Recombination dynamics of a localized exciton bound at basal stacking faults within
the m-plane ZnO film
S. Yang, H. C. Hsu, W.-R. Liu, B. H. Lin, C. C. Kuo, C.-H. Hsu, M. O. Eriksson, P. O. Holtz, and W. F. Hsieh
Citation: Applied Physics Letters 105, 011106 (2014); doi: 10.1063/1.4887280 View online: http://dx.doi.org/10.1063/1.4887280
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/105/1?ver=pdfcov Published by the AIP Publishing
Articles you may be interested in
Surface related and intrinsic exciton recombination dynamics in ZnO nanoparticles synthesized by a sol-gel method
Appl. Phys. Lett. 102, 013109 (2013); 10.1063/1.4774002
Photoluminescence associated with basal stacking faults in c-plane ZnO epitaxial film grown by atomic layer deposition
Appl. Phys. Lett. 100, 101907 (2012); 10.1063/1.3692730
Exciton localization on basal stacking faults in a -plane epitaxial lateral overgrown GaN grown by hydride vapor phase epitaxy
J. Appl. Phys. 105, 043102 (2009); 10.1063/1.3075596
Complex excitonic recombination kinetics in ZnO: Capture, relaxation, and recombination from steady state Appl. Phys. Lett. 90, 041917 (2007); 10.1063/1.2432259
Radiative recombination centers induced by stacking-fault pairs in ZnSe/ZnMgSSe quantum-well structures Appl. Phys. Lett. 75, 3944 (1999); 10.1063/1.125502
P. O. Holtz, and W. F. Hsieh
1
Scientific Research Division, National Synchrotron Radiation Research Center, Hsinchu 30076, Taiwan
2
Department of Photonics and Advanced Optoelectronic Technology Center, National Cheng Kung University, 701 Tainan, Taiwan
3
Department of Photonics and Institute of Electro-Optical Engineering, National Chiao Tung University, Hsinchu 30010, Taiwan
4
Department of Physics, Chemistry and Biology (IFM), Link€oping University, Link€oping, Sweden (Received 12 March 2014; accepted 24 June 2014; published online 7 July 2014)
We investigated the carrier dynamics near basal stacking faults (BSFs) inm-plane ZnO epitaxial film. The behaviors of the type-II quantum wells related to the BSFs are verified through time-resolved and time-integrated photoluminescence. The decay time of the emission of BSFs is observed to have a higher power law value and longer decay time than the emission of the donor-bound excitons. The spectral-dependent decay times reveal a phenomenon of carriers migrating among band tail states, which are related to the spatial distribution of the type-II quantum wells formed by the BSFs. A high density of excited carriers leads to a band bending effect, which in turn causes a blue-shift of the emission peak of BSFs with a broadened distribution of band tail states.VC 2014 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4887280]
ZnO, a wide direct band gap material, has attracted much attention due to its potential for applications in ultra-violet photonic devices.1–4The presence of the built-in elec-tric field in c-plane-grown ZnO reduces the emission efficiency. To avoid this problem, much research has been focused on the growth of non-polar (a- or m-planes) ZnO epitaxial films. Since ZnO substrates are very rare and ex-pensive, other foreign substrates, such as sapphire, SiC, and LaAlO3, are the most commonly used substrates to grow
nonpolar ZnO and related alloys. Therefore, this heteroepi-taxy results in the generation of a high density of structural defects, such as dislocations and stacking faults. The hetero-epitaxial non-polar hetero-epitaxial films suffer from a high density of basal stacking faults (BSFs). From the viewpoint of crys-tal growth, defects such as BSFs should be eliminated or at least significantly reduced. Several growth methods have been proposed to reduce the density of BSFs in ZnO and III-Nitride materials.5 In contrast, some of these defects have exposed unusual properties, which have potential to be employed in particular applications. Interestingly, BSFs exhibiting optical activity strongly influence the lumines-cence properties of the materials with wurtzite structure such as GaN and ZnO.6–10 Through ab initio calculations, the BSF is predicted to have the zinc-blend (ZB) lattice structure embedded in the wurtzite (WZ) ZnO matrix, which forms a type-II quantum well (QW).11Besides possible applications as lighting devices, the type-II QW structures in semiconduc-tor materials are promising for photovoltaic devices as a result of their spatial separation of photoexcited electrons and holes.12,13Consequently, this ZnO-based spontaneously formed type-II QW system has become alternatively impor-tant. Although it is essential to understand the luminescence
mechanism and to investigate the photoexcited carrier dy-namics, published reports on the experimental aspects of the carrier dynamic properties of this material system have been limited up to now. In this article, we present a detailed emis-sion mechanism of the BSFs in them-plane ZnO layers. We report on some peculiar behaviors such as S-shaped peak shifting due to the composition fluctuation and the migration of photoexcited carriers among the band tail states. The high density of electrons accumulated at the interface between the potential well and barrier result in a band bending effect that broadens the distribution of these energy levels and decreases the decay times.
An epitaxial m-plane ZnO layer with a thickness of 130 nm was grown on m-plane sapphire substrates by the atomic layer deposition (ALD) method (for details, see Ref. 14). The TEM results show that BSFs are the predominant defect with a density of106cm1. For the transient
photo-luminescence (PL) measurements, the excitation from the third harmonic (kexc¼ 266 nm) of a Ti:sapphire femtosecond
pulsed laser (pulse width 150 fs) with the repetition rate of 76 MHz was employed, and the laser spot was focused down to about 1 lm. The time-resolved PL (TR-PL) signal was detected by a UV-sensitive Hamamatsu streak camera with a temporal resolution better than 20 ps.
The low-temperature time-integrated PL (TI-PL) spectra of the sample with different excitation powers are shown in Figure1(a). The dominating emission at 3.32 eV results from the BSFs, and the emission centered at 3.37 eV is attributed to the near-band-edge (NBE) emission bands consisting of the donor-bound exciton (D0X) and the free exciton (FX).7The DoX emission has a rather broad width (FWHM 22 meV) as compared with those of the pulsed-laser-deposition-grown ZnO films (FWHM 10 meV and typical 9–15 meV).15,16
Some of the literature17–19 has reported that DoX emissions with a lifetime range from 230 to 370 ps were observed in
a)Authors to whom correspondence should be addressed. Electronic addresses: hsuhc@mail.ncku.edu.tw and wfhsieh@mail.nctu.edu.tw
high-quality ZnO thin film, with a line width of less than 5 meV. Compared to these reported results, the rather large line width and short lifetime of the DoX emission in this paper could be due, we believe, to defects such as Zinc vacancy and/or BSFs, which may have enhanced the non-radiative transition from the DoX. The energy peak of the BSF emission reveals a blueshift from 3.319 to 3.322 eV as the excitation power is increased, and the peak energy as a function of the excitation power is shown in Figure1(b). In type-II hetero-structures of other material systems, a blueshift of the peak energy due to excitation power increment has been reported.20,21In our case, the electrons and holes are spatially separated in different regions; that is, the electrons are con-fined in the ZB structure of the ZnO layer, and the holes are localized in the WZ structure of the ZnO layer. The well width formed by BSFs is 0.8 nm (exactly 1.5 times that of the c-axis lattice constant)9less than the exciton Bohr radius of 2.3 nm, which leads to the zero Stokes-shift that originated from the screening effect.22,23 Therefore, the blueshift for increasing excitation power may not be a result of screening the electric fields. Such spatial separation between electrons and holes leads to an electric field, which gives rise to the bending of the valence and conduction band. The electrons and holes accumulated at the interface result in the formation of the potential well, which can be treated as approximately triangular. Thus, the electron quantization energy is expected to be proportional to the third root of the excitation power.24 As shown in Figure1(b), the PL peak position energy can be fitted well by the third root of the excitation power, which indicates that the blueshift indeed originates from the recom-bination in the type-II QWs.25,26
In Figure 1(c), the energy peak of the BSF emission (marked by solid black circles) has a peculiar shift with rais-ing temperature: the peak energy as a function of the temper-ature is shown in Figure 1(d) and termed as a "S-shape behavior," which is attributed to the localization effect
observed on the excited carriers in the QWs.27The S-shape behavior can be explained by band filling in the tails of the density-of-states (DOS), which implies that band tail states exist in the m-plane ZnO thin film. The model assumes that there are a certain amount of trap states forming the band tails with the DOS of Gaussian distribution, and the temperature-dependent peak energy is given by28
h0¼ E0
r2
kBT
; (1)
whereE0 is the transition energy determined by the centers
of the DOS Gaussian distributions of the electrons in the band tail and the holes in the valance band, and r2=k
BT is
the Stokes-type redshift, and where r is the dispersion of the Gaussian of the band tail, which is correlated to the localiza-tion energy. The quantity E0 relates to the band gap
described by Varshni’s equation; therefore, the temperature-dependent equation of the peak energy can be described as
E0ð Þ ¼ ET 0ð Þ 0 aT2 Tþ b r2 kBT : (2)
From the fitting of the S-shaped curve of the BSF peak energy as a function of the temperature to Eq. (3) (see Fig. 1(d)), the parameter values of E0ð0Þ 3.334 eV,
a 1.02 meV, b 920 K, and r 7.04 meV are obtained. The fitting results reveal that the BSF emission at low temperatures originates from the transition between the electrons trapped in the QWs formed by the BSFs and the holes in the valance band. Then the electrons in the QWs are further trapped into the band tail states. Subsequently, the electrons in the QWs and the holes in the valance band are considered to form BSF-bound excitons, which are further confined into the band tail states. The NBE emission is composed of the DoX (marked by hollow red circles) and the FX (solid red circles) transitions. In FIG. 1. (a) Power dependent TI-PL spectra recorded at 5 K of the ZnO film grown by ALD and annealed at 600C.
(b) The intensity versus the excitation power of the BSF and NBE emissions. The dashed lines represent the outcome of the power-law fitting work. (c) The temperature dependent TI-PL spectra of the ZnO film taken between 10 and 200 K. (d) The energy versus tempera-ture plot of the BSF and NBE emis-sions. The dashed lines depict the fitting results to the Varshni’s law.
011106-2 Yang et al. Appl. Phys. Lett. 105, 011106 (2014)
Figure1(d), the peak energy of the FX transition as a function of temperature is fitted to Varshni’s formula, which is shown as a red dashed line and determines the transition energy at 0 K to be about 3.375 eV. The FX tends to be bound by a donor to form DoX below 100 K, which is consistent with previous observations.29To confirm this assumption, we performed TR-PL measurements. In Figure2(a), the TR-PL spectra at a tem-perature of 5 K are monitored after a pulsed excitation. With increasing time after the pulse, the BSF emission band shows a redshift from 3.322 to 3.315 eV, while the NBE emission band remains at about 3.362 eV. Such an observed behavior is a typ-ical feature for a relaxation process of the BSF-bound electrons in the band tail states.30To determine the average decay times of the BSF and NBE emission bands, the decay curves are plotted in Figure 2(b). The decay curve of the BSF (NBE) emission band is obtained by integrating the intensity in the energy range from 3.280 to 3.345 eV (from 3.345 to 3.397 eV). Each decay curve is fitted by a bi-exponential function: IðtÞ ¼ A1expðt=s1Þ þ A2expðt=s2Þ. For the NBE
emis-sion, the decay times are determined to be s1 52 ps, s2 250
ps, and the ratio ofA1=A2is about 8.3, which means that s1 is
the decay time of the NBE in which the DoX transition domi-nates. The decay time of the NBE band agrees with the reported lifetime of DoX, and s2 is the decay time resulting
from the free exciton transition.31,32For the BSF emission, the decay times are s1 55 ps and s2 450 ps. Corfdir et al.
reported the decay curves of the emissions in BSFs of the GaN a-plane films.33 They found that bi-exponential decay curves only occur in the region of the BSFs bundles. The origin of the shorter decay time s1 is attributed to the complex overlapping
of electron and hole wave function in the different distributions of inter-BSF distances. The longer decay time s2 is due to the
radiative recombination of the BSF-bound electrons with the holes in the W-ZnO.26Using the rate equation analysis,34the rate equations of the transitions of FX-to-D0X and FX-to-BSF could be written as follows:
dnDX=dt¼ nFE=sFE!DX nDX=sDX; (3)
dnBSF=dt¼ nFE=sFE!BSF nBSF=sBSF; (4)
wherenFE is the density of the free excitons, sFE is the
life-time of the free exciton, and the constants sFE!DX and
sFE!BSF are the lifetimes of the capture transitions of
FX-to-D0X and FX-to-BSF, respectively. In the steady state, the dnDX=dt and dnBSF=dt in the rate equations are zero. Besides,
the intensities of FX, D0X, and BSF are presented as IFE
¼ nFE=sFE,IDX¼ nDX=sDX, andIBSF¼ nBSF=sBSF. Therefore,
we could obtain the lifetimes of the capture transitions of FX-to-D0X (sFE!DX) and FX-to-BSF (sFE!BSF) from Eqs. (3)
and(4)
sFE!DX¼ IFEsFE=IDX; (5)
sFE!BSF¼ IFEsFE=IBSF: (6)
Based on the PL results, the intensity ratio of D0X to FE is about 8.3 and the lifetime of the FX is250 ps. Therefore, the value of sFE!DX can be roughly estimated to be about 6
ps. Similarly, the intensity ratio of BSF to FX is about 1.67 104, so the value of s
FE!BSFcan be roughly estimated
to be 15 fs.
To gain more information on the details of the band tails, spectrally dependent decay times are plotted. In Figure 3, the PL decay times at different monitored emission ener-gies are plotted together with the corresponding TI-PL spec-tra (solid curves). The decay times of the BSF emission gradually increase with decreasing emission energy, while the NBE emission appears to have fixed decay times of the DoX transition for an excitation power of 100 lW in Figure 3(a). Such a significant difference implies that carriers migrating among band tail states will occur for the BSFs. In the energy range of the BSF emission band, the density of band tail states is estimated to follow expðE=EoÞ, and the
radiative recombination time sr as a constant does not vary
with altered emission energies. The decay time sðEÞ can be expressed as30,35
s Eð Þ ¼ sr
1þ exp E Eð meÞ=Eo
; (7)
where Eme is the characteristic energy, which plays a role
analogous to that of a mobility edge, and the inverse value of the radiative recombination time corresponds to the transfer rate for electrons from the band tail states to the valance band. The parameter Eo represents the characteristic energy
of the density distribution of the band tail states, which also plays a role analogous to the depth of the band tail state.31,36 Spectrally dependent decay time curves with excitation powers from 10 to 1020 lW at 5 K (not shown here) reveal an ambiguous cross at 3.33 eV. This observation implies that the value of Eme is consistent with 3.33 eV, evaluated
from Eq. (4). For the case of an excitation laser power of 100 lW, the best fits in Figure 3(a) are sr¼ 758 ps,
Eme¼ 3.33 eV, and E0¼ 18 meV. As the temperature is
time interval between each other is 25 ps. (b) The decay curves of the inten-sity integrated from the BSF and NBE emissions in energy ranges of 3.280–3.345 eV and 3.345–3.397 eV, respectively. The decay curve at the bottom is the response of the instrument.
raised to 90 K, the curve of the decay times versus the moni-tored photon energy (in Fig.3(b)) in the BSF emission band reveals an almost constant value of about 450 ps. Such result indicates that the BSF-bound excitons are delocalized from the band tail states by thermal excitation at 90 K. Further, we performed TR-PL with excitation powers of 10 lW and 1020 lW at 5 K, and the decay times versus the photon energy are shown in Figures3(c) and3(d). The best fitting results yield sr¼ 723 ps (535 ps) and Eo¼ 19 meV (30 meV)
for an excitation power of 10 lW (1020 lW). The band pro-file of the QWs structure grown along the polar axis should be tilted by the internal electric field that leads to QWs of tri-angular shape, which is in agreement with the results of power-dependent TI-PL. Therefore, a high density of excited carriers is produced by the high excitation power, and they accumulate at the interface between the potential well and barrier due to the Coulomb attraction, which results in a band bending effect that raises the electron quantization energy as mentioned above and enhances the overlap of the wave functions of the electron and the hole. Such an enhancement of the wave function overlap is in agreement with the decrease of the radiative recombination time sr at
high excitation power in Figure3(d). Besides, the electron quantization energy level increases, which means an increased separation between the energy levels among the band tails states followed by a broadening of the distribution. This fact raises the value ofEo, as observed in Figure3(d).
The separation between the energy levels among the band tail states is also considered to increase with high excitation laser power due to the band bending effect, which results in a high density of photon-excited carriers.
In conclusion, we have studied BSFs-related localized carrier recombination dynamics in the m-plane ZnO epitaxial films. The randomly distributed type-II QWs are formed by the high density of BSFs in the ZnO materials, leading to a wave function coupling between the BSF-bound electrons. By TI-PL and TR-PL measurements, which represent the carrier dynamics in the BSFs, reveal the migration behavior of the BSF-bound excitons in the band tails states. High excitation
laser power leads to a band bending effect, which in turn broadens the distribution of the band tail states, and subse-quently increases the depth of the distribution as well as caus-ing an energy shift of the emission band due to the type-II QWs.
The authors are thankful to P. Bergman in IFM for providing the time-resolved facilities. This work was partly supported by National Science Council of Taiwan under Grant Nos. NSC-99-2112-M-006-017-MY3, NSC-100-2112-M-213-002-MY3, and NSC-102-2112-M-006-012-MY3.
1U. €€ Ozg€ur, Ya. I. Alivov, C. Liu, A. Teke, M. A. Reshchikov, S. Dogan, V. Avrutin, S.-J. Cho, and H. Morkoc¸,J. Appl. Phys.98, 041301 (2005). 2
D. C. Look, D. C. Reynolds, C. W. Litton, R. L. Jones, D. B. Eason, and G. Cantwell,Appl. Phys. Lett.81, 1830 (2002).
3T. Wang, H. Wu, H. Zheng, J. B. Wang, Z. Wang, C. Chen, Y. Xu, and C. Liu,Appl. Phys. Lett.102, 141912 (2013).
4
W.-R. Liu, Y.-H. Li, W. F. Hsieh, W. C. Lee, M. Hong, J. Kwo, and C.-H. Hsu,J. Phys. D: Appl. Phys.41, 65105 (2008).
5R. Ravash, P. Veit, M. M€uller, G. Schmidt, A. Dempewolf, T. Hempel, J. Bl€asing, F. Bertram, A. Dadgar, J. Christen, and A. Krost,Phys. Status Solidi C9, 507 (2012).
6S. Yang, C. C. Kuo, W.-R. Liu, B. H. Lin, H. C. Hsu, C. H. Hsu, and W. F. Hsieh,Appl. Phys. Lett.100, 101907 (2012).
7
W. H. Lin, U. Jahn, H. T. Grahn, L. Chang, M. M. C. Chou, and J. J. Wu, Appl. Phys. Express6, 061101 (2013).
8W. Rieger, R. Dimitrov, D. Brunner, E. Rohrer, O. Ambacher, and M. Stutzmann,Phys. Rev. B.54, 17596 (1996).
9
P. Corfdir, P. Lefebvre, J. Levrat, A. Dussaigne, J.-D. Ganiere, D. Martin, J. Ristic, T. Zhu, N. Grandjean, and B. Deveaud-Pledran,J. Appl. Phys. 105, 043102 (2009).
10Y. J. Sun, O. Brandt, U. Jahn, T. Y. Liu, A. Trampert, S. Cronenberg, S. Dhar, and K. H. Ploog,J. Appl. Phys.92, 5714 (2002).
11
Y. Yan, G. M. Dalpian, M. M. Al-Jassim, and S.-H. Wei,Phys. Rev. B. 70, 193206 (2004).
12Y. Zhang, L.-W. Wang, and A. Mascarenhas,Nano Lett.
7, 1264 (2007). 13
J. Schrier, D. O. Demchenko, L.-W. Wang, and A. P. Alivisatos, Nano Lett.7, 2377 (2007).
14S. Yang, B. H. Lin, C. C. Kuo, H. C. Hsu, W.-R. Liu, M. O. Eriksson, P.-O. Holtz, C.-S. Chang, C.-H. Hsu, and W. F. Hsieh,Cryst. Growth Des. 12, 4745 (2012).
15
W.-R. Liu, W. F. Hsieh, C.-H. Hsu, K. S. Liang, and F. S.-S. Chien, J. Appl. Cryst.40, 924 (2007).
16W.-R. Liu, Y.-H. Li, W. F. Hsieh, C.-H. Hsu, W. C. Lee, Y. J. Lee, M. Hong, and J. Kwo,Cryst. Growth Des.9, 239 (2009).
FIG. 3. The spectrally dependent decay times are plotted with the same excitation power, 100 lW, at different temperatures; (a) 5 and (b) 90 K, and with different excitation power; (c) 10 and (d) 1020 lW at the same tempera-ture 5 K.
011106-4 Yang et al. Appl. Phys. Lett. 105, 011106 (2014)
19
Y. Chen, H. J. Ko, S. K. Hong, and T. Yao,Appl. Phys. Lett76, 559 (2000). 20S.-Y. Lin, C.-C. Tseng, W.-H. Lin, S.-C. Mai, S.-Y. Wu, S.-H. Chen, and
J.-I. Chyi,Appl. Phys. Lett.96, 123503 (2010). 21
T. Kawazu, T. Mano, T. Noda, and H. Sakaki, Appl. Phys. Lett. 94, 081911 (2009).
22T. Kuroda and A. Tackeuchi,J. Appl. Phys.92, 3071 (2002).
23S. F. Chichibu, A. C. Abare, M. P. Mack, M. S. Minsky, T. Deguchi, D. Cohen, P. Kozodoy, S. B. Fleischer, S. Keller, J. S. Speck, J. E. Bowers, E. Hu, U. K. Mishra, L. A. Coldren, S. P. DenBaars, K. Wada, T. Sota, and S. Nakamura,Mater. Sci. Eng. B59, 298 (1999).
24T. T. Chen, C. L. Cheng, Y. F. Chen, F. Y. Chang, H. H. Lin, C.-T. Wu, and C.-H. Chen,Phys. Rev. B.75, 033310 (2007).
25
S. D. Singh, V. K. Dixit, S. Porwal, R. Kumar, A. K. Srivastava, T. Ganguli, T. K. Sharma, and S. M. Oak,Appl. Phys. Lett.97, 111912 (2010). 26D. Alonso- Alvarez, B. Alen, J. M. Gercıa, and J. M. Ripalda,Appl. Phys.
Lett.91, 263103 (2007).
30
C. Gourdon and P. Lavallard,Phys. Status Solidi B153, 641 (1989). 31Y. Narukawa, Y. Kawakami, S. Fujita, and S. Fujita,Phys. Rev. B 55,
R1938 (1997). 32
S. F. Chichibu, T. Onuma, M. Kubota, and A. Uedono,J. Appl. Phys.99, 093505 (2006).
33P. Corfdir, J. Ristic, P. Lefebvre, T. Zhu, D. Martin, A. Dussaigne, J. D. Ganiere, N. Grandjean, and B. Deveaud-Pledran,Appl. Phys. Lett.94, 201115 (2009).
34
T. Schmidt, K. Lischka, and W. Zulehner, Phys. Rev. B 45, 8989 (1992).
35M. Strassburg, M. Dworzak, H. Born, R. Heitz, A. Hoffmann, M. Barteis, K. Lischka, D. Schikora, and J. Christen, Appl. Phys. Lett. 80, 473 (2002).
36I. A. Buyanova, J. P. Bergman, G. Pozina, W. M. Chen, S. Rawal, D. P. Norton, S. J. Pearton, A. Osinsky, and J. W. Dong,Appl. Phys. Lett.90, 261907 (2007).
