of the indirect excitons is determined as 12 meV for the thinnest sample. A reduction in PL efficiency was found to be greatly suppressed by employing a high excitation power.
Fig. 1 (a) Low-T PL spectra of ZnSe0.8Te0.2/ZnSe MQWs. (b)Spectral distributions of the decay times for the ZnSe0.8Te0.2/ZnSe MQWs. (c)Evolution of PL peak intensity as a funtion of time.
Fig. 2 PL spectra at various excitation powers for the sample with Lw = 3 nm. The inset plots the dependence of the PL peak energy on the cube root of the excitation power.
Fig. 3 (a) T-dependence of τPL (solid circles), τr (open circles) and τnr (open squares).
The solid line represents the fit of τr to Eq. 2, yielding a characteristic energy of 12±3 meV. (b) Arrhenius plot of the normalized PL peak intensity.
Fig. 4 Arrhenius plot of the normalized PL peak intensity for the ZnSe0.8Te0.2/ZnSe MQW (Lw = 1 nm), under excitation powers of 40 mW (solid triangles) and below 0.1 mW (solid squares). Activation energies with the fitting errors were also shown.
Chapter 4
ZnTe/ZnSe quantum dots
In this chapter, the optical properties of type-II semiconductor ZnTe 0-D quantum dots (QDs) were investigated by temperature-dependent and time-resolved photoluminescence (PL) spectroscopy. The initial decrease then increase with temperature for the full width at half maximum (FWHM) of PL is attributed to the hole thermal escape from the smaller QDs then transfer and re-capture to the neighboring-larger QDs. The non-mono-exponential decay profiles reflect the processes of carrier transfer and recapture. We show that the Kohlrausch’s stretching exponential well fits the decay profile of ZnTe/ZnSe QDs.
The samples studied in this chapter were grown on GaAs (100) substrates with a Vecco Applied EPI 620 molecular beam epitaxy (MBE) system. Solid sources Zn, Te, and Se were used for the growth of self-assembled ZnTe QDs and ZnSe buffers. Prior to the growth procedure, GaAs (100) substrate was etched in a H2O2 : NH4OH : H2O (1:5:50) solution for one minute at room temperature, rinsed in flowing de-ionized water about two minutes and dried with high purity N2. The substrate temperatures were set at 300 ℃. The growth process started with several monolayers (MLs) of
conventional MBE. Immediately after the deposition of ZnSe buffer layer, the alternating supply method of ZnTe growth was performed. The root mean square roughness of the ZnSe buffer layer, determined from atomic force spectroscopy, is approximately 0.5 nm. The coverage of the single ZnTe QDs layer, grown on the flat ZnSe buffer layer, was varied from 1.8 to 3.0 ML. A 50 nm ZnSe capping layer was grown on the QDs for optical measurements. The strong PL was observed even in the 3.0 ML ZnTe QDs sample. It implies less defects in this sample. The excitation source for the conventional PL spectroscopy was the 325 nm-line of a He-Cd laser and the emissions were analyzed using the SPEX 1403 double grating spectrometer in conjunction with a thermoelectrically cooled photomultiplier tube. A pulsed GaN diode laser (405 nm) with pulse duration of 50 ps and a repetition rate of 2.5 MHz was used as an excitation source for the time-resolved measurement. The PL decay spectra were analyzed by a high-speed photomultiplier tube, followed by a personal computer plug-in time-correlated counting card. The overall time resolution of the detection system was about 300 ps.
Figure 1 shows the low-temperature PL spectra of ZnTe QDs with different values of coverage. The 1.8 eV to 2.2 eV energy emission band is sensitive to the change of coverage. The emission energies are lower than that (2.4 eV) of the ZnTe epilayer because of the type II band alignment of ZnTe/ZnSe QDs as reported
previously [29]. For the ZnTe/ZnSe QDs, holes are confined in ZnTe and electrons are localized in ZnSe with the Coulomb attraction from holes. The red-shift in energy is attributed to the decrease in the quantum confinement of the holes in ZnTe. Noted that there exists a critical coverage (near 2.4 MLs) for which the slope of the peak energy versus coverage changes, as shown in the inset of Fig. 1. This indicates a change from the 2-D layer growth to 0-D ZnTe QD formation. It is confirmed by the atomic force microscopy (AFM) measurements and reflection high energy electron diffraction (RHEED) patterns. In the case of type I CdSe/ZnSe QDs, smooth slope changes to abrupt slope as the growth varying from 2-D layer to 0-D dot [48]. The abrupt slope is attributed to the sudden increase in QD volume. It further results in a faster decrease in energy with the coverage MLs. However, for the type II ZnTe/ZnSe QDs, the formation of QD results in a decreasing electron-hole wave-function overlap and the exciton binding energy. Therefore, the decreasing of emission energy with coverage for 0-D QDs is not as fast as the 2-D layers. For the 3.0 ML sample, in addition to the main peak, a weak PL structure appears at 2.2 eV. It is attributed to the emission from the underlying wetting layer.
The temperature dependent PL spectra of 2.6 ML sample is shown in Fig. 2 (a).
The peak energy exhibits red shift with increasing temperature. Whereas, the full
The FWHM as a function of the sample temperature is shown in Fig. 2(b). The initial decrease in FWHM at low temperature is attributed to the exciton thermal activation and escape from smaller QDs then recaptured by the larger QDs [49]. It results in a quench of the higher energy part of the PL spectra, which is contributed by the smaller QDs. The broadening of FWHM at high temperature regime is due to the exciton-longitudinal optical (LO) phonon interaction and can be fitted by line-width broadening model [50].
Figure 3 reveals the relation between PL intensity and inverse temperature. The dotted line is fitting curve using the equation
1 1 2 2 0 K, and Ea1 and Ea2 are the activation energies. The smaller activation energy is due to the activation of smaller QDs. The higher activation energy is attributed to the dissociation of exciton. This phenomenon corroborates the result of FWHM versus temperature plot, because the lower activation energy Ea1 is about several meV which is very close to the temperature where the minimum FWHM and exciton activation/escape become significant.
The coverage dependence of the decay characteristics of the PL signal from
ZnTe/ZnSe QD samples detected at their PL peak energies are shown in Fig. 4 with semi-log plot. The non-linear dependence is found. Such non-linear dependence could not be fitted by a mono-exponential function. As discussed in above mentioned temperature dependent FWHM and PL intensity, the PL spectra involve emissions from smaller and larger QDs and exciton transfer between them. Recently, the hopping-transport model has been proposed as the origin of the stretched exponential relaxation in complex condensed-matter systems [28]. Therefore, we propose that the
decay profiles should be better fitted by the stretched exponential.
The stretched exponential, known as the Kohlrausch’s law, is the following equation:
IPL( )t =I0exp{ ( /− t τPL) }β (2) where ( )IPL t and I are the PL intensity at time t and 0, respectively. 0 τPL and β are the mean lifetime and stretched parameter, respectively. The higher energy part of PL originates from the emission of smaller QDs. Part of the holes relax to the smaller QDs then recombine with the attracted electrons immediately and give rise to a fast decay time. Part of the holes are thermally excited and transfer to the larger QDs as discussed in the above mentioned temperature dependant FWHM and PL intensity.
The thermal activation and carrier transfer create other decay channels for exciton
sites of larger QDs. The lower parts of PL spectra are emissions due to the recombination from the larger QDs. It is better fitted by the stretched exponential function with a decay time of several hundred ns. The energy dependent decay times obtained by the stretched exponential are also plotted in Fig. 5. The increasing decay time with decreasing energy indicates the exciton thermal activation and transfer between smaller and larger QDs. This phenomenon is similar to the case of exciton transfer between localized states in CdSSe alloys [51]. The excitons in smaller QDs have higher energy and were thermally excited then transferred to larger QDs of lower energy localized states. As a result, the lower energy states have much longer decay time.
Figure 6 presents the PL lifetime and stretching exponent β of different values of coverage. The lifetimes of 2-D layers are longer than 100 ns. On the other hand, as the 0-D QDs are formed, the lifetime decreases to values less than 100 ns. It decreases with the coverage thickness. The stretching exponent β also drops as the coverage above 2.5 ML. The long decay time of 2-D layers could be attributed to the long traveling (transfer) time before recombination. However, for the 0-D case, the formation of QDs results in an increase in numbers of recombination centers and a decrease in the traveling time. Furthermore, the decreasing stretching exponent β reflects the increasing recombination centers and the size of QDs. This result is
closely related to the numerical analysis of hopping-transport models by Sturman et al.
[28].
In conclusion, optical properties of MBE-grown type-II ZnTe QDs in ZnSe matrix were investigated by temperature-dependent and time-resolved PL spectroscopy. The PL results showed that the critical thickness for dot formation of SK ZnTe/ZnSe QDs is around 2.4 MLs. The two activation energies obtained from the temperature-dependent PL are consistent with the phenomenon of initial decrease in FWHM then increase with temperature. The PL decay profiles were reasonably fitted by the Kohlrausch’s stretched exponential which describes the hole activation from smaller QDs then transfer and re-capture to larger QDs. Current work shows that the Kohlrausch’s stretched exponential is very useful to analyze the decay profiles of the semiconductor QDs which involve carrier activation, transfer and then re-capture process.
1.8 2.0 2.2 2.4
Fig. 1: Low temperature PL spectra of ZnTe QDs with different coverages. The inset shows the PL peak energy of QD as a function of ZnTe coverage. The solid line is just a guide for eyes.
0 50 100 150 200 250 300
Fig. 2: (a) Temperature dependent PL spectra of 2.6 ML sample. (b) The FWHM as a function of the sample temperature for 2.6 ML samples.
0.00 0.02 0.04 0.06 0.08 0.10 Ea1 = 4.5+0.4 meV
Ea2 = 60.6+0.6 meV
Integrated PL Intensity (a.u.)
1/T(1/K)
2.6 ML
Fig. 3: PL intensity as a function of the inverse temperature.
0 100 200 300 400 3.0 ML 2.8 ML 2.6 ML 2.5 ML 2.4 ML 2.2 ML 2.0 ML
Normailized PL Intensity (a.u.)
Time (ns)
1.8 ML
Fig. 4: The decay profiles of the PL signal from ZnTe QDs detected at their PL peak energies with different coverage.
1.8 2.0
0 50 100 150 200 250 300 350
Lifetime (ns)
PL intensity (a.u.)
Photon energy (eV)
2.6 ML
Fig. 5: The energy dependent decay times obtained by the stretched exponential.
1.8 2.0 2.2 2.4 2.6 2.8 3.0 50
100 150 200
0.0 0.2 0.4 0.6 0.8 1.0
Decay time (ns)
Coverage (ML)
Beta
Fig. 6: Coverage dependences of decay times and β values.
Chapter 5
ZnMnTe/ZnSe quantum dots
In this chapter, optical properties of type-II diluted magnetic semiconductor (DMS) ZnMnTe quantum dots (QDs) were investigated by temperature-dependent, time-resolved and magneto-photoluminescence (PL) spectroscopy. The time-resolved magneto-PL was analyzed by σ+ and σ– circular polarization to study the electron and hole spin dynamics. Both σ+ and σ– time-decay PL profiles can be well fitted by the Kohlrausch’s stretching exponential function. In addition, the magnetic polarons are detected at temperature up to 100K.
The samples studied in this chapter were grown on GaAs (100) substrates by a Vecco Applied EPI 620 molecular beam epitaxy (MBE) system. After the deposition of ZnSe buffer layer, two kinds of alternating supply MBE methods were performed to grow the ZnMnTe QDs of 2.5 mono-layers (MLs). A 2.5 ML ZnTe QD sample
The samples studied in this chapter were grown on GaAs (100) substrates by a Vecco Applied EPI 620 molecular beam epitaxy (MBE) system. After the deposition of ZnSe buffer layer, two kinds of alternating supply MBE methods were performed to grow the ZnMnTe QDs of 2.5 mono-layers (MLs). A 2.5 ML ZnTe QD sample