Physica B: Condensed Matter 558 (2019) 127–130

Contents lists available at ScienceDirect

Physica B: Condensed Matter

journal homepage: www.elsevier.com/locate/physb

Temperature dependent localization dynamics of excitons in Mg0.14Zn0.86O T

alloyed semiconductor
Md. Sherajul Islama,∗, Md. Soyaeb Hasana, E. Kobayashib, Takayuki Makinob
a
Department of Electrical and Electronic Engineering, Khulna University of Engineering & Technology, Khulna, 9203, Bangladesh
b
Department of Electrical and Electronic Engineering, University of Fukui, 3-9-1 Bunkyo, 910-8507, Fukui, Japan

A R T I C LE I N FO A B S T R A C T

Keywords: This paper explicates the temperature induced localization dynamics of exciton in Mg0.14Zn0.86O alloyed
Photoluminescence semiconductor. The photoluminescence (PL) spectroscopy measurement has been performed at the temperatures
Exciton localization ranging from 16 K to 180 K. The linewidths of the PL bands reveal the W-shaped behavior with temperature. We
Monte Carlo simulation also observe the temperature induced S-shaped characteristic of Stokes shift. These experimental findings are
MgZnO alloy
after that modeled using the Monte Carlo simulation of exciton hopping and relaxation. The simulation results
Stokes shift
quantitatively agree with both the Stokes shift and the PL linewidth for the whole temperature regime while
accounting for the band potential fluctuation, σ = 17 meV and the additional inhomogeneous broadening,
Γ = 8 meV. The band potential fluctuation is attributed here to the inhomogeneous distribution of Mg in
Mg0.14Zn0.86O alloyed semiconductor. These results may provide a meaningful reference to the development and
design of ZnO based optoelectronic devices.

1. Introduction knowledge of the electronic structures of the disorder-induced localized

ZnO is a direct, wide band gap semiconductor material with nu-
merous potential applications in spintronics, sensors, transparent elec-
tronics, and especially in optoelectronic devices which operate in the
violet and ultraviolet (UV) regime [1–6]. However, for competent op-
toelectronics, the realization of suitable hetero-junction is the key re-
quirement in order to expedite the confinement of charge carriers and
photons. Ternary MgxZn1-xO alloys have been shown to be an appro-
priate material as the band gap energy of this alloy can be made ad-
justable [5,7] by substituting Zn ions with Mg ions in the host lattice of
ZnO. Thus, for the fabrication of hetero-structures with optimized op-
tical features, the detailed knowledge of the fundamental optical
properties of MgxZn1-xO alloyed semiconductor is essential. Although,
the thin-film ZnO [3,4] along with its Quantum Wells (QWs) [8–11]
have been subjected of many fundamental physics investigations,
however, it is only recently that MgZnO alloy commenced to be actively
investigated both experimentally [5,7,12–15] and theoretically [16].
The above mentioned advantages renders urgent to increase our
knowledge of the optical properties of these alloy materials.
In general, semiconductor alloys suffer from the microscopic con-
centration fluctuation which causes the generation of the localized
states, prevailing from the band edge. The full understanding of the
optical properties of the alloyed semiconductors thus requires the
states. The theoretical efforts have been mainly based on the effective
Hamiltonian approach, where the disorder potential is replaced with
the complex self-energy (the coherent-potential approximation). The
density of states (DOS) of the localized state was obtained through its
numerical solution [6]. Quite recently, the first-principles-based ap-
proach is also adopted to take the electronic localization effect into
consideration [17].
It is well known that the photocurrent spectroscopy enables us the
quantitative evaluation of the line-shape of the localized states ex-
perimentally using the Laplace transformation technique [18]. On the
other hand, Kazlauskas and coworkers deduced the line-shape of the
localized states by analyzing the temperature-dependent photo-
luminescence (PL) data with the technique of Monte Carlo simulation
[19]. This process has successfully clarified the experimentally ob-
served anomalous ‘S-shaped’ and ‘W-shaped’ temperature dependence
of the PL peak energy and line-width, respectively in InGaN/GaN QWs.
Using the similar technique, one of the authors of the present article
evaluated the characteristics of the line-shape of the localized states in
ZnO/MgZnO QWs [20]. Moreover, in case of Zn1-xCdxSe quantum is-
lands [21], the Monte Carlo technique has been employed to simulate
the phonon-assisted redistribution processes of excitons which revealed
the anomalous ‘S-shaped’ temperature behavior of PL peak energies.
Anomalous ‘Λ-shaped’ PL line-width is observed in the temperature

∗
Corresponding author.
E-mail address: sheraj_ruet@yahoo.com (Md. S. Islam).

https://doi.org/10.1016/j.physb.2019.01.049
Received 20 June 2018; Received in revised form 20 December 2018; Accepted 30 January 2019
Available online 30 January 2019
0921-4526/ © 2019 Elsevier B.V. All rights reserved.
Md. S. Islam et al.
range of 0–200 K. Likewise, the evidence of localized exciton has also
been demonstrated in the ternary MgZnO alloyed semiconductor, re-
vealing the remarkable enhancement in emission spectra at certain
temperature and unusual behavior of PL line-width (increase-decrease-
increase) with increasing temperature [13]. Moreover, Neumann et al.
[22] reported the red-shifting behavior of the PL spectra with tem-
perature in this alloy. The signature of exciton motion and their con-
tribution to the inhomogeneous Stokes shift and the full width at half
maximum (FWHM) of PL bands have also been analyzed for different
Mg contents [22]. The unusual behavior of PL bands in MgZnO alloy
can be elucidated adopting the above mentioned technique, which si-
mulates the exciton hopping and relaxation. However, there is no prior
report on the quantitative interpretation of the exciton hopping to ex-
plain the anomalous PL properties in MgZnO alloyed semiconductor.
Thus, the theoretical authentication of the experimental results of the
MgxZn1-xO alloy is more desirable.
It should be noted that the above mentioned technique based on the
combination of the simulation and the PL spectroscopic measurement
allows us the acquisition of the information on the dynamical aspects,
in addition to the ‘static’ information. Accordingly, here in this work we
have investigated the PL spectroscopy of epitaxially grown
Mg0.14Zn0.86O alloyed semiconductor crystals in details. The dynamics
of localized excitons in this alloy have been studied using the technique
of Monte Carlo simulation at the temperature range of 16–180 K.
Considering an additional inhomogeneous broadening with the band
potential fluctuation, the phonon-assisted exciton hopping and relaxa-
tion has been quantified in the Mg0.14Zn0.86O alloy for the first time.
2. Experimental details
The sample measured in this study is 180-nm-thick Mg0.14Zn0.86O
epilayer grown on ScAlMgO4 substrate fabricated by pulsed-laser de-
position. The PL were excited by a 325 nm continuous-wave He-Cd
laser, dispersed by a monochromator, and detected by a liquid-N2
cooled charge coupled device (CCD). The PL measurements were per-
formed at temperatures ranging from 16 K to 180 K.
3. Results and discussion
Fig. 1 depicts the temperature dependent PL spectra of
Mg0.14Zn0.86O alloyed semiconductor. The dots indicate the PL peak
energy positions for each temperature. It is found that the PL peak
Mg0.14Zn0.86O
PL Intensity (a.u)
T=16K
T=30K
T=50K
T=60K
T=90K
T=120K
T=150K
T=180K
3.55 3.6 3.65
E (eV)
Fig. 1. The PL spectra from Mg0.14Zn0.86O alloyed semiconductor measured at
different temperatures. The PL intensities are normalized with respect to the
highest intensity and shift vertically. Dots are the peak energies of each tem-
perature.
128
Physica B: Condensed Matter 558 (2019) 127–130
3.62 50
PL Peak Energy (eV)
FWHM (meV)
3.61 40
PL Peak Energy
PL FWHM
3.60 30
0 50 100 150 200
Temperature (K)
Fig. 2. Temperature dependence of the experimental PL maximum and the
FWHM.
positions show the shifting behavior with temperature. Up to the tem-
perature of 60 K, a small shift in the PL peak energy is observed.
However, a large shift is found till the temperature of 120 K even
though the peak energy shifts again to a smaller value with further
increase of the temperature. In addition, an anomalous characteristic of
PL linewidth is visible with temperature. The PL linewidth remains
unchanged while T < 50 K, however, a sharp increase is observed up
to the temperature of 90 K. After remaining unchanged till T = 140 K, it
increases even more with the further increase of the temperature. The
anomalous behaviors of both the PL maximum and the linewidth have
been illustrated clearly in Fig. 2. It is found that the PL band peak
position exhibits a well-known S-shaped variation [23–25] with tem-
perature. Moreover, the W-shaped behavior of PL line-width is seen
with the change of temperature. These anomalous behaviors of the PL
spectra reveal the signature of exciton hopping over randomly dis-
persed localized states with a crossover from a non-thermalized to a
thermalized distribution function of the excitons [26] in Mg0.14Zn0.86O
ternary alloyed semiconductor.
The exciton hopping and relaxation over the localized states in the
Mg0.14Zn0.86O ternary alloy have been modeled here using the two
dimensional (2D) Monte Carlo simulation. The algorithm was given
elsewhere in details [18], and we will only explain it briefly at this time.
The hopping transition rate of excitons from an occupied site i to an
unoccupied site j is described by the Miller-Abrahams expression [27]:
2rij εj − εi + |εj − εi |
υij = υ0 exp(− − )
α 2kT (1)
where υ0 is the attempt-to-escape frequency, rij is the distance between
site i and j, α is the decay length of the exciton center-of-mass wave
function in the localized states, εi and εj are the energies of state i and j.
We considered the energy of the band tails, which follows a Gaussian
DOS as follows:
N 1 ε2
g (ε ) = exp(− 2 )
σ 2π 2σ (2)
with σ as the scale of energy distribution. Here, N is the concentration
of the localized states. The reference energy is ε = 0 , which is the
middle of the distribution. All the energies in our simulation are below
this value which indicates that we have a “half” distribution of the DOS
[19]. For each generated exciton, the hopping process terminates by
recombination with the probability, τ0−1, here, τ0 is the radiative lifetime
of exciton. The energy of the localized state where the recombination
has taken place is scored to the emission spectrum. Combining the
spatial fitting parameter Nα 2 and temporal fitting parameter τ0 υ0 with
the scale of band potential fluctuation σ , the specific process is de-
termined.
The calculated Stokes shift and the PL FWHM for Mg0.14Zn0.86O
Md. S. Islam et al.
-20
2 3
(a) N =1, 0 0=10
Stokes Shift (meV)
-30 =16meV
=17meV
-40
=18meV
2 3
(b) N =1, 0 0=10
50 =18meV, =7meV
FWHM (meV)
40
=17meV, =8meV
30
=16meV, =8.5meV
0 50 100 150 200
Temperature (K)
Fig. 3. Temperature dependence of the Stokes shift and PL linewidth. Circles
indicate the experimental results. Solid line represents the best fitting with the
fitting parameters σ = 17 meV , Nα 2 = 1, υ0 τ0 = 1 × 103 , Γ = 8 meV, respec-
tively. Dashed line and dash dot line are the calculated results with different
sets of σ and Γ .
ternary alloy have been compared with the experimental results in
Fig. 3(a) and Fig. 3(b), respectively. The solid line in Fig. 3(a) illustrates
the simulated Stokes shift with most relevant set of fitting parameters
(σ = 17 meV , Nα 2 = 1 and τ0 υ0 = 1 × 103 ). It shows a good quantitative
agreement with the experimental Stokes shift except in the temperature
range of 30 K–50 K. In this regime, the calculated Stokes shift is slightly
smaller than the experimental observation. The smaller Stokes shift
may be attributed to the reduced value of PL peak position. The si-
mulated PL peak position shifts to the smaller value owing to adopting
the constant lifetime of exciton in the whole temperature regime.
However, the temperature dependence of the lifetime of excitons is not
monotonous, it first increases at low temperatures and decreases con-
tinuously with increasing temperatures. Moreover, the increase of
lifetime of excitons at low temperature in Mg0.14Zn0.86O epilayer is
higher than that of other materials.
The temperature behavior of the Stokes shift is deduced from the
peak position of the simulated PL spectra. The PL peak position shows
the redshift behavior at low temperature due to the increased mobility
of the excitons. With the increase of temperatures, the excitons become
competent to move long distances to reach deeper localized states.
During this process, the excitons can be trapped by lower energy states
up to T > 60 K. Thus, the lower energy states become populated gra-
dually and the PL peak energy shows the blue shift behavior. After
120 K, the redshifts of the PL peak energy is mainly caused by the band
gap shrinkage.
Fig. 3(b) depicts the W-shaped temperature behavior of the PL
linewidth. The anomalous broadening of the PL linewidth is also at-
tributed to the increased mobility of the localized excitons. At low
temperature regime, only the weakly localized states are occupied and
the PL linewidth is narrow. By raising the temperature, the weakly
localized excitons become more mobile and recombine in a broader
energy distribution of the localized states. In the range of 90 K–120 K,
the excitons achieve thermal equilibrium and the PL linewidth remains
129
Physica B: Condensed Matter 558 (2019) 127–130
Table 1
Comparison of the potential profile fluctuations and additional inhomogeneous
broadening for different structures extracted by double-scaled potential profile
model.
Materials σ (meV) Γ (meV) Ref.
In0.22Ga0.78N/GaN MQWs 31 29 [19]
In0.27Ga0.73N/GaN MQWs 38 47 [19]
Mg0.27Zn0.73O/ZnO QWs 20 12 [20]
Mg0.12Zn0.88O/Cd0.04Zn0.96O QWs 20 29 [20]
Al0.35Ga0.65N/Al0.49Ga0.51N QWsa 46 29 [28]
Al0.26Ga0.74N epilayersa 19 20 [29]
Mg0.14Zn0.86O epilayers 17 8 Our sample
a
3D Monte-Carlo simulation results.
unchanged. The FWHM increases again after T > 120 K, which is re-
lated to the influence of phonon coupling. To get a quantitative
agreement with the experimental linewidth, an additional in-
homogeneous broadening, Γ is incorporated based on the double-scaled
potential profile model proposed by K. Kazluskas et al. [26]. We have
presented the data regarding the relevant values of potential profile
fluctuation scales σ and Γ in Table 1. These parameters were used to
perform the double-scaled potential profile model in several samples,
including epitaxial layer and QW structure. The additional in-
homogeneous broadening, Γ in the ternary alloy is significantly smaller
than that in the QWs. It is suggested that the excitonic localization ef-
fect can be more simply interpreted in the alloy system because the Γ is
related to the effect that cannot be taken into account with the present
model. This is somehow unexpected if we consider about the well-width
dependence of the quantized energy. Thus, the comparative study
suggested that the Γ can be a descriptive parameter for universally
understanding the contributions from the well-width and microscopic
concentration fluctuations against the (localization) mechanism of the
material-informative perspectives.
It is evident that each sample possesses a certain degree of in-
homogeneity due to their constituent elements and structural differ-
ences. The QWs structures where both wells and barrier consist of al-
loyed semiconductor materials possess more roughness than the usual
QWs. The inhomogeneity in epilayer turned out to be much smaller
than multiple QWs and QWs structures. Moreover, to increase the
content of the dopant material leads to the more potential fluctuation.
For Mg0.14Zn0.86O epilayer, the repeated simulation shows that the
realistic fitting of the PL linewidth with the experimental results is
obtained while considering the additional inhomogeneity parameter, Γ
in the calculation. Thus, the actual density of localized states with en-
ergy incorporating the additional inhomogeneous broadening becomes:
(E − E0)2
D (E ) ∝ exp[− ]
2(σ 2 + Γ 2) (3)
We have successfully reproduced the PL spectra and estimated the
optimized linewidth by varying the additional inhomogeneous broad-
ening parameter, Γ although this result shows some discrepancies with
the experiment. The calculated FWHM is narrower than the experi-
mental results in the low temperature regime which is mainly due to the
assumption that all the excitons have a same lifetime.
The calculated results show that for any temperature the estimated
band potential fluctuation, σ of the Mg0.14Zn0.86O epilayer is about
17 meV. The energy of the potential fluctuation, σ is estimated from the
kink of the temperature dependence of experimental linewidth using
the expression, σ = 2KBTkink. The characteristic kink of the PL linewidth
of this alloy is observed at T = 90 K, which indicates a crossover from
non-thermalized to thermalized energy distribution of excitons. The
additional inhomogeneous broadening, Γ is found to be 8 meV.
However, at high temperature, the simulated PL FWHM is narrower
than that of the experimental FWHM. The discrepancy of the calculated
and experimental linewidth at higher temperature may be originated by
Md. S. Islam et al.
the following reasons: (i) the longitudinal optical phonons participate
in the radiative transition, (ii) excitons are thermally activated from
localized to delocalized states and (iii) with the increase of temperature
the process of nonradiative recombination is initiated. Moreover, the
decrease of the lifetime of exciton at high temperature contributes
partially to the discrepancy.
4. Conclusion
In summary, we studied here the temperature induced anomalous
behavior of the PL spectra in Mg0.14Zn0.86O alloyed semiconductor
crystal using the Monte Carlo simulation technique where the exciton
hopping was modeled with Gaussian DOS. We successfully deduced the
Stokes shift and the FWHM of the PL bands in the temperature range of
16 K–180 K. The calculated PL maximum and linewidth depicted the S-
shaped and W-shaped behavior, respectively with temperature. These
data were shown to be in a fair agreement with the experiment for the
whole temperature regime while considering the characteristic energy
scale of the distribution of the localized states, σ = 17 meV and the
additional inhomogeneous broadening, Γ = 8 meV. Through this ana-
lysis, it was conspicuous that the inhomogeneity in epilayer is lower
compared to QW structures. Finally, the simulation of exciton hopping
in Mg0.14Zn0.86O epilayer is expected to be a useful tool for the char-
acterization of ZnO based intricate QW structure.
Acknowledgement
The authors are pleased to acknowledge that the sample used in this
study was provided by Prof. M. Kawasaki of the University of Tokyo.
References
[1] C. Jagadish, S.J. Pearton, Zinc Oxide Bulk, Thin Films and Nanostructures:
Processing, Properties, and Applications, Elsevier, 2011.
[2] P. Capper, S. Kasap, A. Willoughby, Zinc Oxide Materials for Electronic and
Optoelectronic Device Applications, John Wiley & Sons, 2011.
[3] D.C. Reynolds, D.C. Look, B. Jogai, C.W. Litton, G. Cantwell, W.C. Harsch, Phys.
Rev. B 60 (1999) 2340.
130
Physica B: Condensed Matter 558 (2019) 127–130
[4] C. Klingshirn, Phys. Status Solidi (b) 244 (2007) 3027.
[5] A. Ohtomo, M. Kawasaki, T. Koida, K. Masubuchi, H. Koinuma, Y. Sakurai,
Y. Yoshida, T. Yasuda, Y. Segawa, Appl. Phys. Lett. 72 (1998) 2466.
[6] M. Ueta, H.B. Kanzaki, K. Kobayashi, Y. Toyozawa, E. Hanamura, Excitonic
Processes in Solids vol. 60, Springer Series in Solid State Sciences, 1986.
[7] J. Zhu, A.Y. Kuznetsov, M.S. Han, Y.S. Park, H.K. Ahn, J.W. Ju, I.H. Lee, Appl. Phys.
Lett. 90 (2007) 211909.
[8] T. Makino, C.H. Chia, Nguen T. Tuan, Y. Segawa, M. Kawasaki, A. Ohtomo,
K. Tamura, H. Koinuma, Appl. Phys. Lett. 77 (2000) 1632.
[9] T. Makino, K. Tamura, C.H. Chia, Y. Segawa, M. Kawasaki, A. Ohtomo, H. Koinuma,
J. Appl. Phys. 93 (2003) 5929.
[10] T. Makino, C.H. Chia, Nguen T. Tuan, H.D. Sun, Y. Segawa, M. Kawasaki,
A. Ohtomo, K. Tamura, H. Koinuma, Appl. Phys. Lett. 77 (2000) 975.
[11] G. Coli, K.K. Bajaj, Appl. Phys. Lett. 78 (2001) 2861.
[12] M. Wang, E.J. Kim, S. Kim, J.S. Chung, I.K. Yoo, E.W. Shin, S.H. Hahn, C. Park, Thin
Solid Films 516 (2008) 1124–1129.
[13] H. Zhu, C.X. Shan, B.H. Li, Z.Z. Zhang, J.Y. Zhang, B. Yao, D.Z. Shen, X.W. Fan, J.
Appl. Phys. 105 (2009) 103508.
[14] S.S. Chen, X.H. Pan, Y.G. Li, W. Chen, H.H. Zhang, W. Dai, P. Ding, J.Y. Huang,
B. Lu, Z.Z. Ye, Phys. Lett. A 379 (2015) 912–915.
[15] I. Gorczyca, H. Teisseyre, T. Suski, N.E. Christensen, A. Svane, J. Appl. Phys. 120
(2016) 215704.
[16] Y. Hu, B. Cai, Z. Hu, Y. Liu, S. Zhang, H. Zeng, Curr. Appl. Phys. 15 (2015)
15423–15428.
[17] C.E. Ekuma, V. Dobrosavljević, D. Gunlycke, Phys. Rev. Lett. 118 (2017) 106404.
[18] S.A. Studenikin, N. Golego, M. Cocivera, J. Appl. Phys. 84 (1998) 5001–5004.
[19] K. Kazlauskas, G. Tamulaitis, P. Pobedinskas, A. Žukauskas, M. Springis,
C.F. Huang, Y.C. Cheng, C.C. Yang, Phys. Rev. B 71 (2005) 085306.
[20] T. Makino, K. Saito, A. Ohtomo, M. Kawasaki, R.T. Senger, K.K. Bajaj, J. Appl. Phys.
99 (2006) 066108.
[21] B.D. Don, K. Kohary, E. Tsitsishvili, H. Kalt, S.D. Baranovskii, P. Thomas, Phys. Rev.
B 69 (2004) 045318.
[22] M.D. Neumann, C. Cobet, N. Esser, B. Laumer, T.A. Wassner, M. Eickhoff,
M. Feneberg, R. Goldhahn, J. Appl. Phys. 110 (2011) 013520.
[23] D. Monroe, Phys. Rev. Lett. 54 (1985) 146–149.
[24] S.A. Tarasenko, A.A. Kiselev, E.L. Ivchenko, A. Dinger, M. Baldauf, C. Klingshirn,
H. Kalt, S.D. Baranovskii, R. Eichmann, P. Thomas, Semicond. Sci. Technol. 16
(2001) 486–492.
[25] Y.-H. Cho, G.H. Gainer, A.J. Fischer, J.J. Song, S. Keller, U.K. Mishra, S.P. DenBaars,
Appl. Phys. Lett. 73 (1998) 1370.
[26] K. Kazlauskas, G. Tamulaitis, A. Žukauskas, M.A. Khan, J.W. Yang, J. Zhang,
G. Simin, M.S. Shur, R. Gaska, Appl. Phys. Lett. 83 (2003) 3722.
[27] A. Miller, E. Abrahams, Phys. Rev. 120 (1960) 745.
[28] J. Mickevičius, J. Jurkevičius, K. Kazlauskas, A. Žukauskas, G. Tamulaitis,
M.S. Shur, M. Shatalov, J. Yang, R. Gaska, Appl. Phys. Lett. 101 (2012) 041912.
[29] G. Tamulaitis, K. Kazlauskas, A. Žukauskas, J. Mickevičius, M.S. Shur, R.S. Fareed,
J.P. Zhang, R. Gaska, Phys. Status Solidi (c) 3 (2006) 2099.