III. RESULTS AND DISCUSSION

Figures 1(a) and 1(b) show AFM images of the uncapped surfaces for the CdTe/ZnTe QDs with thicknesses of 3.0 and 4.5 ML, respectively. When the thickness of the CdTe layer is 3.0 ML, uniformly sized CdTe QDs are formed on the ZnTe buffer layer. The average heights of the formed CdTe QDs are approximately between 8 and 10 nm, and their diameters are approximately between 45 and 55 nm. The density of the CdTe QDs is approximately 3 × 10¹⁰/cm². However, when the thickness of the CdTe layers is 4.5 ML, large- and small-sized QDs are formed on the ZnTe buffer layer owing to the coalescence of the QDs. The average heights of the formed CdTe QDs are approximately between 9 and 13 nm, and their diameters are approximately between 60 and 85 nm. The density of the CdTe QDs is approximately 1 × 10¹⁰/cm².

FIG. 1.

AFM images for CdTe/ZnTe QDs with thicknesses of (a) 3.0 and (b) 4.5 ML grown on Si substrates. (c) PL spectra at 20 K for CdTe/ZnTe QDs with thicknesses of 3.0, 3.5, 4.0, and 4.5 ML grown on Si substrates. (d) Peak position and FWHM of the PL spectra for CdTe/ZnTe QDs grown on Si substrates as a function of CdTe thickness.

To clarify the observed AFM behavior, we investigated the optical properties of CdTe/ZnTe QDs with various CdTe thicknesses. Figure 1(c) shows the PL spectra at 20 K for CdTe/ZnTe QDs with thicknesses of 3.0, 3.5, 4.0, and 4.5 ML grown on Si substrates. The dominant peaks shown in Fig. 1(c) correspond to the exciton transition from the ground-state electronic subband to the ground-state heavy-hole band (E₁–HH₁) in the CdTe/ZnTe QDs. The dominant peak was assigned to the E₁–HH₁ transition of the CdTe/ZnTe QDs using the theoretical E₁–HH₁ transition value, which was calculated using the finite differential method.^{Reference Woo, Song, Lee, Kim, Yoo, Lee and Park17} The peak position corresponding to the E₁–HH₁ transition in the CdTe/ZnTe QDs shifts to a lower energy level with increasing CdTe thickness. Figure 1(d) shows the peak position and the FWHM of the E₁–HH₁ transition for the CdTe/ZnTe QDs grown on Si substrates as a function of CdTe thickness. A shift in the E₁–HH₁ transition peak for the CdTe/ZnTe QDs as the CdTe thickness increases is attributed to the size variation of the CdTe QDs. The redshift in the PL peak position for the CdTe/ZnTe QDs with increasing CdTe thickness is attributed to a decrease in the energy gap of the CdTe QDs resulting from an increase in the size of the QDs.^{Reference Karczewski, Maćkowski, Kutrowski, Wojtowicz and Kossut18} The FWHM of the E₁–HH₁ transition for the CdTe/ZnTe QDs increases with increasing CdTe thickness, owing to the inhomogeneous broadening resulting from the size distribution of the QDs.^{Reference Leon, Kim, Jagadish, Gai, Zou and Cockayne19} The PL results obtained above clearly confirm the observed AFM behavior.

For understanding the carrier dynamics of CdTe/ZnTe QDs with various CdTe thicknesses grown on Si substrates, we performed time-resolved PL measurements. The time-resolved PL spectra at 20 K for the CdTe/ZnTe QDs with thicknesses of 3.0, 3.5, 4.0, and 4.5 ML grown on Si substrates are shown in Fig. 2(a). PL decay curves exhibit different decay behaviors depending on the CdTe thickness. For determining the PL decay time, we performed reconvolution fitting with the IRF for the time-resolved PL results, which fit well with a biexponential function. The fitting results revealed that the decay times of the fast components are 173, 232, 262, and 308 ps, whereas those of the slow components are 490, 646, 879, and 1203 ps for the CdTe/ZnTe QDs with thicknesses of 3.0, 3.5, 4.0, and 4.5 ML, respectively. A biexponential decay is present in all recorded spectra, in agreement with previous observations on CdTe/ZnTe QDs.^{Reference Couteau, Moehl, Tinjod, Gérard, Kheng and Mariette20} As reported in the literature, the fast component can be attributed to “bright” exciton decay, and the slow component, to “dark” exciton decay^{Reference Labeau, Tamarat and Lounis21}; hereafter, we focus only on the fast component. The decay times of the CdTe/ZnTe QDs grown on Si substrates as a function of CdTe thickness are shown in Fig. 2(b). As mentioned above, the size of the CdTe/ZnTe QDs increases with increasing CdTe thickness. Consequently, the decay time of the CdTe/ZnTe QDs increases with increasing CdTe thickness. This increase is caused by a reduction of the exciton oscillator strength owing to a strong built-in electric field in the larger QDs.^{Reference Kako, Miyamura, Tachibana, Hoshino and Arakawa22} These results of carrier dynamics for the CdTe/ZnTe QDs with various CdTe thicknesses, which are grown on Si substrates, are similar to those for CdTe/ZnTe QDs grown on GaAs substrates.^{Reference Lee, Yim, Lee and Kim16}

FIG. 2.

(a) Time-resolved PL spectra at 20 K for CdTe/ZnTe QDs with thicknesses of 3.0, 3.5, 4.0, and 4.5 ML grown on Si substrates. (b) Decay time of CdTe/ZnTe QDs grown on Si substrates as a function of CdTe thickness.

For determining the activation energy of the electrons confined in the CdTe/ZnTe QDs grown on Si substrates as a function of CdTe thicknesses, we performed temperature-dependent PL measurements. Even though these measurements were performed for the CdTe/ZnTe QDs with thicknesses of 3.0, 3.5, 4.0, and 4.5 ML, only the results of the QDs with thicknesses of 3.0 and 4.5 ML are reported in this article, because those QDs had the minimum and maximum activation energies. The PL spectra measured at several temperatures for the CdTe/ZnTe QDs with thicknesses of 3.0 and 4.5 ML are shown in Figs. 3(a) and 3(b), respectively. Given that the energy gaps of the QDs decrease with increasing temperature, the PL peaks corresponding to the E₁–HH₁ transitions shift to a lower energy level with increasing temperature.^{Reference Benyoucef, Rastelli, Schmidt, Ulrich and Michler23} The occurrence of PL quenching at higher temperatures is attributed to the escape of carriers via a nonradiative recombination path. As the temperature is increased, the integrated PL intensity corresponding to the bound electron peak is given by^{Reference Williams, Bebb, Willardson and Beer24}

(1)$$I = I_0 /[1 + C\exp ( - \Delta E_{\rm{A}} /k_{\rm{B}} T)]\quad,$$

where I ₀ is the integrated PL intensity at 0 K, C is the ratio of the thermal escape rate to the radiation recombination rate, ΔE _A is the activation energy that corresponds to the barrier height for the nonradiative recombination path, and k _B is the Boltzmann constant. From Eq. (1), the activation energies ΔE _A of the electrons confined in the CdTe/ZnTe QDs with thicknesses of 3.0 and 4.5 ML as determined from the solid lines in Figs. 4(a) and 4(b) are 28 and 46 meV, respectively. The activation energies of the electrons confined in the CdTe/ZnTe QDs grown on Si substrates as a function of CdTe thickness are shown in Fig. 4(c). The behavior of the activation energy for the CdTe/ZnTe QDs with various CdTe thicknesses is similar to their decay times. The activation energy of the electrons confined in the CdTe/ZnTe QDs increases with increasing CdTe thickness. This increase is related to an enhancement of the quantum confinement effect resulting from an increase in the energy difference between the electronic state and the conduction band edge.^{Reference Lee, Park and Kim25}

FIG. 3.

PL spectra at several temperatures for CdTe/ZnTe QDs with thicknesses of (a) 3.0 and (b) 4.5 ML grown on Si substrates.

FIG. 4.

Integrated PL intensities as a function of reciprocal temperature for CdTe/ZnTe QDs with thicknesses of (a) 3.0 and (b) 4.5 ML grown on Si substrates. (c) Activation energies of CdTe/ZnTe QDs grown on Si substrates as a function of CdTe thickness.