ecombination process for this type-II QD s
exp( / ) exp( / )2
I t I t
In order to understand the mechanism of carrier r
ystem, the TRPL was investigated. The experimental PL decay profiles (open circle) measured for the ZnTe QDs sample with 2.4 ML coverage is shown in Fig. 3-7. The decay time is composed of a faster initial component and a slower tail component. In order to resolve the time constants, all of the experimental data were fitted with two-exponential function described by the experimental data is shown by the solid line. The time constants 1
two diffe T fo
τ andτ2are 3.2 ns and 74.7 ns, respectively. The lifetime of carrier recombination is inverse proportion to the wavefunction overlap of electrons and holes [34]. The observed two-exponential decay can be explained by the spatial separation of charges. After the photoexcitation, a dipole layer is formed between the holes in the ZnTe layer and the electrons attracted from the surrounding ZnMnSe regions. This field-induced band-bending effect makes the electron wavefunction closer to the ZnTe region. The initial faster time constant is attributed to the increasing spatial overlap due to the band-bending effect. When the majority of the carriers have recombined, the wavefunction overlap decreases and the band-bending effect is negligible. This will lead to a slow radiative recombination process and results in a longer decay time. Therefore, the slow decay time is attributed to the decay time of the ground state exciton. Because of the nature of type-II QDs structure of ZnTe/ZnMnSe, holes are confined in the ZnTe layer and electrons are localized in ZnMnSe layer around the dots. Due to the less electron-hole overlap in type-II band structure, the measured lifetimes are much longer than the radiative lifetime of type-I QDs [35-37].
In Fig. 3-8 we show coverage thickness dependence of the exciton decay times. The excito
Ds with cover
n decay time increases with increasing coverage and starts to decrease above 2.7 ML.
We explain the result by separating the data into two regions: 2D layer and 0D QD. In the 2D layer case, the lifetime increases with the coverage thickness due to the increasing electron and hole separation. However, in the QD case, the recombination coefficient depends crucially on the portion of hole wavefunction confined in QD structure. Figure 3-9 shows the schematic representations for the wavefuntion distribution for different case. For the 2D layer case, the wavefunction of holes is located at the center of 2D layer. When the coverage thickness increases, the portion of hole wavefunction decreases at the edge and causes a less electron-hole wavefunction overlapping shown in Fig. 3-9 (a) to (b). Thus, the decay time of the ground state transition becomes longer with increasing coverage. On the contrary, the wavefunction of holes starts to move in quantum dot as the dot formed as shown in Fig. 3-9 (c). Increasing dot size makes the ground state energy become lower and leads the wavefunction of holes closer to ZnMnSe capping layer shown in Fig. 3-9 (d). Hence, the overlap of electron-hole wavefunction enhances and results in a shorter decay time.
Figure 3-10 shows the spectral distribution of decay time of ZnTe/ZnMnSe Q
age thickness of 2.7 ML. As shown in the figure, the exciton decay time depends on the monitored emission energy and is shorter for higher energies. Carriers in the small QDs, whose ground state energy is larger, can transfer to the lower energy state in the adjacent larger QDs [37]. The extra transfer channels will result in a faster decay rate for higher-energy transition. The PL in the lower-energy transition has a longer decay time due to the increase in the number of carriers that relaxed from the smaller QDs. We also examined the temperature effect on the indirect type-II transition in ZnTe/ZnMnSe QD structures. Figure 3-11 shows the temperature dependence of lifetime at peak energy with 2.7 ML coverage. The ground state exciton decay time is almost constant with temperature below 50 K. This is not surprising because the thermal distribution of radiative rate was greatly suppressed in QD structures. The
exciton decay time, however, decreases quickly with increasing temperature above 50 K. It is because nonradiative recombination dominates the transition above 50K.
3.3 Polarization of zero magnetic field
In order to study the spin polarization degree without an applied magnetic field (B), the circular polarization PL was investigated. Figure 3-12 shows the PL spectra with σ+(dash line) and σ−(solid line) circular polarization for the 2.7 ML ZnTe/ZnMnSe QDs without external m etic field. We can clearly see that the polarization rate (P) is nonzero, where
P is defined as
calculated value of is about -7 % in this sample. It’s quite a sp ial enomenon because the spin orientation of carriers should be randomized before recombination for B=0 [38]. This result implies that some built-in B exists in the type-II structures. The Mn spin alignment is not essential for the built-in B. The results show that the nonzero polarization is less sensitive to thermal energy at least for temperature below 100 K. However, the spin polarization induced by Mn should be greatly suppressed with increasing temperature [39]. We believe that this phenomenon is similar to those found in type-II semiconductor heterostructures [40], where a nonzero spin splitting is possible with a built-in B induced by the motion of carriers under the presence of nonuniform electric field across the interfaces of ZnTe QDs and ZnMnSe layer. The nonuniform strain induced internal magnetic field might be in the minus z-direction because the negative polarization rate changes to positive by an external magnetic field applied to z-direction [30]. The schematic diagram was shown in Fig. 3-13. Because of the built-in magnetic field, the carriers of higher-energy spin state will easily flip to the lower-energy spin state and result in a negative polarization rate as shown in Fig. 3-13(b).
P
This investigation shows that the carrier recombination process of type-II structures dominated by the interface electric field enormously. It also provides a possibility to manipulate the spin degree by tuning the internal strain for zero applied magnetic field.
QD
ZnSe band edge
3.0 ML H
2.7 ML
2.4 ML
Fig. 3-1. Low temperature PL spectra of ZnTe/ZnMnSe QDs for ZnTe coverages of 3.0, 2.7, 2.4, 2.2, and 1.8 MLs.
1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2
P L In te n s ity (a .u .)
phot gy(eV)
2.2 ML
ZnTe epilayer
1.8 ML
on ener
Type-II emission
1.8 eV
ZnMnSe ZnTe ZnMnSe
Fig. 3-2. Schematic diagram of type II band alignment for ZnTe/ZnMnSe QDs.
1.85 1.90 1.95 2.00 2.05 2.10 2.15 2.20
Fig. 3-3. Peak energy versus ZnTe coverage.
1.8 2.0 2.2 2.4 2.6 2.8 3.0
PL peak energy(eV)
Coverage(ML) 2D
layer
0D QD
rent coverage.
(a) 1.8 ML
(b) 2.4 ML
(c) 2.7 ML
Fig. 3-4. RHEED pattern of ZnTe QDs with diffe
0 2 4 6 8 10 12 0.00
0.01 0.02 0.03 0.04
2.2 ML 2.4 ML 2.7 ML s=0.30 s=0.32 s=0.10
PL peak energy shift(eV)
Power(mW)
Fig. 3-5. Excitation power dependence of PL peak energy shifts for samples with coverage of 2.2 ML ,2.4 ML and 2.7 ML.
1.8 2.1
Fig. 3-7. (a) Schematic diagram of the excitation power dependence of the carrier filling distribution for different QD sizes. (b) The estimated peak energy blue-shifts.
0 50 100 150
PL Intensity(a.u.)
Time(ns)
1 3.2( )ns τ =
fast ( 1
Fig. 3-7. TRPL of ZnTe QDs sample with 2.4 ML monitored at peak energy (1.93 eV).
The solid line is fitted by a two-exponential decay function.
2 74.7( )ns τ =
)
slow ( 2 τ
τ )
40 50 60 70 80 90
Fig. 3-8. ZnTe coverage dependence of the ground state exciton decay time.
1.8 2.0 2.2 2.4 2.6 2.8 3.0
Decay time(ns)
Coverage(M )
ground state exciton decay time
2-D layer 0-D QD
L
Fig. 3-9. Schematic representations for the wavefuntion distribution
shadow) and hole (dark shadow) for (a) thin 2D layer, (b) thicker 2D
of electron (light layer, (c) small 0D QD, and (d) larger 0D QD without considering wetting layer.
(a)
(c) (d)
(b)
ZnTe ZnTe
ZnTe ZnTe
ZnMnSe ZnMnSe ZnMnSe ZnMnSe
ZnMnSe
ZnMnSe ZnMnSe ZnMnSe
Fig. 3-10. Broad band time-resolved PL of ZnTe QDs with 2.7 ML coverage.
1.8 2.0
ground state exciton decay time
40 50 60 70 80
90 PL
Photon energy(eV)
Decay time(ns) PL Intensity(a.u.)
0 20 40 60 80 100 120 140 40
45 50 55 60 65 70 75 80
Exciton decay time(ns)
Temperature(K)
ZnTe/ZnMnSe 2.7 ML ground state decay time
Fig. 3-11. Temperature-dependent lifetime at peak energy for QD of 2.7 ML coverage.
0 10000 20000
Fig. 3-12. PL spectra with σ+(dash line) and σ−(solid line) circular polarization for the 2.7 ML ZnTe QDs at external magnetic field B=0.
1.65 1.70 1.75 1.80 1.85 1.90 1.95 2.00 2.05 2.10
PL Intensity(a.u.)
photon energy(eV)
σ -σ+
polarization~-7%
nMnSe structure (a) and spin relaxation process (b).
-1/2
-Fig. 3-13. The schematic diagram of built-in magnetic field in the ZnTe /Z
3/2 +1/2
+3/2
σ- σ+
ZnMnSe (capping layer) 500 Å
GaAs (substrate)
B
inZnTe QDs layer 1.8~3.0 ML) (
ZnSe (buffer layer) 500Å ZnMnSe (buffer layer) 500 Å
B
exChapter 4：Conclusions
The optical properties of t olecular beam
epitaxy were investigated by conventional PL nd TRPL in this study. The type-II emission peak
ant is attributed to the increasing spatial overl
ype-II DMS ZnTe/ZnMnSe QDs grown by the m a
energies of samples with 1.8, 2.2, 2.4, 2.7, and 3.0 ML coverage are 2.175, 2.005, 1.963, 1.917, and 1.888 eV, respectively. S-K growth mode was identified by the RHEED patterns and different red-shift slopes of the PL peak energy with increasing ZnTe coverages. The excitation power dependence of spatially indirect transitions of ZnTe/ZnMnSe QDs shows a significant blue-shift for PL peak energy. The blue-shift of 0D QDs and 2D layer are caused by two different mechanisms. The QD size distribution and band-bending effect result in the blue-shift for 0D QDs and 2D layer, respectively.
In addition, the radiative decay time is composed of a faster initial component and a slower tail component. The initial faster time const
ap due to the band-bending effect and slower decay time is attributed to the recombination time of ground state transition. The exciton decay time of samples with different coverages is well-explained by wave function overlap. In addition, the temperature dependent time-resolved PL shows that the carrier recombination process dominated by radiative channel in low temperature but by nonradiative channel above 50K. Finally, a non-zero circular polarization of PL at the absence of magnetic field is attributed to the accumulation of interface charges confined in adjacent layers. This type-II QD system could be a potential candidate for the spintronics devices.
Refe nce
[1] L. V. Titova, J. K. Furdyna, M. Dobrowolska, S. Lee, T. Topuria, P. Moeck, and N. D.
[2] L. Landin, M.S. Miller, M.E. Pist uelson, Science 280, 262 (1998).
[4] D. J. Norris, N. Yao, F. T. Charnock, and T. A. Kennedy, Nano. Lett. 1, 3 (2001).
[7] S. Makino, T. Miyamoto, T. Kageyama, N. Nishiyama, F. Koyama, K. Iga, J. Cryst.
[8] D. Ga ook, D. S. Katzer, and D. Park, Science 273, 87
[9] S. Mackowski, T. A. Nguyen, H. E. Jackson, and L. M. Smith, Appl. Phys. Lett. 83, 5524
[10] m, M. Kim, S. Lee, J. Kossut, J. K. Furdyna, and M. Dobrowolska, J. Cryst.
[11] P. R. Kratzert, J. Puls, M. Rabe, and F. Hennenberger, Appl. Phys. Lett. 79, 2814 (2002).
[15] A. V. Chernenko, P. S. Dorozhkin, V. D.
re
Browning, Appl. Phys. Lett. 80, 1237 (2002).
ol, C.E. Pryor, L. Sam
[3] G. Bacher, T. Kummell, D. Eisert, A. Forchel, B. Konig, W. Ossau, C. R.Becker, and G.
Landwehr, Appl. Phys. Lett. 75, 956 (1999)
[5] P. R. Kratzert, J. Puls, M. Rabe, and F. Henneberger, Appl. Phys. Lett. 79, 2814 (2001).
[6] S.H. Xin, P.D. Wang, Aie Yin, C. Kim, M. Dobrowolska, J.L. Merz, J.K. Furdyna, Appl.
Phys. Lett. 69, 3884 (1996).
Growth 221, 561 (2000).
mmon, E. S. Snow, B. V. Shanabr (1996).
(2003) C. S. Ki
Growth 214, 395 (2000).
[12] Heremans, J. and Partin, D. L., Phys. Rev. B 37, 6311 (1988).
[13] H. Ohno, A. Shen, H. Munekata et al., Appl. Phys. Lett., 69, 363 (1996).
[14] S. Mackowski, S. Lee, J. K. Furdyna, M. Dobrowolska, G. Prechtl, W. Heiss, J. Kossut and G. Karczewski, Phys. Stat. Sol. (b) 229, 469 (2002).
Kulakovskii and A. S. Brichkin, Phys. Rev. B
72, 045302 (2005).
T. Gurung, S. Mac
[16] kowski, H. E. Jackson, L. M. Smith, W. Heiss, J. Kossut and G.
[17] Y. Terai, S. Kuroda, and K. Takita, Appl. Phys. Lett. 76, 2400 (2000).
nd W. Heiss, Appl. Karczewski, J. Appl. Phys. 96, 7407 (2004)
[18] S. Mackowski, H. E. Jackson, L. M. Smith, J. Kossut, G. Karczewski a Phys. Lett. 83, 3575 (2003).
G. Prechtl, W. Heiss, S. Mack
[20] J. Seufert, G. Bacher, M. Scheibner, A. Forchel, S. Lee, M. Dobrowolska and K.Furdyna, Phys. Rev. Lett. 88, 027402 (2002).
E. Oh, K. J. Yee, S. M. Soh, J. U. Lee, J. C. W
Furdyna, H. C. Ri, H. S. Chany and S. H. Park, Appl. Phys. Lett. 83, 4604 (2003).
G. Bacher, H. Schomig, M. K. Welsch, S. Zaitsev, V. D. Kulakovskii, A. Forchel, S M. Dobrowolska, J. K. Furdyna, B. Konig and W. Ossau, Appl. Phys. Lett. 79, 524 (2001).
K. Shib
(Amsterdam) 10, 358 (2001).
[25] P. Recher, E. V. Sukhorukov, and D. Loss, Phys. Rev. Lett. 85
[26] K. Chang, K. S. Chan, and F. M. Peeters, Phys. Rev. B 71, 155309 (2005).
[27] A. Murayama, T. Asahina, K. Nishibayashi, I. Souma, and Y. Oka, J. App 084327 (2006).
Chou, M. Yasar, R. Mallory, A. Petrou, and H. Luo, Appl. Phys. Lett. 89, 263111 (2006).
D. J. Jang, and Y. C. Cheng, J. Appl. Phys. 97, 033514 (2005).
[30] . Uen, W. H. Lan, and S.
[31] 稀磁性半導體量子點」，私立中原大學，博士
[32] C. M. Lin, Y. F. Chen, Appl. Phys. Lett. 85, 2544 (2004).
. Grundmann, D. Bimberg, S.
[34]
[33] N. N. Ledentsov, J. Bohrer, M. Beer, F. Heinrichsdorff, M
V. Ivanov, B. Ya Meltser, S. V. Shaposhnikov, I. N. Yassievich, N. N. Falev, P. S.
Kopev, and Zh. I. Alferov, Phys. Rev. B 52, 14058 (1995).
H. Born, L. Muller-Kirsch, R. Heitz, phys. stat. sol. (b) 228, No.3, R4-R5(2001) [35] T. Y. Lin, D. Y. Lyu, J. Chang, J. L. Shen, and W. C. Chou, Appl. Phys. Lett. 88, 1
(2006).
Appl. Phys. Lett. 68, 1543 (1996).
Physica E 26, 422 (2005).
[39] S. Mackowski, T. Gurung, T. A. Nguyen, H. E. Jackson, L. M. Smith, G and J. Kossut, Appl. Phys. Lett. 84, 3337 (2004).
G. Lommer, F. Malcher, and U. Rössler, Phys. Rev