INFLUENCE OF MORPHOLOGY AND SURFACE CONDITIONS ON DEFECT
PROPERTIES OF NANOCRYSTALLINE ZINC OXIDE
by
JORGE ANTONIO PARAMO
Bachelor of Science, 2004, 2005
Midwestern State University
Wichita Falls, Texas
Submitted to the Graduate Faculty of the
College of Science and Engineering
Texas Christian University
in partial fulfillment of the requirements
for the degree of
Doctor of Philosophy
August 2012
ii
Acknowledgements
My most sincere and deep thanks to Dr. Yuri Strzhemechny for all his help and
patience during my graduate studies at TCU. This dissertation could not have been
written without Dr. Strzhemechny, who not only served as my supervisor but also
encouraged and challenged me throughout my academic program. Dr. Strzhemechny and
other faculty members guided me through the academic process, never accepting less
than my best efforts. I thank them all.
A special thanks to all the members of our research group in the Nanoelectronics
Optical Spectroscopy lab over the years. In particular, I would like to appreciate my
colleague Dr. Raul Mugabe Peters for all his support, understanding and brotherhood. Dr.
Peters was more than instrumental in plasma processing support.
I would like to extend my heartfelt gratitude to all the faculty and staff at the TCU
Physics Department. Specifically, I would like to thank Dr. T.W. Zerda for his valuable
support and guidance as well as for his instrumentation. His assistance and suggestions
have made a difference. This work would not have been possible without the samples
supplied by Dr. Z. Crnjak Orel and her team from the National Institute of Chemistry,
Slovenia. I would also like to thank M. Murdoch, D. Yale (TCU Machine Shop), J.
Katchinska (TCU Electronics Shop), and J. Cuanzon (TCU Glass Shop) for all the
technical support provided.
My greatest level of gratitude I extend to my friends and family. I would like
express gratitude to my parents, Jorge Fabian Paramo and Blanca Esmeralda Villacis, for
their faith in me. Thank you for everything. Also I would like to thank my brothers, Juan
iii
and Daniel Paramo, for the support that they have lent me during my life in Fort Worth. I
would also like to thank my friends, especially, Ms. Aline Mora, Mr. Carlos Crespo, Mr.
Gonzalo Ferrer and Mr. Paco Cornejo for their support during my studies and for their
encouragement.
Finally, I would like to thank all the professors and staff in both the TCU College
of Science and Engineering and the TCU Neeley School of Business. I would like to
thank Mr. John Singleton and the staff at the TCU Office of International Students
Services (OISS) for their support in my development as a professional. I enjoyed the time
working as an administrative assistant in the OISS.
iv
Contents
Acknowledgements……………………………………………………………………..…ii
Contents……………………………………………………………………………….….iv
List of Figures…………………………………………………..………………………...vi
List of Tables………………………………………..…………………………………..xiii
List of Abbreviations…………………………...……………………………………….xiv
Chapter I. Introduction………………………………………………………………….1
I.1 Crystal Defects………………………………………………………………………...4
I.2 Basic Principles of Photoluminescence………………………………………………6
I.3 Photoluminescent Transitions in ZnO……………………………………………….8
I.3.1 Near Band Gap Transitions…………………………………………………………………9
a. Free and Bound Exciton Luminescence……………………………………………..9
b. Two Electron Satellites……………………………………………………………….16
c. Phonon Replicas………………………………………………………………………17
d. Donor-Acceptor Pair Emission……………………………………………………..17
I.3.2 ZnO Luminescence in the Visible Spectral Region……………………………………..17
Chapter II. Experimental Setup……………………………………………………….19
II.1 Photoluminescence System…………………………………………….…….……..19
II.2 Modifications and Troubleshooting Procedures…………………………………..23
Chapter III. Results and Discussion…………………………………………………...26
III.1 Correlation between Morphology and Defect Luminescence in ZnO
Nanopowders…………………………………………………..….……………………..26
v
III.2 Correlation between Morphology and Defect Luminescence in Precipitated ZnO
Nanorods…………………………………………..………….……………….…...........35
III.3 Correlation between Defect Luminescence and Quasi-fractal Dimensionality in
ZnO Nanostructures…………………………………………………………………….43
III.4 Growth-morphology-luminescence Correlation in ZnO-containing
Nanostructures Synthesized in Different Media………………………………………..52
III.5 Studies of BEx Luminescence in ZnO Nanopowders…………………………….64
III.6 Effects of Plasma Processing on PL Spectra of ZnO Nanopowders……………..78
III.7 Effects of Polymer Embedding on ZnO Nanopowders…………………………...91
Chapter IV. Conclusions……………………………………………………………...106
Chapter V. Future Plans……………………………………………………………...109
References……………………………………………………………………...………111
Vita
Abstract
vi
List of Figures
Figure 1.1. Diagram illustrating a photoluminescent transition involving across-the-gap
electron-hole pair recombination. An electron excited by a photon jumps from the
valence into the conduction band and then relaxes (via thermalization) to the bottom of
the conduction band. This electron then encounters a hole at the top of the valence band
and recombines with it producing a photo……...…………………………………………7
Figure 1.2. Diagram illustrating several different photoluminescent transitions from/to
states inside the forbidden band gap……………………………………………………………8
Figure 1.3. Diagram illustrating a FEx photoluminescent transition involving ground
levels of an electron and a hole bound in an exciton…………………………………………11
Figure 1.4. Diagram illustrating a photoluminescent transition involving ground levels of
a bound exciton……………………………………………………………………………………11
Figure 1.5. Excitonic range of a LT PL spectrum of a bulk single-crystalline ZnO……..13
Figure 2.1. Experimental setup for photoluminescence measurements…………………20
Figure 2.2. Diagram of the experimental layout……………………………………………..20
Figure 2.3. Screenshot of the custom-designed LabView interface………………………..23
vii
Figure 3.1. SEM images of nanocrystal samples (A)AE25; (B)AE31; (C)ZB; (D)ZA…..27
Figure 3.2. RT PL spectra for the as-received ZnO nanopowders…………………………28
Figure 3.3. Gaussian-resolved PL spectrum of the AE31 sample………………………….29
Figure 3.4. Gaussian-resolved PL spectrum of the AE25 sample………………………….30
Figure 3.5. Gaussian-resolved PL spectrum of the SA sample……………………………..31
Figure 3.6. Gaussian-resolved PL spectrum of the ZA sample……………………………..32
Figure 3.7. Gaussian-resolved PL spectrum of the ZB sample……………………………..33
Figure 3.8. LT PL spectra for the as-received ZnO nanopowders…………………………34
Figure 3.9. SEM images of the precipitated ZnO nanorods………………………………...37
Figure 3.10. Length-to-width (L/W) aspect ratio of the ZnO nanorods obtained from the
quantitative analysis of the SEM images (ca. 100 nanocrystals per sample were analyzed
in CorelDraw)……………………………………………………………………………………..38
viii
Figure 3.11. Comparison between room temperature PL spectra of the samples obtained
after 30 min and 4 hrs of synthesis in the W/EG 1/1 mixture……………………………….39
Figure 3.12. Comparison between low temperature (8K) PL spectra of the samples
obtained after 30 min and 4 hrs of synthesis in the W/EG 1/1 mixture…………………….40
Figure 3.13. Comparison between low temperature (8K) and room temperature PL
spectra of the ZnO nanorod samples obtained after 4 hrs of synthesis in the W/EG 1/1
mixture……………………………………………………………………………………………...41
Figure 3.14. SEM image of the sample obtained after 48 hrs of synthesis………………..46
Figure 3.15. SEM image of the sample obtained after 4 hrs of synthesis…………………46
Figure 3.16. SEM image of the sample obtained after 2 hrs of synthesis…………………47
Figure 3.17. Comparison between room temperature PL spectra of the samples of
variable dimensionality…………………………………………………………………………..48
Figure 3.18. Comparison between low temperature (8 K) PL spectra of the samples of
variable dimensionality…………………………………………………………………………..49
ix
Figure 3.19. Comparison between low temperature (8 K) PL spectra of the 2D sample
collected within a 1 hour interval during a continuous laser beam irradiation………….50
Figure 3.20. SEM images of samples DJ2 (A), DJ7 (B), and DJ8 (C)…………………….54
Figure 3.21(a). SEM images of sample DJ12 at lower (A) and higher (B)
magnification………………………………………………………………………………………54
Figure 3.21(b). SEM images of sample DJ13 at lower (A) and higher (B)
magnification………………………………………………………………………………………55
Figure 3.21(c). SEM images of sample DJ14 at lower (A) and higher (B)
magnification………………………………………………………………………………………55
Figure 3.22. SEM images of samples obtained after thermal treatments at two different
temperatures: 350°C (A-C) and 500°C (D-F) for 2 hours…………………………………..57
Figure 3.23. PL spectra of samples DJ2, DJ7, and DJ8…………………………………….58
Figure 3.24. PL spectra of samples DJ12 and DJ13………………………………………...59
Figure 3.25. PL spectra of samples DJ18 and DJ22………………………………………...60
x
Figure 3.26. PL spectra of samples DJ16 and DJ20………………………………………...61
Figure 3.27. PL spectra of samples DJ17 and DJ21………………………………………...62
Figure 3.28. Comparison between BEx spectral regions for ZnO nanopowders………...66
Figure 3.29. Incident power dependence of the LT PL spectra for the ZA sample………68
Figure 3.30(a). Temperature-dependent PL spectra of the ZA sample……………………69
Figure 3.30(b). Temperature-dependent PL spectra of the ZB sample……………………70
Figure 3.30(c). Temperature-dependent PL spectra of the AE31 sample…………………71
Figure 3.30(d). Temperature-dependent PL spectra of the AE25 sample…………………72
Figure 3.31. Peak intensity vs. temperature dependence for the AE31 sample…………..73
Figure 3.32. Peak position vs. temperature dependence for the AE31 sample…………..74
Figure 3.33. Peak intensity vs. temperature dependence for the ZA sample……………...75
Figure 3.34. Peak position vs. temperature dependence for the ZA sample………………76
xi
Figure 3.35. FWHM vs T dependence for the AE31 sample………………………………..77
Figure 3.36. FWHM vs T dependence for the ZA sample,,,,………………………………...78
Figure 3.37. Nitrogen plasma effects on (a) ZA, (b) ZB and (c) AE31 samples………….81
Figure 3.38. Oxygen plasma effects on the ZA sample………………………………………84
Figure 3.39. Hydrogen plasma effects on the ZA sample……………………………………85
Figure 3.40. Hydrogen plasma effects on the ZB sample……………………………………86
Figure 3.41. Hydrogen plasma effects on the AE31 sample………………………………...87
Figure 3.42. RT PL results for hydrogen plasma effects on (a) AE31 and (b) AE25
samples……………………………………………………………………………………………..88
Figure 3.43. Effects of nitrogen plasma on the AE25 sample………………………………90
Figure 3.44. STEM micrographs of ZnO nanoparticles prepared in (a) EG; (b) PD; (c)
DEG…………………………………………………………………………………………………93
xii
Figure 3.45. STEM micrographs of a ZnO/PMMA nanocomposite section with ZnO
nanoparticles (0.05 wt.%) prepared in PD........................................................................95
Figure 3.46. (a) Room temperature PL spectra of the as-received ZnO nanoparticles
prepared in different diols; (b) Low temperature (8 K) PL spectra of the as-received ZnO
nanoparticles prepared in different diols………………………………………………………96
Figure 3.47. (a) Room temperature PL spectra of the ZnO/PMMA composites with ZnO
nanoparticles prepared in different diols; (b) Room temperature PL spectrum of a pure
PMMA plate; (c) Room temperature PL signal only from the ZnO nanoparticles
embedded in the PMMA matrix. The shown spectra result from subtracting the PMMA
background luminescence, (Fig. 3.47(b)) from the “raw” spectra shown in Fig.
3.47(a)……………………………………………………………………………………………...99
Figure 3.48. Effects of incident laser intensity on the near band-gap PL spectra of a
ZnO/PMMA nanocomposite sample (with ZnO nanoparticles prepared in PD); (a) T = 8
K; (b) T = 250 K..............................................................................................................102
Figure 3.49. Integrated near band-gap PL luminescence at 9 K and 250 K as a function
of incident laser intensity for a ZnO/PMMA nanocomposite sample (with ZnO
nanoparticles prepared in PD). Straight lines represent linear fits……………………...104
xiii
List of Tables
Table 3.1. PL transitions in the ZnO nanopowders obtained from Gaussian fits………..33
Table 3.2 Brief summary of the experimental parameters-prepared at 90°C…………….53
Table 3.3. Brief summary of the experimental parameters………………………………….56
Table 3.4. Summary of the observed PL features. RT – observed in room temperature
spectra; LT – observed in low temperature (8 K) spectra…………………………………...63
Table 3.5. Fitting results for bound excitons observed in AE31, ZA and ZB samples…...74
Table 3.6. Employed remote plasma parameters……………………………………………..79
xiv
List of Abbreviations
AE – American Elements
BEx – Bound exciton
CW – Continuous wave
D – Dimensionality
DAP – Donor acceptor pair
DEG – Di(ethylene glycol)
EG – Ethylene glycol
FEx – Free exciton
FWHM – Full width half maximum
LO – Longitudinal optical
LT – Low temperature
NBE – Near band edge
PD – Propane diol
PL – Photoluminescence
PMMA – Polymethyl methacrylate
PMT – Photomultiplier tube
RF – Radio frequency
RP – Remote plasma
RT – Room temperature
SA – Sigma Aldrich
SEM – Scanning electron microscopy
SMC – Stepper motor controller
xv
TEM – Transmission electron microscopy
TES – Two-electron satellite
TSA – Toluene-sulfonic acid
TTL – Transistor-transistor logic
UV – Ultraviolet
W/EG – Water/ethylene glycol
Z (ZA, ZB) – Zochem
1
Chapter I. Introduction
Although ZnO has been a well known semiconductor with many traditional
applications (varistors, phosphors), a new thrust in research on this material started
during the 1990’s when novel growth techniques were developed to produce relatively
large quantities of high quality single-crystalline ZnO [1-6]. These studies in turn have
spun off an active development of newly discovered optoelectronic [7-9], spintronics [10-
12], biomedical [13], and nanodevices [14, 15] applications.
There are many clear advantages of ZnO making it a material of choice on many
occasions. ZnO is inexpensive, nontoxic, and its constituent elements are in abundance.
Crystal growth is reproducible and yields a high quality product. ZnO is capable of
sustaining high temperatures and high power loads making it a good candidate for
applications in harsh environments. As a semiconductor, ZnO has a direct wide band
gap, Eg ~ 3.3 eV at 300 K, very close to that of such mature materials as GaN or SiC,
which makes ZnO suitable for modern optoelectronic applications (e.g., fabrication of
light emitting devices [9-11]). An unusually large energy of free exciton dissociation in
ZnO (Ex ~ 60 meV, see below, section I.3.1) sustains existence of free bulk excitons even
above room temperature, which is very appealing for the creation of UV excitonic lasers
operating at room temperature.
Many of the beneficial characteristics of ZnO as well as hindrances with their
implementation are rooted in the crystal defects. Defect states can fundamentally affect
such physical properties as conductivity, luminescence, etc. However, as of today,
despite decades of vigorous research, the nature of defects in ZnO is still an open and
2
actively debated topic with many challenges to clear. For example, as-grown ZnO
always exhibits an unintentional n-type conductivity. Historically, defects like oxygen
vacancies and zinc interstitials (see below, section I.1) were assumed to determine the
conductivity type [16]. However, recent first-principle calculations suggested that native
point defects may not be the reason for an unintentional n-type conductivity, [17] with
hydrogen and group III contaminants being among prime candidates for shallow donors
[18]. Synthesis of a high-performance p-type ZnO remains a challenge detrimental for
further practical advances, since numerous potential ZnO applications (such as excitonic
lasers or spintronics devices) require a p-type material [19].
Recently studies of nanoscale ZnO have become one the most popular fields in
materials research [20-22]. The reduction in the crystal size down to the nanoscale brings
about qualitatively new phenomena in this material. As the surface and sub-surface states
become more prevalent compared to bulk states. Nanocrystalline ZnO can be
synthesized in a wide range of sizes, morphologies, dimensionalities, doping
stoichiometries, and surface functionalities [23-26]. The scope of applications for
nanoscale ZnO is growing by the day in such technology domains as piezoelectric
nanoactuators [27, 28], random lasing [29, 30], nanocantilevers [31-33], sensors [34, 35],
etc. Remarkably, most of these novel applications do not require a p-type conductivity.
For example, employment of different nanoscale morphologies (nanorods in ref. 24
multipod or star-like nanostructures in refs. 25, 30) reveals an outstanding dependence of
luminescent effects on the predominant crystalline shapes. By the same token, nanoscale
ZnO reveals exceptional piezoelectric characteristics [27, 28]. Thereby, arrays of aligned
ZnO nanorods could serve as piezoelectric power supplies for nanoscale and microscale
3
devices [36] (such as nanobots injected into human blood vessels using natural variation
of blood pressure [37]). Also, in ZnO-based biosensors biological and inorganic
components are mixed to detect toxic substances, cancer cells, etc. Due to its nontoxisity,
nanoscale ZnO can be implanted into various biological systems to act as a transducer.
ZnO nanobelt cantilevers can be used in, e.g., scanning probe microscopy. They have
been shown to significantly increase the cantilever sensitivity. Also, the range of lengths
in which ZnO nanowires are grown allows effective tuning of the resonance frequency
depending on the application.
There are many types of crystal defects affecting these nanotechnology
applications. One of the most important sources of defects in nanosize ZnO is the surface
of the material. Generally, the surface of the crystal could be a significant generator of
lattice defects and contaminating impurities due to the formation of dangling bonds,
surface reconstruction and exposure to the environment. The influence of the surface
defect states often extends also into the sub-surface vicinity. Hence, it is natural to
expect that in ZnO nanocrystals, on the one hand, the influence of the surface-related
defect states grows with the decreasing crystal dimensions, and, on the other hand, the
quality of the surface and the subsurface vicinity is an important control parameter.
Despite their importance, understanding of surface defects and defect-related properties
in ZnO nanostructures is far from complete despite certain progress in recent years [38-
43]. Very few studies addressed the question of how to controllably manipulate the
surface states in nanoscale ZnO [44-47].
Thereby, in our work we set the goal to advance understanding of the surface
defect properties in ZnO nanocrystal systems, in particular their dependence on the
4
geometry of the nanostructure (size, morphology, dimensionality) as well as on the
modifications of the surface. We investigated primarily optoelectronic properties of
defects in nanoscale ZnO, and, specifically, defect-mediated radiative recombinations as
measured by the photoluminescence (PL) spectroscopy. PL is a powerful nondestructive
experimental technique, easy to setup, and capable of yielding detailed information on a
variety of processes involving electronic structure of the studied material.
I.1 Crystal Defects
In this section we provide some definitions and assertions helpful for general
understanding of defects in ZnO.
ZnO is a II-VI compound semiconductor with four atoms per primitive cell of a
hexagonal wurtzite structure, which is thermodynamically predominant under ambient
conditions. In reality, the periodicity of the ZnO lattice (as of any crystal structure) is not
perfect and is usually disrupted by crystal defects. There are several classification
categories of defects as follows.
For example many defects can be classified as either donors or acceptors. Donors
share one or more electrons with the crystal, whereas acceptors capture electrons,
increasing the concentration of holes. On the other hand, native defects are those formed
by the constituent atoms of the material, while those formed with the participation of
foreign atoms are called extrinsic (or impurity) defects. Therefore, in ZnO native defects
will be associated with deficiency, abundance, or misplacement of zinc and oxygen
species in the lattice. Defects could be also categorized based on the dimensionality (0D,
1D, 2D, etc.) of their occurrence in the crystal. For example the 0D defects do not extend
5
in any spatial dimension and thus are called point defects. In turn, point defects can be
sub-classified as vacancies, interstitials, substitutional impurities, or anti-sites. Vacancies
are vacant sites in a crystal, interstitial defects are atoms placed outside the lattice pattern,
substitutional impurities refer to native atoms being replaced by foreign ones, and anti-
sites refer to replacement of native atoms into a wrong lattice cite in ordered alloys. E.g.,
native point defects in ZnO would be oxygen vacancies (VO), zinc interstitials (Zni), etc.
Higher-dimensional defects are also called extended defects and clusters. Edge
dislocations (terminations of a plane of atoms in the middle of a crystal) or screw
dislocations (structures, in which a helical path is traced around linear defects by atomic
planes) are examples of extended defects. The grouping of two or more adjacent defects
is called a defect cluster.
Defects in crystals originate from numerous sources, and crystal surface is one of
them. The surface conditions in nanostructures play a remarkable role. Crystalline
surface is the largest disruption of the periodic lattice at which the broken (dangling)
chemical bonds have to terminate by attracting impurities from the environment or to
reconstruct thus accumulating structural imperfections. Moreover, the surface defect
states and their influence may extend into the sub-surface layers of the crystal. Given
this, the increase of the surface-to-volume ratio with the crystal size reduction increases
the contribution of the surface defect states making it predominant in nanocrystals. Thus,
in ZnO nanostructures, the manifestation of surface defects should increase with the
decreasing crystal size. Similarly, the relative contribution of the bulk defects should
decrease. This aspect coupled with the fact that ZnO is a very luminescent material
6
makes PL spectroscopy a very suitable probe of surface and sub-surface defect properties
in nanoscale ZnO.
I.2 Basic Principles of Photoluminescence
Operation of most semiconductor devices is based on creation of charge carriers.
One of the mechanisms to create excess carriers is optical excitation. On the other hand,
luminescence is the emission of light by a material when excited electrons lose their
energy via photon emission. If carriers are excited by photon absorption, the radiation
resulting from recombination of excited carriers is called photoluminescence (PL).
Light can be emitted by a semiconductor via numerous different channels. The
simplest example of light emission by a semiconductor is a direct recombination of a free
electron-hole pair across the band gap (Fig. 1.1). Otherwise carrier recombination and
emission of light may involve transitions from/to states inside the forbidden band gap.
As those defects may act as acceptors or donors, free electrons and holes seek abundance
or shortage of charge. Given suitable conditions, free carriers can be trapped at these gap
states. If the carriers recombine further and emit radiation, the emitted light could be
used to determine the energy of the transition involving the defect state. Defect states
may be shallow or deep. If the temperature is low enough to prevent thermal activation
of the carriers from the traps, radiative recombination is more favorable for shallow
defects, since they are closer to the band edges. Figure 1.2 illustrates several different
scenarios of such transitions.
7
e-
Ev
Ece-
h+
Exchange
with Latticee-
Figure 1.1 Diagram illustrating a photoluminescent transition involving across-the-gap electron-hole pair
recombination. An electron excited by a photon jumps from the valence into the conduction band and then
relaxes (via thermalization) to the bottom of the conduction band. This electron then encounters a hole at
the top of the valence band and recombines with it producing a photon.
8
Shallow Donors
Shallow Acceptors
Deep Levels
e-
e- e- e-
h+
e-
e- e-
e-h+
h+ e-
h+h+
h+h+ h+ h+
e- e-
h+
EC
EV
e-
e-
Laser Beam
Exchange
with Lattice
Figure 1.2. Diagram illustrating several different photoluminescent transitions from/to states inside the
forbidden band gap.
The time scales of luminescent events depend on the probability of thermal
excitation into/from the trap. If the probability of thermal excitation is small, the time for
excitation and recombination is long. Even longer times can occur when the electrons are
trapped several times before recombination. If the trapping probability is greater than the
probability of recombination, an electron may make several trips between the trap and the
valence band.
I. 3 Photoluminescent Transitions in ZnO
In ZnO many defect properties are detectable employing PL spectroscopy. In this
section we will describe the most PL common transitions observed in ZnO and their
qualitative description.
9
In general, ZnO is a highly luminescent material. PL spectra for ZnO can be
collected in a wide range of temperatures, with various transitions dominating different
temperature domains. For example, room temperature (RT) PL shows relatively narrow
near band gap peaks in the UV and broad emission bands in the visible range of the
spectrum. Low temperatures yield narrow luminescent features related to bound exciton
luminescence, two electron-satellite transitions, phonon replicas, etc. PL peak positions,
widths, and intensities as well as dependencies of these on temperature and incident light
power provide rich information about the defect properties. Below is a brief description
of the most frequently observed PL transitions in ZnO.
I.3.1 Near Band Gap Transitions
a. Free and Bound Exciton Luminescence
When a semiconductor absorbs a photon with energy high enough to move an
electron from the valence band to the conduction band, a free electron and a free hole are
created. However, if the temperature is low enough and if the hole and the electron are
close enough, they can become bound via a modified Coulomb interaction thus forming a
hydrogen-like system called an exciton. Excitons carry no net charge. Excitonic energy
levels are discrete and can be approximately determined using the hydrogen atom model
with several modifications accounting for the effective masses of electrons and holes,
screening of the Coulomb interaction, etc. Since band gap edges represent a boundaries
of the quasi-vacuum continuum for electrons and holes, discrete excitonic energy levels
appear inside the forbidden gap.
10
Free excitons (FEx) are defined as those not localized by the crystal lattice. Their
ionization (dissociation) energy is remarkably high in ZnO – 60 meV – in excess of
energy provided by room temperature (~ 25 meV), thus leaving many free excitons intact
at room temperature, a peculiarity very valuable for potential applications. Ground state
FEx recombination process is shown in Fig. 1.3.
On the other hand, at relatively low temperatures, disruptions of crystal lattice due
to defects can create potential wells for the free excitons and trap them in bound states –
bound excitons (BEx). Since the nature of such localization is primarily dipole, the
energy required to liberate BEx from the trap is smaller than the ionization energy of
FEx.
At low temperatures, upon raditative recombination of an electron and a hole in
BEx, the binding defect is usually found in its ground state. Such transitions in ZnO give
rise to spectral lines in the energy range of 3.348 – 3.374 eV [48]. LT FEx
recombinations are usually found in the 3.375 – 3.380 eV [49] range and are normally
orders of magnitude weaker than those of BEx. Such dominance of BEx luminescence
occurs because at low temperatures most of the excitons “freeze out” via binding with
defects and thus the probability of finding a FEx capable of producing a luminescent
transition is small. Fig. 1.4 shows BEx recombination process involving ground states of
the exciton proper and those of the binding potential well.
11
Figure 1.3. Diagram illustrating a FEx photoluminescent transition involving ground levels of an electron
and a hole bound in an exciton.
Figure 1.4. Diagram illustrating a photoluminescent transition involving ground levels of a bound exciton.
h
Ec
Ev
Eo, FEx Electron
h
Ec
Ev
Eo, BEx Electron
Eo, BEx Hole
Eo, FEx Hole
Modified Coulomb
Interaction
Dipole-Dipole Interaction
h+e-
D
h+
h
Ev
Ec
Eo, FEx Electron
Eo, FEx Hole
Modified Coulomb
Interaction
e-
12
Most detailed and up-do-date treatise of the BEx PL transitions in ZnO is given in
recent reviews [19]. It is worth mentioning some important BEx recombination related to
common impurities. According to Van de Walle [18], hydrogen may produce n-type
conductivity inherent to ZnO. Hydrogen forms a strong bond with oxygen and entering
the ZnO crystals is relatively easy. Moreover, in ZnO, hydrogen is non-amphoteric [18]
and is always a donor. Several transitions due to hydrogen have been identified, the most
prominent being the BEx line located at 3.3628 eV in the LT PL spectrum of ZnO [48].
Luminescence results presented for nitrogen-doped ZnO show an acceptor bound exciton
at 3.32 eV. The acceptor was identified as substitutional nitrogen on oxygen. This is
probably a shallow acceptor ~ 165 meV above the valence band. If the nitrogen acceptor
is activated, e.g. through annealing, it may lead to a reduction of the donor concentration
[48]. Hydrogen has, also, been identified as the reason for the passivation of the 3.3313
eV emission [50].
In Fig. 1.5 we show a LT PL spectrum collected by us at 10 K on a high-quality
single crystalline ZnO sample from Cermet Inc. One can see typical rather narrow
excitonic features: several FEx features corresponding to different dispersion branches of
the band gap as well a number BEx peaks related to recombination on various defects.
In principle, peak shape of the PL luminescence stems from a convolution of
several factors, such as instrumental resolution, thermal broadening, microscopic nature
of radiative transitions, etc. Although more precise fitting can be obtained by employing
multi-peak Voigt functions, decent results can be also obtained by a much more
straightforward fitting using a superposition of Gaussian curves. Temperature strongly,
13
sometimes dramatically, affects such parameters of luminescent peaks as shape,
integrated intensity, position, full width at half maximum (FWHM), etc. Substantial
3.30 3.32 3.34 3.36 3.38
PL
In
ten
sit
y,
a.u
.
E, eV
Figure 1.5. Excitonic range of a LT PL spectrum of a bulk single-crystalline ZnO.
amount of information can be obtained from the analysis of evolution of these parameters
with temperature. For example, information about energetics and mechanisms of BEx
dissociation can be derived from the temperature dependence of the integrated intensity
of their PL emission.
In our work to fit the experimental results we employed the following dependence
of the BEx peak intensities on temperature [19]:
kTE
o
aAe
ITI
/1
)(
14
Here I0 is the peak intensity at 0 K, A – a fitting parameter, k – the Boltzmann
constant and Ea – the activation energy. Such temperature dependence can be derived as
follows.
BEx in the ground state can obviously be thermally activated into another, higher
energy, states. In a simplified model we assume that there exists only one predominant
channel of exciton activation. Within this model, at certain temperature, the number of
BEx in the ground state Ng(T) plus the number of excitons in the higher level state Ne(T)
will give the total number of excitons:
Defining energies of the ground state and the excited state levels as Eg and Ee
respectively, we will apply Boltzmann statistics to describe the number of excitons with
these energies at temperature T:
Here ae and ag are the degeneracies of the states with energies Ea and Eg
respectively and Z is the partition function. If we presume that the total number of
excitons is conserved (which is a good assumption for ZnO in a wide range of
temperatures), then the value of Nt(T) is a constant, since at 0 K all excitons are in the
ground state: Nt(T) = Nt(0) = Ng(0). Combining these expressions one obtains:
)()()( TNTNTN teg
ZeaTNTNkTE
etee /)()(
/
ZeaTNTNkTE
gtgg /)()(
/
15
kTE
kTE
g
kTE
e
g
eeg
g
T
g
a
g
e Ae
ea
ea
TN
TNTNTN
TN
TN
TN/
/
/1
1
1
1
)(
)(1
1
)()(
)(
)(
)(
Here A = ae/ag and Ea = Ee – Eg, the activation energy necessary to activate BEx
in its ground state to a higher energy state. In PL experiments the intensity of the BEx
luminescence is proportional to the number of bound excitons at certain temperature.
Therefore
By the same token, peak positions of the BEx emission are changing with
temperature due to several factors [51]. Comprehensive theoretical description of the
peak position vs. temperature dependence is very complex, and oftentimes this
dependence is approximated with an empirical formula [52],
Here E0, α and β are fitting parameters. Normally, parameter β is directly related to the
Debye temperature.
This formula reflects the fact that the variation of the energy gap with temperature
is linear at high temperatures and nonlinear at low temperatures.
Also, analysis of the BEx peak broadening with temperature yields account of the
exciton-phonon scattering mechanisms.
Most of the time BEx are trapped by relatively deep potential wells created by
either donor-or acceptor-like defects. In such case the localization radius of BEx and the
T
TETE
2
0)(
kTE
t
g
oaAeN
TN
I
TI/
1
1
)0(
)()(
16
corresponding PL FWHM are rather small for both neutral and ionized (charged) defects.
However, sometimes the impurity element has the same number of outer shell electrons
as the surrounding crystal matrix – the so called isoelectronic impurity. Naturally,
isolelectronic traps are shallow and excitons bound by such defect will be less localized.
Thus the expected linewidth of the related BEx luminescence will be greater and the
corresponding activation energy – smaller.
If an exciton becomes bound by the defect at or in the vicinity of the surface, the
effective depth of the potential well in such case is somewhat shallower leading to
weaker localization and a broader linewidth, similar to the case of isoelectronoic
trapping.
b. Two Electron Satellites
Ordinary BEx transitions, described above, happen when the binding defect is left
in its ground state after the recombination. However, there is a non-zero probability that
this recombination will instead result in the defect left in an excited state after the
recombination. This gives rise to a red shifted satellite line, the energy of which
corresponds to the BEx energy minus the energy required to excite the defect into its
excited state. Such transitions are commonly referred to as two-electron satellite
transitions (TES). They appear in the 3.32 – 3.34 eV range of the LT PL spectra for ZnO
[53]. TES transitions can be very instrumental in determining the ionization energy of the
binding defect. E.g., if the recombination leaves the defect atom in its first excited state,
and we know the energy for the BEx recombination, we can determine the ionization
energy Ed of the binding defect itself. Given that the defect spectrum levels are
17
hydrogen-like, we can approximate its energy levels as En = – Ed/n2. The ground state
energy is E1 = – Ed and the first excited state energy is E2 = – Ed/4. Thus, Ed is 4/3 of the
E2 – E1 interval (the red-shift of TES with respect to BEx).
c. Phonon Replicas
Numerous optical transitions may produce more than just a residual photon.
Luminescence is often accompanied by emission of multiple longitudinal optical (LO)
phonons resulting in the so called spectral phonon replicas extending sometimes over a
substantial ranges of energy.
d. Donor-Acceptor Pair Emission
Closely spaced donors and acceptors in the crystal can form pairs interacting via
Coulomb attraction. Optical transitions between such pairs are called the donor-acceptor
pair (DAP) recombination. For the DAP case, the transition energy will be determined
by the separation between the donor and acceptor levels as well the distance-dependent
Coulombic interaction. In ZnO these transitions appear in the 2.95 – 3.25 eV range [4].
DAP luminescence provides in many cases additional insight into defect properties.
I.3.2 ZnO Luminescence in the Visible Spectral Range
In ZnO, there is a variety of defect-related optical transitions in the visible part of
the spectrum. A number of different luminescent bands were reported, such as a green
band at ~ 2.5 eV, a yellow band at ~ 2.2 eV, an orange band at ~ 1.9 eV, a blue band at ~
2.9 eV, etc.
18
Usually, the most prominent band is the 2.5 eV green luminescence. Various
models were proposed to explain it. Vanheusden, et al. [54], modeled the green emission
as a recombination of oxygen vacancies with holes in the valence band. Kohan, et al. [49]
suggested that the green emission originates only from Zn interstitial defects. Garces, et
al. [55] assigned the green PL band to an electron-hole radiative recombination involving
a DAP. Su [56] referred the green band to copper-related defects. Other models were
suggested as well.
The yellow and orange emission bands at ~1.9 eV and ~2.2 eV may originate
from deep levels with different initial states (conduction band and shallow donors).
These bands exhibit dependency on wavelength of excitation [57]. Oxygen interstitials as
well as Li impurities were suggested to be responsible for the yellow luminescence [58].
The orange-red emission, on the other hand, was attributed to the presence of excess
oxygen in the samples [59]. Other hypotheses include surface dislocations and zinc
interstitials [60].
Li and Cu are cations that act as acceptors in ZnO, according to Kasai [60]. In
Ref. 58 it was shown that Li introduces deep acceptor states. Similar observations were
reported for Na-doped ZnO. The PL of Li- and Na-doped crystals gives rise to DAP
recombinations in the green and yellow spectral ranges.
19
Chapter II. Experimental Setup
As mentioned previously, in our studies we employ PL spectroscopy as a
powerful yet relatively simple nondestructive optical technique to probe defect-related
and other optoelectronic properties of the systems of interest. In this section, we describe
the experimental system designed with my direct participation to run PL experiments at
the TCU Optical Spectroscopy Lab, as well as the modifications and troubleshooting
devised to enhance the performance and functionality of the experimental setup.
II.1 Photoluminescence System
As a rule, photoluminescence experiments require an optical bench assembly
consisting of a laser, as an excitation source, optical components (mirrors, lenses, filters,
etc.), a monochromator to differentiate the light wavelengths, a detector of the
luminescence signal, and an optional cryostat for temperature-dependent measurements.
Additionally, the setup needs a computer interface, a signal amplifier and an optional
lock-in amplifier coupled with an optical chopper for noise reduction. An optical bench
provides stability, minimizes vibration-related perturbations and locks the optical train
alignments. Lenses, mirrors, etc. guide primary (laser) and secondary (luminescence)
photons.
At this moment we have a working PL optical train that includes all the above
mentioned components. A picture of the complete assembly is shown in Fig. 2.1 and the
diagram of our experimental layout – in Fig. 2.2.
20
Figure 2.1. Experimental setup for photoluminescence measurements
Figure 2.2. Diagram of the experimental setup
The laser employed in the experiments discussed here is a HeCd Kimmon
continuous wave (CW) laser, model IK5452R-E. It is capable of producing a UV
21
radiation of 200 mW at 325 nm and a visible radiation of 300 mW at 442 nm. This laser
can also be used in a dual wavelength mode (UV and visible beams at the same time).
There are other lasers, available in our setup (such as pulsed nitrogen laser), that were not
used in our experiments.
The Janis CCS-150 closed cycle helium gas cryostat provides temperatures
controllably between 8 K and 325 K. A water cooled compressor unit cools down the
working helium. The cryostat requires a mechanical vacuum pump for startup at
pressures of less than 10 mtorr. The cryostat stabilizes the temperature via a built-in
heater. For the cryostat we designed a sample holder that can be screwed into the
original copper cold finger of the cryohead. The sample holder is multi-purpose and
allows mounting a variety of different types of samples (bulk fragments, powders,
liquids, etc.). In addition, we constructed a stand for the cryostat. It was made of pre-
fabricated aluminum parts and two stainless steel rods elevating the cryostat 5 inches,
making the sample holder stand at 17.5 inches above the surface of the optics table (the
exact height of the monochromator entrance slits).
The monochromator used in our studies is SPEX 1401. Its spectral resolution is
0.18 cm-1
with an optical path of 4 meters and an f# of 7.8. The gratings turret of this
Czerny-Tuner monochromator is run by an 8 phase stepper motor coupled to a stepper
motor controller (SMC), and this combination provides high precision and spectral
accuracy control of ± 1 cm-1
. The output slit assembly of the monochromator is
connected to a PR C31034 photomultiplier tube (PMT) detector in a water cooled
housing (PC104CE) from Products for Research Inc. The PMT receives the light coming
from the monochromator and transforms it to a current output read by a lock-in amplifier
22
from Stanford Research. The amplifier converts this signal into a DC voltage signal
which is fed to a computer.
We developed an in-house software interface based on the LabView programming
language to control the experiment and collect the data. This arrangement is aided by an
optical chopper Oriel/Newport coupled to the lock-in amplifier capable of processing the
signal and performing mathematical calculations to reduce the experimental noise. The
computer receives the signal through a GPIB National Instruments card which is
controlled by our custom-made software. The screenshot of the software interface is
presented in Fig. 2.3. The program sends a transistor-transistor logic (TTL) signal to the
SMC for a specific period of time. The times needed to gather a specific number of
points depend on the desired range of wave numbers and a time constant. Real time data
can be seen on the monitor, while it is archived in a matrix form for further processing.
The program has an additional input to record into the data file all the relevant
experimental information (type of material, experimenter, time constant, range of wave
numbers, etc.). All this information along with the data matrix is then sent to an ASCII
file retrievable with most programs handling spreadsheets (Excel, Origin, WordPad, etc.).
The smaller optical components (lenses, mirrors, component holders, etc.)
complying with the spectral range of our experimental needs were acquired from several
vendors (such as Oriel/Newport, Edmunds). Most of the lenses and mirrors were of a UV
grade.
In addition, our setup is rather versatile and can be relatively simply rearranged to
accommodate other experiments (such as, e.g., Raman scattering spectroscopy).
23
Figure 2.3. Screenshot of the custom-designed LabView interface
II.2 Modifications and Troubleshooting Procedures
In the process of design, assembly and subsequent running of the current PL setup
we encountered multiple problems and holdups. They were all resolved via successful
troubleshooting and modifications.
The original design included a fiber optics assembly collecting luminescence and
guiding it into the monochromator. A fiber optics bundle was employed because of its
flexibility and ease of tune-up. In order to couple the fiber optics to the spectrometer,
several different fixtures were custom-made for the assembly. However, it was
discovered that the scattered laser light, focused into the fiber bundle, excited parasitic
luminescence in the material of the fiber (fused silica). Also, the f# of the fiber optics
assembly differed greatly from the one of our monochromator. As a result, the observed
24
spectra were inadequate. Thereby, the optical bench required major changes after the
problem was detected: a relocation of the monochromator on the optics table and building
a mirror/lens assembly instead of the fiber optics setup. This modified optical train is
currently used in the TCU Optical Spectroscopy Lab, featuring excellent f# matching
with the monochromator as well as negligible parasitic luminescence in the mirrors and
lenses.
To generate the most adequate experimental automation control and data
acquisition we tried several different versions of the SMC software: (a) running the
monochromator without a stop-control or time calculations; (b) calculating the time
required to collect data points without an integration time constant; (c) applying the time
constant used to improve signal-to-noise ratio and gathering data within selected time
periods. These versions of the program used a coaxial line communication from a
LabView interface box (BC 2110) and sent a TTL level signal for interfacing. The first
version of the program was intended to allow the user to control the collection time
thorough a physical coaxial connection from a National Instruments interface box and the
LabView software. In other words, a time constant would be used to collect data points
for a fixed wavenumber. The user would have the option to select the desired
wavenumber range, spectral step and the time constant. The software was intended to
integrate the signal for a given wavelength depending on the time constant. After
integration, the monochromator would move to the next point. However, the TTL signal
intended to actuate the controller produced noticeable mechanical failure on the stepper
motor. Thus, the next version was created in which the time constant feature was
removed. Instead, the data collection was intended to happen for a user-specified range
25
and speed, with no stoppage. For low speeds, however, the mechanical failure of the
motor was still apparent. These software solutions were abandoned because of the bad
reception of the signal by the SMC, apparently due to inaccurate translation of the signal
by the interface box.
In an effort to solve this problem we employed a direct communication with the
RS-232 (24 pin assembly) port of the controller box instead of the coax line. The
programming became a lot more intricate, since each pin represented a different function
of the stepper motor speed. The movement, however, was erratic and the signal through
each pin was hard to control due to several conflicting protocols between the software
and the controller box.
The current final version of the interface software is a simplified protocol
combining automated data gathering and processing with a manual control of the spectral
range settings.
We refurbished (twice) the design of the PMT housing, which initially had inside
a focusing lens made of a UV-opaque material (plastic), hindering collection of the PL
spectra for ZnO. The lens assembly was replaced with a new one made of a UV grade
fused silica with the transmittance complying with the spectral ranges of interest.
Finally, there were spurious spectral lines observed in numerous experiments. It
was eventually surmised that these lines originated from an internal tune-up flaw of the
monochromator. The alignment of the monochromator was not trivial and eventually
required a professional service person to complete.
26
Chapter III. Results and Discussion
As stated above, the goal of this research effort was to advance understanding of
the surface defect properties in nanoscale ZnO, their correlation with the geometry (size,
morphology, dimensionality) of the nanostructure as well as the influence of surface
modifications. This chapter discusses results obtained within the outlined program.
III.1 Correlation between Morphology and Defect Luminescence in ZnO
Nanopowders [51, 62]
Commercially sold ZnO nanopowders made a first natural choice of the samples
to study due to their availability in large quantities. They represent an important class of
nanoscale ZnO due to the numerous applications in electronics, agriculture, chemical
industry, energy production, cosmetics, etc. [19, 32] These nanopowders can be
purchased from multiple vendors in a wide range sizes and morphologies.
Nanopowder ZnO samples were obtained from several suppliers: Sigma Aldrich
(SA), American Elements (AE), and Zochem (Z). American Elements and Zochem
provided powders with different average nanocrystal sizes. The nomenclature for the
five samples studied is: SA, AE25, AE31, ZA, and ZB. Scanning electron microscopy
was used to characterize nanocrystal size distribution and morphologies of the samples
(see Fig. 3.1). The average sizes were found to be as follows: AE25 < 25 nm, AE31 ~ 80
nm, SA ~ 90 nm, ZA ~ 180 nm and ZB ~ 210 nm (30 to 50 nanocrystals per sample
analyzed).
27
Figure 3.1. SEM images of ZnO nanopowder samples (A) AE25; (B) AE31; (C) ZB; (D) ZA.
Room temperature (RT) PL spectra of the as-received samples are shown in figure
3.2. One can see a substantial discrepancy between the spectra of different nanopowders,
indicating significant sample-to-sample differences in the abundance of optically active
defects. In addition to the near-band edge (NBE) transitions observed in all the samples
at ~ 3.2-3.3 eV, the defect emission bands with variable relative intensities appear at ~
3.1. eV (AE25), ~2.9. eV (SA), ~ 2.4. eV (AE31, ZA, ZB, SA), ~ 2.1. eV (AE25). One
can see only a limited correlation between the RT PL spectra and the average grain size
in a given nanopowder, pointing to the prevailing influence of the growth history and the
quality of specific specimens over generic scaling features. Notably, powders from the
same vendor with different size distributions (AE25 and AE31) reveal different spectral
shapes. We performed multi-peak Gaussian fitting of the obtained RT PL spectra, the
results of which are shown in Figures 3.3-3.7 and also summarized in Table 3.1.
28
2.00 2.25 2.50 2.75 3.00 3.25
PL
In
ten
sit
y,
a.u
.
E, eV
AE31
AE25
SA
ZA
ZB
Figure 3.2. RT PL spectra for the as-received ZnO nanopowders.
29
1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6
AE31
~ 3.2 eV
~ 2.9 eV
~ 2.4 eV
~ 2.2 eV
PL
In
ten
sit
y , a
.u.
E, eV
Figure 3.3. Gaussian-resolved PL spectrum of the AE31 sample.
30
1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6
~ 3.3 eV~ 3.13 eV
~ 2.9 eV
AE25~ 2.20 eV
PL
In
ten
sit
y , a
.u.
E , eV
Figure 3.4. Gaussian-resolved PL spectrum of the AE25 sample.
31
1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6
SA ~ 3.2 eV~ 2.9 eV
~ 2.3 eV
PL
In
ten
sit
y , a
.u.
E , eV
Figure 3.5. Gaussian-resolved PL spectrum of the SA sample.
32
1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6
~ 3.20 eV
~ 2.9 eV
~ 2.3 eV
ZA
PL
In
ten
sit
y , a
.u.
E ,eV
Figure 3.6. Gaussian-resolved PL spectrum of the ZA sample.
33
1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6
~ 3.2 eV
~ 3.0 eV
~ 2.3 eV
ZB
PL
In
ten
sit
y , a
.u.
E , eV
Figure 3.7. Gaussian-resolved PL spectrum of the ZB sample.
Sample Approximate transition energy (eV)
AE31 2.2, 2.4, 2.7, 2.9, 3.1, 3.2
AE25 2.2, 2.9, 3.1, 3.2
SA 2.4, 2.9, 3.0, 3.2
ZA 2.4, 2.9, 3.1, 3.2
ZB 2.4, 3.0, 3.1, 3.2
Table 3.1. PL transitions in the ZnO nanopowders obtained from Gaussian fits.
34
Low temperature (8 K) PL spectra of the studied nanopowders are presented in Fig. 3.8,
where one can observe a similar limited correlation between the average nanocrystal size
and the luminescent features.
2.00 2.25 2.50 2.75 3.00 3.25
PL
In
ten
sit
y, a.u
.
E, eV
SA
AE25
AE31
ZB
ZA
Figure 3.8. LT PL spectra for the as-received ZnO nanopowders.
From the measurements performed on the commercial ZnO nanopowders we
concluded that although the studied objects were of essentially the same morphology and
chemical composition, the variance of their origin and growth history might have
overshadowed observation of a surface-related scaling behavior of the defect
luminescence (or, paraphrasing: “not all the ZnO is created equal”). It became evident
that in order to avoid the aforementioned impediments and to be able to observe the
35
expected scaling effects, nanoscale ZnO with variable geometries should come from a
single source.
Through collaboration with the National Institute of Chemistry in Ljubljana,
Slovenia (Dr. Crnjak Orel’s group) we gained access to an assortment of high-quality
nanoscale ZnO synthesized in a wide variety of morphologies with a high degree of
growth control.
III.2 Correlation between Morphology and Defect Luminescence in
Precipitated ZnO Nanorods [63]
A low-temperature wet precipitation growth approach, initially suggested in [64],
satisfied the requirements of a low-budget multi-parameter controlled synthesis of high-
quality hexagonal ZnO nanorods, necessary for our research program. In this section we
demonstrate that variation of the growth time and the composition of the solvents,
provide an accurate control of the size and morphology of ZnO nanocrystals.
Furthermore, we established a direct correlation between the morphology of the particles
and their defect emission within a set of samples synthesized using a procedure with a
consistent size-control.
ZnO nano- and submicrometer sized rod-shaped structures were prepared by a
homogeneous precipitation method at a relatively low temperature (90°C) from zinc
nitrate and urea in a water/ethylene glycol (W/EG) mixture. Fresh stock solutions were
prepared from Zn(NO3)2×6H2O (Aldrich) and urea (Aldrich) in MilliQue water. All
reagents were analytically pure. The concentration of urea was kept constant at 0.05 mol
dm-3
while the concentration of the Zn2+
ions was 0.01 mol dm-3
. The synthesis was
36
performed employing continuous mixing in 250 ml open reactors. Different solvents
were used – water and three W/EG mixtures with the volume ratios of 3/1, 1/1 and 1/3.
Different growth times were applied. More detailed discussion of the growth procedure
can be found in Ref. 64.
Morphology and structural properties of the synthesized samples were
characterized by the scanning electron microscopy. Particle sizes and size distributions
were determined from the SEM images using the CorelDraw software (100 nanocrystals
were analyzed per sample). Fig. 3.9 shows SEM images of the studied samples for a
range of different growth times and solvent compositions. One can observe a common
trend – both the longitudinal and transverse average dimensions of the obtained
hexagonal nanorods are increasing with the growth time as well as water content in the
solvent. CorelDraw-based quantitative analysis of the SEM images summarized in Fig.
3.10 confirms such dependence of the growth dynamics (with an exception of a few
outliers). Synthesis of similar ZnO nanorod networks was reported in a few recent
publications [65-67], e.g. in Ref. 65, where the authors employed a low-temperature
seeded growth based on a zinc nitrate hydrate/hexamethylenetetramine solution to
produce nanowire network transistors. However, practically no studies were reported on
the growth parameters controlling dimensions and quality of the nanocrystals obtained.
Our preparation approach provides well-defined size-controlling parameters.
For most of the samples reported here, the majority of the nanocrystals reveal
transverse grooves close to the middle of the rod. At shorter growth times (and hence
smaller rod sizes) this feature is more distinct, yielding an appearance of a dumbbell-
shape bipod. Such feature was previously reported [68, 69] and has been attributed to
37
twinning mechanisms. In Ref. 69, it was suggested that ZnO nanostructures with
dumbbell twinning may be used in applications for second harmonic generation in
nonlinear optics.
Figure 3.9. SEM images of the ZnO samples synthesized with different W/EG concentrations and for
different times: (a1) W/EG 1/3, 30 min; (a2) W/EG 1/3, 45 min; (a3) W/EG 1/3, 4 h; (b1) W/EG 1/1, 30
min; (b2) W/EG 1/1, 45 min; (b3) W/EG 1/1, 4 h; (c1) W/EG 3/1, 30 min; (c2) W/EG 3/1, 45 min; (c3)
W/EG 3/1, 4 h.
Thus, the suggested wet solution-phase approach provides inexpensive and
effective routine for growing well-defined hexagonal ZnO nanorods with an accurate
control of the nanocrystal length and thickness through synthesis time and solvent
contents. Discussion of the possible underlying mechanisms of such dependence on the
time and chemistry is provided in Ref. 64.
38
0 50 100 150 200 2500
2
4
6
8
10
12
14
As
pe
ct
rati
o,
Le
ng
th /
Wid
th
Synthesis time, min
Solvent composition
Water
W/EG 3/1
W/EG 1/1
W/EG 1/3
Figure 3.10. Length-to-width aspect ratio of the ZnO nanorods obtained from the quantitative analysis of
the SEM images (ca. 100 nanocrystals per sample were analyzed in CorelDraw).
We employed PL measurements to verify how the morphology of these nanorod
samples affects their optoelectronic characteristics. Consistent sampling of ZnO
nanostructures provided by the growth procedure allows one to address the issue of size-
scaling of defect luminescence. In the nanorod samples we observed a strong correlation
between their morphology and defect properties as illustrated by Figs. 3.11, 3.12 and
3.13.
Fig. 3.11 shows a comparison between room temperature PL spectra of the 30
min and 4 hr nanorod samples. The 30 min samples reveal a very intense deep defect
39
2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4
RT
30 min
4 hr
PL
In
ten
sit
y,
a.u
.
E, eV
Figure 3.11. Comparison between room temperature PL spectra of the samples obtained after 30 min and
4 hrs of synthesis in the W/EG 1/1 mixture.
40
2.2 2.4 2.6 2.8 3.0 3.2 3.4
T = 8 K
30 min
4 hr
PL
In
ten
sit
y,
a.u
.
E, eV
Figure 3.12. Comparison between low temperature (8K) PL spectra of the samples obtained after 30 min
and 4 hrs of synthesis in the W/EG 1/1 mixture.
41
2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4
4 hr
8 K
RT
PL
In
ten
sit
y, a
.u.
E, eV
Figure 3.13. Comparison between low temperature (8K) and room temperature PL spectra of the ZnO
nanorod sample obtained after 4 hrs of synthesis in the W/EG 1/1 mixture.
emission at ~ 2.2 eV. The deep level defects are most likely oxygen deficiency-related
[36]. One can also see two relatively weak near-band gap emission broad peaks at 3.15
eV and 3.35 eV. On the other hand, for the longer growth times, the relative intensity of
the deep defect luminescence is lower by an order of magnitude. Presence of two NBE
features was observed only in the 30 min samples and requires further elucidation.
Insignificant spectral differences between the 4 hr and 24 hr growth times (not shown)
are consistent with the morphological similarity of these specimens as detected by the
SEM. The results of Fig. 3.11 can be explained within the assumption that in smaller
42
crystals the relative radiative contribution of the deep defect-rich near-surface layers is
substantially greater.
Figure 3.12 juxtaposes low temperature (8 K) PL spectra of the 30 min and 4 hr
samples. The following features appear in the spectra: sharp excitonic emissions at ~
3.31 eV and ~ 3.36 eV, a “blue-violet” recombination at ~ 2.9 eV due to relatively
shallow defects, and a “green” deep defect band at ~ 2.5 eV. In a recent paper [70], the
3.31 eV peak was attributed to a surface defect-related excitonic luminescence.
Accepting this assignment one can see that in smaller (30 min) crystallites, the relative
intensity of this emission is much stronger compared to the one in the well defined
nanorods. Thus, for smaller morphologies, the surface defect layer is likely to become
predominant. Also, the relative intensity of the deep-defect (“green”) and shallower-
defect (“blue-violet”) emission is higher by an order of magnitude in the 30 min sample,
indicating an elevated contribution from the deep and shallower defects compatible with
a greater volume fraction of the defective near-surface layer.
There also exists a significant temperature-mediated rearrangement of
recombination channels in our ZnO nanorods as can be seen from Fig. 3.13, which is
contrasting the low temperature and room temperature PL spectra of the long nanorod
sample. The defect-related recombination is dominated at room temperature by the green
emission (probably related to oxygen deficiency), whereas in the low-temperature spectra
the green emission is suppressed by the blue-violet luminescence, suggesting existence of
competing recombination centers. Such competition in bulk ZnO was described
previously [71], and the nature of the shallow centers still remains disputable.
43
We suggest that the employed growth procedure for ZnO quasi-1D nanostructures
is allowing good control of their size, which in turn affects the defect properties of the
material. Both the growth time and the solvent composition determine morphology and
quality of the samples. Shorter growth times result in smaller particles with rougher
surfaces, lesser morphological anisotropy, and conspicuous mid-rod bipod twining.
Longer growth times yield well defined nanorods with distinct hexagonal symmetry.
We observed a strong correlation between the morphology of our samples and
their defect luminescent properties, confirming a hypothesis that the radiative
contribution of defects should increase with the decreasing nanocrystal size. Smaller
crystallites exhibited intense deep-defect luminescence and enhanced emission related to
near-surface excitonic recombination. Longer growth times lead to formation of well-
defined nanorods with hexagonal symmetry exhibiting reduced defect emission.
III.3 Correlation between Defect Luminescence and Quasi-Fractal
Dimensionality in ZnO Nanostructures [72]
In the previous section, the growth protocol was designed for a size- and shape-
selective growth of ZnO nanorods. Such selectivity was crucial for addressing the
question of how the nanogeometry is affecting the surface defect properties. Crystal size
and morphology of ZnO nanorods were controlled by the growth time and the solvent
composition: both the longitudinal and transverse average dimensions of the obtained
hexagonal nanorods as well as their morphological anisotropy were increasing with the
growth time and water content in the solvent. We observed a strong correlation between
nanocrystals’ size/morphology and their optoelectronic defect-related properties.
44
In order to better elucidate the influence of the surface-to-volume ratio on the
luminescent properties in nanocrystalline systems, not only the pure size scaling effects
have to be investigated but also the effects of dimensionality since the increase in surface
to volume ratio may also happen with the increasing dimensionality of a nanostructure.
In this section we address this question for quasi-fractal-dimensional nanonetworks
containing ZnO and hydrozincite. The growth method used to produce these samples,
similar to that presented in [14], also employs accurate morphology control of the
prepared nanoscale structures. Here we tackle the questions of the correlation between
the optoelectronic properties of the obtained nanosystem and the quasi-fractal
dimensionality of the studied specimens. We observed a strong correlation between the
morphology of the samples and their optoelectronic properties. Our results indicate that a
substantial increase of the free surface in the nanocrystal samples generates higher
relative concentration of defects, consistent with the model of defect-rich surface and
subsurface layers.
The scope of the reported methods to fabricate ZnO nanostructures is truly
remarkable and grows by the day. One of these many methods is a procedure based on
converting precursor hydrozincite (Zn5(OH)6(CO3)2) nanostructures into ZnO
nanocrystals [73-79]. In particular, this technique was shown to be effective for
producing nanoporous nanoscale ZnO useful in potential solar cells [80] and gas sensing
applications [14, 78]. By itself, hydrozincite, a naturally occurring basic zinc carbonate,
has been researched for a long time, mostly in geology-related studies, as well as a
catalyst component [81, 82], and a corrosion byproduct of zinc. Nanoscale forms of
hydrozincite have been known and studied not only as precursors for generating a
45
nanoscale ZnO, but also as, for example, a product of naturally occurring biologically
controlled mineralization [83]. Although optical properties of hydrozincite have been
investigated by many researchers, the nature of its visible luminescence is still under
debate. For example, the characteristic blue luminescence is often attributed to impurities
(such as Pb), although this assignment is not certain [84]. Furthermore, optoelectronic
properties of the nanoscale hydrozincite are largely unknown.
A set of samples with variable dimensionalities was made as follows. The
specimens were prepared from Zn(NO3)2.6H2O (Aldrich) and urea (Aldrich) in Milli-Q
water, with the initial concentrations of Zn2+
ions and urea of 0.01 M and 0.05 M,
respectively. Fresh stock solutions were prepared to avoid hydrolysis upon storage. All
reagents were of an analytical reagent grade. The sample growth was carried out in 600
mL closed glass reactor bottles, placed in an oven preheated at 90°C, with the total
volume of reaction mixture of 425 mL. The temperature was measured in the center of
the reactor by a Pt100 sensor. The maximum temperature, 84°C, was reached after
around 120 minutes and was kept constant during the synthesis. Three different growth
times, 2, 4 and 48 hours, were applied. The resulting solids were filtered off, washed
with water and dried in air.
Obtained samples were characterized by a scanning electron microscope. The
SEM images of the obtained samples are shown in Figures 3.14 – 3.16. As one can see,
the material grown for the longest (48 hrs) time interval exhibits a distinct 2D
morphology. The morphology of the sample grown during the 2 hrs period is quasi 1D,
consisting of a network of nanowires, whereas the dimensionality of the 4 hrs sample is
of intermediate (“1.xD”) dimensionality. One can see that with the increase of the
46
growth time the nanowires start to expand sideways and acquire first a leaf-like and then
a platelet-like morphology. Notably, the characteristic scale of the obtained
nanostructures does not change significantly with the growth time. Thus, our growth
procedure provides consistent sampling of the hydrozincite/ZnO nanostructures to
address the effects of a quasi-fractal dimensionality and shape on the defect
luminescence, where the growth time serves as a control parameter.
Figure 3.14. SEM image of the sample obtained after 48 hrs of synthesis.
Figure 3.15. SEM image of the sample obtained after 4 hrs of synthesis.
47
Figure 3.16. SEM image of the sample obtained after 2 hrs of synthesis.
Composition of the specimens was determined using the Fourier transform
infrared (FTIR) spectroscopy. The abundance of the carbonate groups from the
precipitated Zn5(OH)6(CO3)2 in the final product was confirmed in all samples. ZnO
vibrational bands were also observed.
We ran room and low temperature PL experiments on these samples (Figures 3.17
and 3.18). The ubiquitous hydrozincite blue emission band can be observed at ~ 2.9 eV
in all the spectra with approximately the same location independent of the temperature
and morphology. On the other hand, in the spectra of the 2D and 1.xD samples one can
also see a band at ~ 2.4 eV at both room and low temperatures. For the 1D sample, the
blue ~ 2.9 eV luminescence band is accompanied by a broad spectral feature at ~ 2.2 eV
for both figures. Finally, at room temperature, in the higher-dimensional samples there is
a single peak at ~ 3.3 eV, and at low temperature – a series of relatively narrow peaks in
the high-energy part of the spectrum. We submit that these features are consistent with
common ZnO luminescence spectra, where the 2.2 eV and 2.4 eV features are the deep
defect-related bands and the UV peaks are associated with the near-band edge (NBE)
48
transitions, such as excitonic luminescence and its phonon replicas. Moreover, one can
see a significant correlation between the sample dimensionality, on the one hand, and the
relative intensity of the defect vs. NBE emission ratio, on the other. For the room
temperature PL, the relative intensity of the ZnO deep defect emissions and the blue
hydrozincite luminescence grows with the decrease of the dimensionality, and the band
gap luminescence disappears for the 1D sample. Similarly, for the 8K PL spectra, the
spectral features in the visible have a visibly higher intensity for the 1D structures
compared to those of the 2D sample, while the NBE emissions observed in the 2D system
are reduced by an order of magnitude in the 1D material.
2.0 2.2 2.4 2.6 2.8 3.0 3.2
2D
1.XD
1D
PL
In
ten
sit
y, a.u
.
E, eV
Figure 3.17. Comparison between room temperature PL spectra of the samples of variable dimensionality.
49
2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4
2D
1D
PL
In
ten
sit
y, a.u
.
E, eV
Figure 3.18. Comparison between low temperature (8 K) PL spectra of the samples of variable
dimensionality
The observed spectral behavior can be explained in terms of a likely coexistence
of two phases, those of zinc oxide and hydrozincite, in the studied samples. The
significant correlation between the quasi-fractal dimensionality of the samples and the
ZnO- and Zn5(OH)6(CO3)2-related emissions could be indicative of an elevated defect
contribution in crystals with smaller dimensionalities and, therefore, in a good agreement
with the assumption that the surface and near-surface defect contribution increases with
the decrease of the quasi-fractal dimensionality and hence the surface-to-volume ratio (a
greater volume fraction of the defect-rich near-surface layers). The two-phase structure
of the specimens is further confirmed by the dependence of the spectral shape on the time
50
of exposure to the laser beam. Figure 3.19 illustrates a typical example of such
dependence, where the two spectra of the 2D sample shown were collected at 8 K within
a 1 hour interval of a continuous laser beam irradiation. One can see a relative constancy
of the ZnO-related emission, whereas the intensity of the blue hydrozincite emission band
is reduced by almost 20%. It is well known that annealing of nanoscale hydrozincite
leads to its transformation into ZnO [73-79]. Local heating produced by the laser beam
may contribute to partial reduction of the relative abundance of the hydrozincite phase.
2.2 2.4 2.6 2.8 3.0 3.2 3.4
PL
In
ten
sit
y, a.u
.
E, eV
1st scan
2nd scan
Figure 3.19. Comparison between low temperature (8 K) PL spectra of the 2D sample collected within a 1
hour interval during a continuous laser beam irradiation.
Time dependence shown in Figure 3.19 was observed consistently for all other
samples (not shown). An additional corroboration for the presence of the ZnO phase in
51
the material is a shift in the position of the deep defect band from 2.4 eV to 2.2 eV
occurring at both temperatures with the decreasing dimensionality of the nanostructures.
A very similar shift (2.4 eV to 2.2 eV) of the deep-defect emission was observed in PL
spectra of ZnO nanopowders with the decrease of an average grain size [51, 62]. It should
be noted that an unambiguous assignment of the optical transitions to specific defects in
ZnO is still under debate. A similar statement can be applied to the blue hydrozincite
luminescence. For example, it is unlikely that the lead impurity (cf. Ref. 84) is the
primary source of this emission band in the studied samples since our growth procedure
did not employ any Pb content.
It should be noted that, in principle, different morphologies of the samples may
affect scattering and/or absorption in PL experiment. However, this factor is not likely to
be the main source of the spectral differences observed in our work. ZnO-related PL
features similar to those reported here were observed by us in nanoscale ZnO specimens
with completely different morphologies [63]. Moreover, time dependent spectra (cf. Fig.
3.19) indicate that the observed spectral transformations result primarily from variance in
composition and/or relative abundance of radiative centers.
Described in this section PL experiments on the quasi-fractal-dimensional
hydrozincite/zinc oxide nanonetworks revealed a strong correlation between the
dimensionality/morphology of the nanocrystals and their defect optoelectronic properties,
confirming hypothesis that the contribution of defects should increase with the increasing
surface-to-volume ratio of the nanocrystal. Careful control of the quasi-fractal-
dimensionality of such nanonetworks and their surface properties can be vital for efficient
operation in underlying applications. To the best of our knowledge there were no
52
published reports on the luminescent properties of nanoscale quasi-fractal-dimensional
hydrozincite.
III.4 Growth-Morphology-Luminescence Correlation in ZnO-Containing
Nanostructures Synthesized in Different Media [85]
In preceding sections it was shown that visible luminescence in nanoscale ZnO
can be controlled by size scaling and quasi-fractal dimensionality. One can state that, in
general, the surface-to-volume ratio of a nanostructure is affected by its morphology. In
this section we demonstrate, on the one hand, how further modifications of the wet
precipitation growth protocol described above allow production of even more diverse sets
of nanomorphologies and, on the other hand, how well the PL spectra collected on these
materials correlate with their geometries.
Zinc hydroxide particles were prepared by a two-step process employing zinc
nitrate hexahydrate, urea, ethylene glycol, water, and p-toluene-sulfonic acid
monohydrate (p-TSA). Different concentrations of the reactants as well as different
volume ratios of the solvents were used. Aqueous solutions were prepared from
Zn(NO3)2·6H2O (Merck), urea (Kemika), ethylene glycol (Sigma-Aldrich) and p-TSA
(Fluka). Chosen concentrations of the Zn2+
ions and urea used in the experiments were
0.01 M, 0.1 M, 1.0 M, and 0.05 M, 5.0 M, respectively. The reaction routine was based
on three different volume ratios of water to ethylene glycol (W/EG) – 1:19, 1:3 and 1:1 as
well as varied additions of p-TSA of 0.5M, 0.25M, 0.025 M, 0.0025 M and 0.00025 M.
Sample growth was performed at different concentrations of reactants and reaction times.
It was carried out in a 250 mL flask under constant mixing at 90°C. First, 200 mL of
53
solvents were heated to 90°C and then Zn2+
-salt, urea and p-TSA were added. Separate
batch of samples was grown in a larger 5L reactor under constant mixing at 90°C. The as-
prepared solids were filtered off, washed with water and dried in air. ZnO particles were
obtained by thermal treatment of the reaction products at two different temperatures,
350°C and 500°C.
A short synopsis of the synthesis conditions, as well as the impact of p-TSA on
the particle morphology is outlined in Table 3.2.
Sam
ple
c Z
n-n
itra
te (
M)
c u
rea
(M)
c (p
-TS
A)
(M)
W v
olu
me
(mL
)
EG
vo
lum
e (m
L)
pH
Rea
ctio
n t
ime
DJ12
(A) 1 5 0.25 10 190 5.46 3 h
DJ13
(B) 1 5 0.5 10 190 5.47 3 h
DJ14
(C) 0.1 5 0.025 50 150 7.82 3 h
DJ2 0.01 0.05 - 1500 1500 - 3h
DJ7 0.01 0.05 0.0025 100 100 - 16h
DJ8 0.01 0.05 0.00025 100 100 - 22h
Table 3.2 Brief summary of the growth parameters for samples prepared at 90°C.
Samples DJ2, DJ7 and DJ8 are shown in Fig. 3.20. Particles presented in Fig.
3.21 are spherically shaped, with the average size of spheres changing slightly with
altering synthesis conditions. For samples DJ12, DJ13 and DJ14 these average sphere
sizes were estimated as follows: 5μm (DJ12, Fig. 3.21(a)), 10μm (DJ13, Fig. 3.21(b)),
and 6μm (DJ14, Fig. 3.21(c)). When comparing surfaces of the particles one can notice
that they are different in each case. For the DJ12 sample, the surface of the particles is
composed of tiny lamellas, which are noticeably separated. Lamellas become larger and
54
less distinct in the DJ13 sample, while for the DJ14 sample they start merging and cannot
be clearly distinguished any more. Thereby we suggest that changes in the synthesis
parameters during the reaction sequence (such as the water/ethylene glycol volume ratio,
the addition of p-TSA, changes in concentration of zinc nitrate, and/or variations of the
amount of urea added) affects both particles’ shape and their size distribution. For
relatively low p-TSA concentrations, the spherical particles are flower-shaped and
composed of lamellas. Higher amounts of p-TSA still yield spherical particles, their size
increases while the inter-lamellar volumes become slowly filled with more solid phase.
Figure 3.20. SEM images of samples DJ2 (A), DJ7 (B), and DJ8 (C).
Figure 3.21(a). SEM images of sample DJ12 at lower (A) and higher (B) magnification.
55
Figure 3.21(b). SEM images of sample DJ13 at lower (A) and higher (B) magnification.
Figure 3.21(c). SEM images of sample DJ14 at lower (A) and higher (B) magnification.
Samples synthesized under the conditions specified in Table 3.2 were thermally
treated at two different temperatures (350°C and 500°C) for 2 hours as shown in Table
3.3.
In this way we tried to determine the influence of the thermal load on the size and
shape of the obtained solid product. First, all three samples, DJ12, DJ13, and DJ14, were
thermally treated at 350°C and then characterized by SEM microscopy and PL
spectroscopy. Consequently those samples were thermally treated at 500°C and then
characterized using the same experimental probes.
56
Sam
ple
c Z
n-n
itra
te (
M)
c ure
a (M
)
c (p
-TS
A)
(M)
W v
olu
me
(mL
)
EG
volu
me
(mL
)
pH
Hea
t tr
eatm
ent
Tem
per
ature
(ºC
)
DJ16
(D) 1 5 0.25 10 190 5.46 2 h 350
DJ17
(E) 1 5 0.5 10 190 5.47 2 h 350
DJ18
(F) 0.1 5 0.025 50 150 7.82 2 h 350
DJ20
(G) 1 5 0.25 10 190 5.46 2 h 500
DJ21
(H) 1 5 0.5 10 190 5.47 2 h 500
DJ22
(I) 0.1 5 0.025 50 150 7.82 2 h 500
Table 3.3. Brief summary of the growth parameters for annealed samples
SEM images revealed that temperature has considerable influence on the
morphology of the product as follows (Fig. 3.22). Individual particles of the DJ16
sample (Fig. 3.22(A)) became smaller than 5 μm after thermal treatment. Those spherical
particles are in fact agglomerates of tiny spheres ~ 50 nm in size. Somewhat similar
morphology can be observed for sample DJ17 (sample DJ13 annealed at 350°C, Fig.
3.22(B)). Star-shaped ~ 5 μm particles (sample DJ18, Fig. 3.22(C)) are agglomerates of
needle-shaped tiny hexagonal rods (which were shown to be made of nano-fibers) [86].
Morphologies shown for samples DJ20 (Fig. 3.22(D)) and DJ21 (Fig 3.22(E)) are similar
to those shown in Figs. 3.22(A) and 3.22(B), respectively – small spherical particles (20-
30 nm) form larger agglomerates. An interesting effect may be noticed for sample DJ21
presented in Fig. 3.22(E) where small (~ 20-30 nm) spherical particles are clumped into
larger (~ 200 nm) ones, which in turn are further agglomerated into even larger (~ 5μm)
clusters. Similarly, in Fig. 3.22(F) (sample DJ22) needle-shaped tiny hexagonal rods (or
57
nano-flakes) form star-like structures (~ 2-3 μm), which are further agglomerated into
larger (~ 4 μm) clusters. Such grouping of the morphologies as detected by SEM is in
excellent correlation with the growth parameters.
Figure 3.22. SEM images of samples obtained after thermal treatments at two different temperatures:
350°C (A-C) and 500°C (D-F) for 2 hours.
Remarkably, the observed differences between the described groups of
nanostructures are retained in the corresponding PL spectra taken at room temperature
and 8 K. Figs. 3.23-3.27 clearly demonstrate that the samples can be grouped according
to the growth/morphology/luminescence correlation as follows.
Samples DJ2, DJ7, and DJ8 (Fig. 3.23). In many respects defect-related
luminescence features in these hexagonal rod-like specimens are similar to those
observed in our earlier work [63] for ZnO nanorods, although there are clear distinctions
58
in both the morphology and the emission spectra. At room temperature all three samples
reveal emission bands at ~ 2.2 eV and ~ 3.1 eV. The DJ2 sample also exhibits a relatively
weak excitonic emission at ~ 3.3 eV. At 8 K the ~ 2.2 eV and ~ 3.1 eV luminescent
bands are preserved in approximately the same proportion, however, the DJ7 sample
shows a strong excitonic emission at ~ 3.3 eV, whereas the DJ8 shows a weak excitonic
emission. At the same time, surprisingly, the excitonic emission for the DJ2 sample
disappears.
2.0 2.4 2.8 3.2
T = 293K
DJ2
DJ7
DJ8
E, eV
2.0 2.4 2.8 3.2
T = 8K
DJ2
DJ7
DJ8
PL
In
ten
sit
y, a.u
.
E, eV
Figure 3.23. PL spectra of samples DJ2, DJ7, and DJ8.
For samples DJ12 and DJ13 (Fig. 3.24) the defect-related luminescence features
are similar to those observed in our work on the fractal hydrozincite/ZnO nanonetworks,
59
although there are certain differences in both the morphology and the luminescence. At
RT both samples reveal emission bands at ~ 2.4 eV and ~ 2.9 eV. The former is attributed
to the ZnO phase and the latter – to the hydrozincite phase. At LT the luminescent bands
are preserved in approximately the same proportion, yet a new narrow emission appears
at ~ 2.65 eV. The origin of this emission in not known and to the best of our knowledge
was not observed in hydrozincite before. The ZnO band gap emission does not appear in
these spectra neither at room temperature nor at 8 K.
2.0 2.4 2.8 3.2
T = 293K
DJ12
DJ13
E, eV
2.0 2.4 2.8 3.2
PL
In
ten
sit
y, a.u
.
T = 8K
DJ12
DJ13
E, eV
Figure 3.24. PL spectra of samples DJ12 and DJ13.
Samples DJ18 and DJ22 (Fig. 3.25) at room temperature both reveal deep defect
emission bands at ~ 2.0 eV and ~ 3.1 eV near band gap peaks in about the same
60
proportion. At low temperature these luminescent features are preserved in approximately
the same proportion, yet a new broad shoulder component appears at ~ 2.4 eV. Also, at 8
K in both samples one observes a distinct ZnO excitonic luminescence at ~ 3.3 eV.
2.0 2.4 2.8 3.2
T = 8K
DJ18
DJ22
PL
In
ten
sit
y, a.u
.
E, eV
2.0 2.4 2.8 3.2
T = 293K
DJ18
DJ22
E, eV
Figure 3.25. PL spectra of samples DJ18 and DJ22.
Samples DJ16 and DJ20 (Fig. 3.26). At room temperature both samples reveal
ZnO excitonic luminescence at ~ 3.3 eV and a near band gap emission at ~ 3.1 eV in
approximately the same proportion. However, the deeper defect luminescence emission is
somewhat different in these samples in both energy and peak intensity. Whereas the DJ20
sample the emission band is centered at ~ 2.2 eV and dominates the entire spectrum, in
the DJ16 sample the deeper emission band is centered at ~ 2.4 eV and its intensity is
61
several times weaker, and is below that of the near-band gap spectral features. At low
temperature the defect luminescent bands are centered at ~ 2.4 eV (shifted) and ~ 3.1 eV
in approximately the same intensity proportion, yet the excitonic luminescence at ~ 3.3
eV is observed only in the DJ16 sample.
2.0 2.4 2.8 3.2
T = 8K
DJ16
DJ20
PL
In
ten
sit
y, a.u
.
E, eV
2.0 2.4 2.8 3.2
T = 293K
DJ16
DJ20
E, eV
Figure 3.26. PL spectra of samples DJ16 and DJ20.
Spectra of samples DJ17 and DJ21 (Fig. 3.27) are somewhat stand-alone. At both
room and low temperature they reveal defect luminescent bands at ~ 2.4 eV and ~ 3.1 eV.
However, at 8 K a narrow emission appears at ~ 2.65 eV (very similar to that observed in
samples DJ12 and DJ13). The excitonic luminescence at ~ 3.3 eV is observed at both
room and low temperature, and its intensity increases significantly at LT.
62
2.0 2.4 2.8 3.2
T = 8K
DJ17
DJ21
PL
In
ten
sit
y,
a.u
.
E, eV
2.0 2.4 2.8 3.2
T = 293K
DJ17
DJ21
E, eV
Figure 3.27. PL spectra of samples DJ17 and DJ21.
Overall, one can see that there is only a relatively short list of luminescent bands
and peaks observed in the studied samples (~ 2.0-2.2 eV, ~ 2.4-2.5 eV, ~ 2.65 eV, ~ 2.9
eV, ~ 3.0-3.1 eV, ~ 3.3 eV), however, their presence in the spectra of a specific group, as
well as their relative intensity and temperature dependence is strongly correlated with the
microscopic morphology of the corresponding group.
It is most likely that the origin of the luminescence in samples DJ12 and DJ13 is
different from the rest of the samples and is produced by the defects in the matrix of zinc
hydroxide carbonate rather than ZnO. The rest of the emission features can be explained
63
in terms of radiative recombinations in zinc oxide crystals. For example, some authors
[87] attribute the yellow-orange luminescence (~ 2.0-2.2 eV) to Zn vacancies.
2.0-2.2 eV 2.4-2.5 eV 2.7 eV 2.9 eV 3.05 eV 3.2-3.3 eV
DJ2 RT/LT RT/LT – – RT/LT –
DJ7,DJ8 RT/LT RT/LT – – RT/LT LT
DJ12, DJ13 – RT/LT LT RT/LT – –
DJ18, DJ22 RT/LT LT – – LT LT
DJ16, DJ20 RT/LT LT – LT RT/LT RT/LT
DJ17 – RT/LT LT RT/LT – RT/LT
Table 3.4. Summary of the observed PL features. RT – observed in room temperature spectra; LT –
observed in low temperature (8 K) spectra.
The 2.0-2.2 eV PL band is observed at both room and low temperatures in all the
ZnO samples, except DJ17. However, there exists an alternative attribution of this band
[88]. Within this model, the ~ 2.0-2.2 eV and ~ 2.4-2.5 eV emissions are related to the
different charge states of the same native defect, oxygen vacancy VO. The ~ 2.0-2.2 eV
luminescence thus originates from the VO++
centers, while the ~ 2.4-2.5 eV green PL
signal – from the singly charged vacancies VO+. In our samples the green band is
observed at low temperatures for all the ZnO samples, however it disappears at room
temperature in samples DJ16, DJ18, DJ20, DJ22. Possible explanation for such behavior
may be similar to that proposed in Ref. 88. These samples have the greatest surface-to-
volume ratio and thus possibly a greater positive surface charge, making it more likely to
find the VO++
centers there. The ~ 2.9 eV band appears primarily at low temperatures
only in a few ZnO samples (DJ16, DJ17, and DJ20) and is most likely associated with the
interstitial zinc [89]. The violet spectral feature at ~ 3.0-3.1 eV appears in most samples
and can be attributed to either hydrogen-related or extended structural defects [19].
Finally, the 3.2-3.3 eV peak is related to excitonic transitions in ZnO. It is absent at room
temperature in samples DJ7, DJ8, DJ18, and DJ22, and even at low temperature in
64
sample DJ2. Such behavior may be associated with the existence of surface traps
quenching excitonic recombination [24]. As we mentioned before, the nature of the
narrow peak ~ 2.65 eV in samples DJ12, DJ13, DJ17, and DJ21 is unknown and requires
further studies.
In summary, for the studied nanostructured samples grown by precipitation from a
water/ethylene glycol solution of a Zn salt in the presence of urea with a subsequent
thermal treatment we observed strong dependence of optoelectronic properties on the
morphology of the grown samples. Morphology on the other hand can be carefully
tailored via the growth control parameters. The following groups of samples clustered
according to their nanomorphology – DJ2/DJ7/DJ8, DJ12/DJ13, DJ18/DJ22, DJ16/DJ20,
DJ17/D21 – are also distinct vis-à-vis their luminescent properties. PL peaks at ~ 2.0-2.2
eV, ~ 2.4-2.5 eV, ~ 2.65 eV, ~ 2.9 eV, ~ 3.0-3.1 eV, ~ 3.3 eV are observed in most
specimens, although their relative intensity and temperature evolution are specific to an
individual group, confirming our assertion about the strong correlation between
morphology and optoelectronic properties. Further studies will be necessary to further
elucidate the nature of the observed luminescent transitions and their explicit association
with the nanoscale geometry and the stoichiometry of the grown species.
III.5 Studies of BEx Luminescence in ZnO Nanopowders [51, 62]
In the previous four sections we investigated rather broad PL spectral ranges of
nanocrystalline ZnO (visible to near UV). We examined shapes and intensities of various
defect-related transitions, observed as rather broad bands, in their relation with the NBE
luminescence as well as with each other. Our observations were instrumental in
determining the decisive influence of morphology on optoelectronic properties. On the
65
other hand, as it was mentioned in the Introduction (Chapter I), an important class of
luminescent transitions occur at low temperatures in the NBE spectral region, namely the
excitonic transitions. These spectrally-narrow recombination channels can and do yield a
substantial amount of information that cannot be obtained from the analysis of the broad
visible peaks.
In this section we focus primarily on this class of radiative transitions in
nanoscale ZnO. A relatively narrow part of the LT PL spectrum corresponding to the
BEx luminescence (in ZnO it is ~ 3.35-3.37 eV) is especially rich in information,
therefore we will concentrate more attention on this energy range. As a reference, we ran
LT PL measurements on a high-quality single crystalline from Cermet Inc. The results
for the BEx range are shown in Fig. 1.5, in which one can see numerous narrow peaks
(some of them just a ~ 100 μeV wide) corresponding to excitons bound to different
defects. The most up-do-date treatise of these peaks is given in a recent detailed review
by Meyer [48].
Our studies of the BEx PL were performed primarily with the commercial ZnO
nanopowders discussed in section III.1. These specimens provide the most adequate
nanoscale ZnO objects for BEx studies because of their rich and intense PL BEx spectra,
as well as their availability in large quantities.
Fig. 3.28 juxtaposes BEx features of the nanopowders studied here. There is a
common 150 meV-wide broad feature with a few narrower excitonic peaks (the most
prominent at ~ 3.357 eV and ~ 3.360 eV) superimposed with it. The 3.363 - 3.369 eV
relatively wide feature, common to all nanopowders, can be attributed to excitons bound
to either surface defects or isoelectronic traps, both giving rise to broader shallower
66
3.354 3.356 3.358 3.360 3.362 3.364
AE25
AE31
ZB
ZA
PL
In
ten
sit
y,
a.u
.
E, eV
Figure 3.28. Comparison between BEx spectral regions for ZnO nanopowders
potential wells. The low-intensity free exciton contribution is observed in all the spectra
(not shown in Fig. 3.28). The ~ 3.357 eV and ~ 3.360 eV BEx luminescence features
were first discussed in Reynolds, et al. [90], and assigned nomenclatures I9 and I6
correspondingly. Relative intensity of the BEx emissions in the studied ZnO powders is
variable and depends on the individual material with little correlation with the deeper-
defect emission intensities or grain size distribution. E.g., the AE25 and AE31 specimens
show different relative intensities of the I9 and I6 peaks. The same statement applies to
the ZA and ZB samples. The origin of these emissions is still under debate. In recent LT
PL studies of BEx in ZnO nanostructures [39], similar emissions, although at slightly
67
different locations, were observed and given rather vague assignments. The question
remains open whether these features originate from the excitonic recombinations on
surface defects or bulk defects. Interestingly, despite the difference in the relative
intensities of the BEx peaks in different nanopowder samples, the background band
centered at ~ 3.363-3.364 eV is common in all the spectra.
It should be noted that the BEx spectra shown above (as well as most other BEx
spectra here) were taken at an incident laser intensities of 2.55 W/cm2. Justification of
this choice becomes clear from Fig. 3.29, where we plot LT PL spectra of the ZA sample.
Low incident intensities result in a low signal-to-noise ratio, whereas relatively high
intensities produce significant broadening of the spectral features and their red shifts,
probably due to local heating leading to effectively higher experimental temperatures.
Thus the chosen optimum experimental condition is the most adequate for a maximum
signal-to-noise ratio and a minimal local heating.
68
3.33 3.34 3.35 3.36 3.37 3.38 3.39 3.40
25.5 W/cm2
12.7 W/cm2
6.37 W/cm2
2.55 W/cm2
1.27 W/cm2
0.255 W/cm2
PL
In
ten
sit
y,
a.u
.
E, eV
Figure 3.29. Incident power dependence of the LT PL spectra for the ZA sample
To extract information about local energetics we performed temperature-
dependent PL measurements in the BEx region for a number of nanopowder samples.
Figures 3.30(a), 3.30(b), 3.30(c) and 3.30(d) illustrate such T-dependent spectra for the
ZA, ZB, AE31 and AE25 specimens, respectively. There are different regions in the
spectra: free excitons (3.369 - 3.378 eV), a broader emission band (3.363 - 3.369 eV),
probably due to surface-bound excitons [91] and narrower BEx lines in the 3.355 - 3.363
eV region. All the spectral elements exhibit relative intensity decrease, broadening, and
red shift with increasing temperatures.
69
3.356 3.360 3.364 3.368 3.372 3.376
10K
15K
20K
25K
30K
35K
40K
45K
50K
55K
60K
PL
In
ten
sit
y, a.u
.
E, eV
Figure 3.30(a). Temperature-dependent PL spectra of the ZA sample.
70
3.355 3.360 3.365 3.370 3.375
10K
15K
20K
25K
30K
35K
40K
45K
50K
55K
60K
PL
In
ten
sit
y, a.u
.
E, eV
Figure 3.30(b). Temperature-dependent PL spectra of the ZB sample
71
3.350 3.355 3.360 3.365 3.370 3.375
10K
15K
20K
25K
30K
35K
40K
45K
50K
55K
60K
PL
In
ten
sit
y,
a.u
.
E, eV
Figure 3.30(c). Temperature-dependent PL spectra of the AE31 sample
72
3.360 3.365 3.370 3.375
10K
15K
20K
25K
30K
35K
40K
45K
50K
55K
60K
PL
In
ten
sit
y,
a.u
.
E, eV
Figure 3.30(d). Temperature-dependent PL spectra of the AE25 sample
For the spectra shown in Figs. 3.30 (a) – (d) we applied a multi-parameter fitting
procedure approximating the peaks to be Gaussians for simplicity. The output of the
fitting is a T-dependence of the manifold of peak widths, intensities and positions. We
performed standard treatment of these data assuming the following T-dependence of the
BEx peak intensities I and positions E discussed in the Introduction section (Chapter I):
Intensity,
Position,
kTE
o
aAe
ITI
/1
)(
T
TETE o
2
)(
73
Here I0, A, E0, α, β are the fitting parameters, k – Boltzmann constant, and Ea –
activation energy. Normally parameter β is directly related to the Debye temperature.
Figs. 3.31 and 3.32 illustrate fitting of the experimental results with the above functional
dependencies of the relative intensities (Fig. 3.31) and peak positions (Fig. 3.32) for the
AE31 sample. Table 3.5 shows fitting results for the samples studied. The obtained
values of the activation energies and Debye temperatures fall within very reasonable
margins for the BEx emissions reported before [39].
0.03 0.06 0.09
Ea = 5.1 meV
Pea
k I
nte
nsit
y,
a.u
.
1/T, K-1
Experiment
Arrhenius Model
Figure 3.31. Peak intensity temperature dependence for the AE31 sample
74
Figure 3.32. Peak position temperature dependence for the AE31 sample
Peak 1 Peak 2
Sample α (eV/K) β (K) E0 (eV) Ea (meV) α (eV/K) β (K) E0 (eV) Ea (meV)
AE31 0.00069 911 3.360 ≥11 0.00065 913 3.365 ≥8
ZA 0.00062 919 3.356 ≥13 0.00071 917 3.360 ≥5
ZB 0.00063 915 3.357 ≥6 N/A
Table 3.5. Fitting results for the T-dependence of BEx peak parameters observed in AE31, ZA and ZB
samples
For the ZA sample, fitting of the peak intensities vs. temperature (Fig. 3.33) for
the 3.361 eV peak indicates activation energy ≥ 16 meV. However, similar fitting for the
3.357 eV BEx line in the AE31 sample shows activation energy of ~ 5 meV (Fig. 3.31).
The energy difference between the BEx and FEx emissions for the ZA sample is ~ 16
10 20 30 40 50 60
3.3575
3.3580
3.3585
3.3590
3.3595
3.3600
3.3605
Experimental data
Varshni fitting
Ph
oto
n E
ne
rgy, e
V
T, K
75
meV, whereas for the AE31 sample the energy difference between the BEx peak and the
center of the broader feature is ~ 5 meV. Looking at these energy separations of the
spectral features one can surmise that the calculated activation energy values reveal
probably dissimilar channels of the BEx state deactivation for different samples – through
a free exciton (ZA sample) and through an exciton bound to either a surface state or an
isoelectronic trap (AE31 sample). We argue that the former assignment should be more
adequate since the ~ 5 meV activation energy is observed in powders with smaller
average grain size where the surface defects are more abundant and thus more likely to
trap an exciton prior to its recombination.
0.02 0.04 0.06
Ea > 16 meV
PL
In
ten
sit
y,
a.u
.
1/T, K-1
Experiment
Arrhenius Model
Figure 3.33. Peak intensity temperature dependence for the ZA sample
76
15 20 25 30 35 40 453.3548
3.3550
3.3552
3.3554
3.3556
3.3558
3.3560
3.3562
Ph
oto
n E
ne
rgy,
eV
T, K
Experimental data
Varshni fitting
Figure 3.34. Peak position temperature dependence for the ZA sample
There are differences observed also for the FWHM vs. T dependence. Thus, for
the 3.357 eV peak of the AE31 sample we see (Fig. 3.35) a nonlinear behavior, much
more pronounced than that of the 3.360 eV peak for the ZA sample (Fig. 3.36). The
nonlinearity of the FWHM vs. T behavior may indicate additional contributions of the
surface phonons in exciton scattering (a linear increase is expected for bulk samples)
[48].
These results for the BEx emissions in nanoscale ZnO further substantiate the
proposition about the crucial impact of the surface-related defect states on the
luminescent properties.
77
10 15 20 25 30 35 40 45 50 55
0.0012
0.0014
0.0016
0.0018
0.0020
0.0022
0.0024
0.0026
FW
HM
, e
V
T, K
Figure 3.35. FWHM temperature dependence for the AE31 sample
78
10 20 30 40 50 60
0.0010
0.0012
0.0014
0.0016
0.0018
0.0020
0.0022
0.0024
0.0026
FW
HM
(e
V)
T, K
Experimental data
Exponential fit
Figure 3.36. FWHM temperature dependence for the ZA sample
III.6 Effects of Plasma Processing on PL Spectra of ZnO Nanopowders [62, 51]
As we mentioned in the Introduction, the conditions of the surface and the near-
surface layers of the ZnO nanocrystals may be performance-defining factors. Thereby it
is important to have the ability to controllably manipulate these conditions.
We suggest remote plasma as one of the tools for controlling and identification of
the surface defect properties. Plasma may activate such processes as ionization,
neutralization, de-excitation, and dissociation within the plasma penetration range and
then beyond due to diffusion and local stresses.
79
Results for bulk ZnO crystals [71, 92-95] demonstrate that remote plasma species
rearrange morphology and composition not only of a surface but also possibly tens of
nanometers below. It is quite natural to presume that the same statement applies to a free
surface of a nanosize ZnO object. In this case insertion of remote plasma species occurs
over a characteristic depth comparable to the typical dimensions of a nanocrystal.
Observation of the plasma-driven changes in the optoelectronic, compositional, and
structural properties could be significantly amplified because several scales – the size of a
nanostructure, the thickness of a defect layer, the plasma penetration depth (as well as the
scale of the space-charge region and the carrier diffusion length) – become comparable.
Thus we expect that remote plasma processing of ZnO nanostructures may potentially be
one of the most sensitive tools of elucidating and tailoring the surface/subsurface
properties in the objects of interest.
A number of ZnO nanopowder samples studied here were subjected to remote
plasma treatment to expose the effects of such processing on the optoelectronic
properties. We employed nitrogen, hydrogen and oxygen/helium plasmas as described in
Table 3.6.
Parameter Nitrogen Hydrogen Oxygen/Helium
Pressure (mTorr) 130 200 150 – 200
RF power (Watts) 38 40 42
Flow rate (sccm) 25 8.8 12/17
Table 3.6. Employed remote plasma parameters
We have to point out that the trial set of plasma parameters described above is the
only set that has been employed so far, which thus may not be the most adequate one.
We have not yet performed systematic studies regarding the optimization of plasma
processing routines.
80
Following plasma treatments we ran RT and LT PL spectra of the processed
samples. Below we provide discussion of some plasma-generated spectral modifications.
The effects of plasma on the defect luminescence were more apparent in the low
temperature (8 K) spectra. We observed that the remote N plasma leads to modification
of the BEx region of the LT PL spectra of our samples. Fig. 3.37 shows these
transformations for the ZA, ZB, and AE31 nanopowders. One can see either appearance
of a new BEx peaks or an increase in the relative intensity of the existing ones. Thus, in
the ZA specimen N plasma treatment produces a defect state giving rise to the BEx
recombination at 3.360 eV, which has been previously assigned to Al or Ga donors [48,
96]. In our samples, however, N plasma can either add nitrogen-related defects or
produce native defects (e.g., theory predicts that substitutional N should produce a
relative shallow acceptor with ionization energy of 165 meV [48]). Thus, we argue that
the defect state responsible for the 3.360 eV emission may not be associated with Al or
Ga.
O-plasma yields no change in the BEx features in the 3.35-3.37 eV range,
however, in the ZA sample (with a generally weak deep defect luminescence) we
observed an increased contribution to the ~ 3.333 eV feature (see Fig. 3.38). The nature
of this emission is not well understood and requires further investigation. In most samples
oxygen treatment produced only slight variations in the deep defect luminescent bands (at
both RT and LT). This is rather a surprising result, in view of the hypothesis that deep
level emission may be related to the oxygen deficiency [54], which was shown to be
altered by a similar O-plasma treatment in high quality bulk ZnO crystal.
81
3.356 3.358 3.360 3.362
ZA N Plasma Treated
ZA As Received
Subtraction Result
PL
In
ten
sit
y,
a.u
.
E, eV
Figure 3.37 (a). Nitrogen plasma effects on the ZA sample
82
3.355 3.356 3.357 3.358 3.359 3.360 3.361 3.362
PL
In
ten
sit
y,
a.u
.
E, eV
ZB N Plasma Treated
ZB As Received
Figure 3.37 (b). Nitrogen plasma effects on the ZB sample
83
3.358 3.360 3.362 3.364 3.366
E, eV
PL
In
ten
sit
y, a
.u.
AE31 N Plasma Treated
AE31 As Received
Figure 3.37 (c). Nitrogen plasma effects on the AE31 sample
84
3.326 3.328 3.330 3.332 3.334 3.336 3.338
ZA As received
ZA O plasma treatedPL
In
ten
sit
y,
a.u
.
E, eV
Figure 3.38. Oxygen plasma effects on the ZA sample
A different group of effects was observed following hydrogen plasma treatment.
In the BEx domain of the ZA and ZB samples the 3.356 eV peak was eliminated
(previously assigned to In or Al donors [48, 97]) as well as the 3.360 eV peak in the ZA
sample (Figs. 3.39, 3.40). However, in the AE31 powder hydrogen treatment probably
generates an additional excitonic emission at ~ 3.361 eV (Fig. 3.41). Moreover, RT PL
in the AE31 and AE25 nanopowders shows a significant increase of the NBE emission
(Fig. 3.42). One should bear in mind that the as-received Zochem and American
Elements specimens exhibit a drastic difference in the deep level emission. Hydrogen ion
(a proton) is an extremely active reagent and was shown to be a non-amphoteric donor in
85
ZnO [18]. Hydrogen is capable of forming a wide variety of defects and defect
complexes in ZnO with very different properties. Since the concentrations of (probably
native) defects for the as-received AE and Z samples are very different, it may lead to a
predominance of dissimilar H-based defects and complexes in the Z and AE materials,
which is reflected in the discrepancy of their reaction to the H plasma treatment. For
example, increased concentration of H-related shallow donors in the AE samples may
explain an elevated intensity of the NBE luminescence.
3.356 3.358 3.360 3.362 3.364
PL
In
ten
sit
y,
a.u
.
E, eV
ZA H Plasma Treated
ZA As Received
Figure 3.39. Hydrogen plasma effects on the ZA sample
86
3.356 3.358 3.360 3.362 3.364 3.366 3.368 3.370
PL
In
ten
sit
y,
a.u
.
E, eV
ZB H Plasma Treated
ZB As Received
Figure 3.40. Hydrogen plasma effects on the ZB sample
87
3.354 3.360 3.366
AE31 Hydrogen Plasma Treated
AE31 As Received
PL
In
ten
sit
y, a.u
.
E, eV
Figure 3.41. Hydrogen plasma effects on the AE31 sample
88
2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4
E, eV
PL
In
ten
sit
y, a.u
. AE31 H Plasma Treated
AE31 As Received
Figure 3.42(a). RT PL results for hydrogen plasma effects on the AE31 sample
89
2.0 2.5 3.0
AE25 H Plasma Treated
AE25 As Received
PL
In
ten
sit
y,
a.u
.
E, eV
Figure 3.42(b). RT PL results for hydrogen plasma effects on the AE25 sample
The most significant changes occurred for the AE25 powder, the finest of all. The
N plasma-processed sample shows (Fig. 3.43) a threefold intensity increase of the ~ 3.1
eV defect band, emergence of a new broad feature around ~ 2 eV, as well as a reduction
of the NBE peaks.
As was mentioned above, the AE25 powder has a significantly smaller average
grain size, which sets this material apart from the rest of the samples discussed. This
circumstance has an explicit reflection in the major spectral transformation following H,
N, and O plasma treatments. Figs. 3.43 demonstrate these changes in the LT PL spectra.
In Fig. 3.43 one can see formation of an additional deep defect band centered at ~ 2 eV
90
2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4
AE25, T = 8 K
As-received
N plasma
PL
In
ten
sit
y, a.u
.
E, eV
Figure 3.43. Effects of nitrogen plasma on the AE25 sample
(or maybe even below that) as well as growth in intensity of a relatively shallow band at
~ 3.1 eV, especially prominent after N plasma treatment. Obviously, plasma treated
AE25 material is much more susceptible to formation of optically active deep and
intermediary defects and it is likely that the crystalline quality is modified to an extent
that hinders formation of bound excitons. In all other samples the plasma-related changes
may look relatively small compared to changes occurring in AE25. This can be
explained by several factors, the most essential of which are scaling phenomena and the
surface-specific nature of the plasma-related phenomena. In the AE25 nanopowder the
effects are most visible likely because the average crystal size is smaller than a certain
91
threshold, below which luminescent transitions are mediated primarily by the surface and
near-surface states.
III.7 Effects of Polymer Embedding on ZnO Nanopowders [98]
As was shown in the previous section, remote plasma treatment can be
instrumental in modifying PL spectra and identifying the origins of some of the
luminescent features. Another possible route to tailor the surface properties in nanoscale
ZnO would be to embed the nanoparticles into a medium capable of transforming the
surface of ZnO. One of such surroundings can be a polymer, thus leading to a formation
of a ZnO/polymer composite material.
Among others, nanoscale ZnO was suggested as a filler in ZnO/polymer
composites. In particular, a distinct attention has turned to the composites of
nanocrystalline ZnO with polymethyl methacrylate (PMMA) [99 – 119] with possible
utilizations in coatings, fibers, cosmetics, medicine, pharmaceutics, etc. Notable
advantages this composite can offer are as follows. A combination of a transparent
matrix and UV absorbing filler is perfect for UV-shielding materials and filters. Easy to
process matrix material with high impact strength is flexible and a good protective layer
of many surfaces as well as the surfaces of the filler nanoparticles themselves. The filler
particles, on the other hand, are non-toxic, luminescent, with a different refractive index,
which, in combination with PMMA, can bring about functionalities suitable for
luminescent films, displays, light-emitting diodes, photo detectors, and other
optoelectronic coatings. Recently, there were also suggestions of using the ZnO/PMMA
composites in the next generation nonvolatile ultrahigh density memory storage [120-
122]. Good dispersion of ZnO is crucial for a PMMA/ZnO composite, in which the filler
92
and the matrix will mutually and complementarily enhance each other’s properties, such
as thermal and optical characteristics of PMMA on the one hand and the optoelectronic
qualities of ZnO on the other. Several research groups have reported recently on the
relatively successful implementation of this program. For example, in Refs. 99-103, 107-
109, 113-115, 119 it was reported on the rather homogenous distribution of nanoscale
ZnO in the PMMA matrix by surface functionalization [108], emulsion polymerization
[115], in-situ bulk polymerization [102], radiation [119], polymer grafting [107], laser
ablation [100], or solution mixing [113]. Thermal stability was shown to improve [99,
101, 103, 107, 109, 113, 114], UV absorbance increased [102, 108, 109, 115], and
transparency retained [101-103, 108, 109].
Most of the time ZnO is discussed as a filler, the addition of which modifies and
improves the performance of the PMMA matrix per se. On the other hand, there were a
few studies presented on how embedding the nanoscale ZnO into the PMMA matrix
modifies the luminescence of ZnO [99, 100, 104, 109, 110, 112, 116-119].
In our studies we used multi-functional ZnO/PMMA composites suitable for mass
production with improved absorption, luminescence, thermal, etc. properties, and several
well-defined control parameters. They were synthesized by our collaborators at the
National Institute of Chemistry, Ljubljana, using a polyol method that satisfies all of the
above criteria and produces ZnO nanopowders with narrow and controlled particle size
distributions [123]. ZnO/PMMA composites were prepared via bulk polymerization with
ZnO nanoparticles homogeneously dispersed in the PMMA matrix. In the obtained
composite materials a substantially enhanced thermal stability and excellent (> 98%) UV
93
absorption were observed. The synthesis method can be easily utilized in an industrial
process.
In this section we report on the luminescent properties of these nanoscale ZnO
materials ZnO/PMMA nanocomposite materials.
Synthesis of ZnO was performed in various diols – ethylene glycol (EG), propane
diol (PD), and Di(ethylene glycol) (DEG) – with the addition of p-TsOH, which acts as a
catalyst. Fig. 3.44 shows TEM images of ZnO nanoparticles synthesized in different
diols with p-TsOH. Particles synthesized in EG and PD are predominantly irregularly-
shaped or quasi-spherical crystallites with sizes between 20 and 50 nm, whereas those
synthesized in DEG are nanorods with lengths between 30 and 150 nm and widths
between 10 and 30 nm. These results are in good agreement with those reported earlier
[124], where growth rates and coarsening of ZnO nanocrystals strongly depended on the
chosen solvent in the reaction media due to difference in viscosity and bulk solubility.
The dispersion of ZnO in the PMMA matrix was studied by TEM on ultramicrotoned
sections. One of the TEM micrographs presented in Fig. 3.45 shows a homogeneous
distribution of ZnO nanoparticles in PMMA with only a small fraction of larger particles.
Figure 3.44(a). TEM image of ZnO nanoparticles prepared in EG
94
Figure 3.44(b). TEM image of ZnO nanoparticles prepared in PD
Figure 3.44(c). TEM image of ZnO nanoparticles prepared in DEG
95
Figure 3.45. TEM image of a ZnO/PMMA nanocomposite section with ZnO nanoparticles (0.05 wt.%)
prepared in PD.
We carried out luminescence experiments on the as-received and PMMA-
embedded ZnO nanopowders synthesized in three different diols (EG, PD, DEG) to
elucidate the influence of interfacing ZnO with PMMA on its optoelectronic properties.
Fig. 3.46(a) shows room temperature PL spectra of the as-received ZnO
nanospecimens. One can see the band-gap (excitonic) emission at 3.3 eV, a shallow
defect emission at ~ 3.1 eV and a broad defect band at ~ 2.4 eV. The relative intensity of
the defect luminescence (both deep and shallow) vs. the band-gap emission depends on
the diol used in fabrication as well as the average nanoparticle size.
96
2.0 2.2 2.4 2.6 2.8 3.0 3.2
EG
PD
DEG
PL
In
ten
sit
y, a.u
.
E, eV
Figure 3.46(a). Room temperature PL spectra of the as-received ZnO nanoparticles prepared in different
diols
97
2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4
T = 8K
EG
PD
DEG
PL
In
ten
sit
y, a.u
.
E, eV
Figure 3.46(b). Low temperature (8 K) PL spectra of the as-received ZnO nanoparticles prepared in
different diols
In Fig. 3.47(a) we plot raw room temperature PL spectra of the ZnO/PMMA
composites corresponding to the three diols used in ZnO synthesis. These spectra
represent an integrated signal from both the ZnO filler and the PMMA matrix. In order
to separate the PMMA spectral contribution we ran a room temperature PL of the pure
PMMA sample (see. Fig. 3.47(b)), which reveals several broad emission features in the
visible range. When this PMMA background luminescence is subtracted from the spectra
shown in Fig. 3.47(a), one obtains the room temperature PL signal originating essentially
from the ZnO nanoparticles embedded in the PMMA matrix. These spectra are shown in
Fig. 3.47(c). When juxtaposed, figures 3.46(a) and 3.47(c) reveal that the room
98
temperature photoluminescence spectra of the as-grown and PMMA-embedded ZnO
nanoparticles exhibit the same set of spectral items – the 3.3 eV excitonic emission, a
shallow defect emission band at ~ 3.1 eV and a broad defect band centered at ~ 2.4 eV.
In our samples PMMA-embedding does not give rise to any new radiative transitions in
ZnO, in contrast to refs. 99, 100, 104, 109, 116, 118, 119. However, multi-Gaussian
fitting of the spectral features reveals a remarkable result. All the PMMA-embedded
particles produce a significant increase of the band-gap luminescence, whereas the
integrated intensity ratio of the 3.1 eV to 2.4 eV bands remains approximately constant.
Moreover, the scaling of the much stronger ZnO excitonic emission from the PMMA-
embedded DEG-synthesized sample to the PMMA-embedded EG-synthesized sample is
approximately the same as in the as-received DEG-synthesized sample to the as-received
EG-synthesized sample.
On the as-received samples we also ran low temperature (8 K) PL experiments,
the results of which are presented in Fig. 3.46(b). One can see, notably, that for the as-
received ZnO nanoparticles, the effect of low temperature is very similar to that of
PMMA-matrix embedding – a significant enhancement of excitonic luminescent
recombination with a fairly unaffected deep and shallow defect PL bands. Naturally, the
enhanced excitonic emission at low temperatures is produced by a much larger number of
excitons at low temperatures (i.e., the reduction of the thermally activated free charge
carriers).
99
2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4
EG
PD
DEGPL
In
ten
sit
y, a.u
.
E, eV
Figure 3.47(a). Room temperature PL spectra of the ZnO/PMMA composites with ZnO nanoparticles
prepared in different diols.
100
2.0 2.2 2.4 2.6 2.8 3.0 3.2
PL
In
ten
sit
y,
a.u
.
E, eV
Figure 3.47(b). Room temperature PL spectrum of a pure PMMA plate
101
2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4
ALC1
ALC2
ALC3
PL
In
ten
sit
y, a.u
.
E, eV
Figure 3.47(c). Room temperature PL signal only from the ZnO nanoparticles embedded in the PMMA
matrix. The shown spectra result from subtracting the PMMA background luminescence, (Fig. 3.47(b))
from the “raw” spectra shown in Fig. 3.47(a).
Our observations could be explained by at least two qualitatively different
mechanisms – either ZnO/PMMA interface-mediated phenomena or the onset of a
random lasing transition due to scattering on the nanoparticles suspended in the PMMA
matrix with a refractive contrast, similar to Ref. 117. To elucidate the nature of the
significantly amplified excitonic luminescence we performed intensity-dependent PL
experiments in the near band-gap spectral range: the intensity of the incident laser beam
was adjusted to observe the dependence of the PL signal on the excitation power at T = 8
K, 50K, 125 K, 200 K, and 250 K. The results of these experiments at T = 8 K, and T =
102
250 K are shown in Fig. 3.48 (similar dependencies were observed for other
temperatures). Analysis of the spectral fitting for the excitonic PL intensity vs. incident
power (see Fig. 3.49) indicates a rather monotonic dependence, without any suggestion of
threshold phenomena. Normalization of the PL spectra for the same temperature (not
shown) to a single incident intensity shows that both the full widths at the half maximum
and the positions of the peaks are independent of the incident intensity, i.e., no indication
of a lasing threshold. These results were observed for T = 8 K, 50 K, 125 K, 200 K, and
250 K. Thus, we did not observe any threshold in the dependence of the intensity,
position or width of the excitonic peak vs. excitation power.
3.1 3.2 3.3 3.4
T= 8 K
Pump intensity
100%
50%
25%
12.5%
10%
5%
Em
issio
n in
ten
sit
y, a.u
.
E, eV
Figure 3.48(a). Effects of incident laser intensity on the near band-gap PL spectra of a ZnO/PMMA
nanocomposite sample (with ZnO nanoparticles prepared in PD) at T = 8 K
103
3.0 3.1 3.2 3.3 3.4
T= 250 K
Pump intensity
100%
50%
25%
12.5%
10%
5%
Em
iss
ion
in
ten
sit
y, a.u
.
E, eV
Figure 3.48(b). Effects of incident laser intensity on the near band-gap PL spectra of a ZnO/PMMA
nanocomposite sample (with ZnO nanoparticles prepared in PD) at T = 250 K
104
0 20 40 60 80 100
8 K
250 K
Em
iss
ion
in
ten
sit
y,
a.u
.
Pump intensity, % of the max
Figure 3.49. Integrated near band-gap PL luminescence at 9 K and 250 K as a function of incident laser
intensity for a ZnO/PMMA nanocomposite sample (with ZnO nanoparticles prepared in PD). Straight lines
represent linear fits.
We conclude thereby that it is likely that the significant increase of the excitonic
luminescence at room temperature in the PMMA-encapsulated ZnO nanoparticles is
associated with the ZnO/PMMA interface phenomena. We suggest that the mechanism
may be similar to that proposed in refs. 110, 112 for the low temperature excitonic
emission in thin films prepared with PMMA-coated ZnO nanowires. The authors of refs.
110, 112 speculated that PMMA encapsulation reduces the density of traps at the surface
of ZnO, and such screening leads to the increased concentration of the near-surface
105
excitons. Our results show that such behavior is, probably, a fairly generic effect, which
occurs in a wide range of temperatures and ZnO nanomorphologies.
Summarizing, in this section we demonstrated that the room temperature UV
luminescence of the PMMA-embedded nanosize ZnO structures is significantly enhanced
because of the essentially ZnO/PMMA interfacial phenomena.
106
Chapter IV. Conclusions
To address the objectives of our studies, we successfully designed and set up a
working multifunctional optical analysis system, equipped with several laser beams, high
resolution/high sensitivity detection system, flexible software interface, and a manifold of
variable experimental parameters (temperature, excitation intensity, polarization, etc.).
PL studies of several commercial ZnO nanopowders revealed substantial sample-
to-sample variations in luminescent properties, in some cases overshadowing possible
scaling behavior. However, many results were consistent with the hypothesis of crystals
with the surface and subsurface layers rich in defects. The discrepancies are observed in
spectral domains corresponding to optical transitions of different origins, as well as in
different temperature ranges. Some microscopic parameters can be derived from the
analysis of the obtained spectra. For example, temperature-dependent PL spectra from
the ZnO nanopowders were analyzed in detail for the bound-exciton range and the
numerical fits of the peak positions yielded activation energies that suggested different
channels of recombination for the BEx. Also, fits for the FWHM show nonlinear
behavior, indicating contribution from surface phonons. The nanopowder samples were
also treated by remote plasmas of hydrogen, nitrogen, and oxygen. We demonstrated that
those plasma species induce a variety of changes in the deep defect visible emission as
well as in the BEx luminescence, most likely associated with the surface/subsurface
states, indicating that modifications in the plasma-affected surface/subsurface layers are
significant enough to be readily observable in their optoelectronic signatures. Most
107
visible plasma-induced changes occur for the samples with the smallest grains, partially
confirming the aforementioned scaling hypothesis.
To create better conditions for observing scaling behavior ZnO nanorods were
prepared at relatively low temperatures by homogeneous precipitation from zinc nitrate
and urea in water/ethylene glycol mixture. Crystal size and morphology were controlled
by growth time and solvent composition. PL experiments on the precipitated ZnO
nanorod samples revealed a strong correlation between the size/morphology of the
nanocrystals and their defect optoelectronic properties, confirming hypothesis that the
contribution of defects should increase with the decreasing size of the nanocrystal.
Smaller crystallites exhibited intense deep-defect luminescence and enhanced emission
related to near-surface excitonic recombination. Longer growth times lead to formation
of well-defined nanorods with hexagonal symmetry exhibiting reduced defect emission.
Since an increase of the surface-to-volume ratio may occur not only with an
average size decrease for a nanocrystal but also with a decrease in its dimensionality,
samples with various quasi-fractal dimensionalities were probed with PL to test this
proposition. These were precipitated quasi-fractal-dimensional nanostructures of
hydrozincite containing a ZnO phase. Crystal morphology of the samples was accurately
controlled by the growth time. We observed a strong correlation between the
morphology of the samples and their optoelectronic properties. Our results again
indicated that an increase of the free surface in a nanocrystal generates higher relative
concentration of defects, consistent with the model of defect-rich surface and subsurface
layers.
108
ZnO nanostructures with rich morphology obtained by thermal treatment of
hydrozincite particles were also studied by us. Luminescence properties of the samples
probed by PL at 8 K and room temperature exhibited a remarkable correlation with
specimens’ nanomorphology. Luminescent features at ~ 2.0-2.2 eV, ~ 2.4-2.5 eV, ~ 2.65
eV, ~ 2.9 eV, ~ 3.0-3.1 eV, and ~ 3.3 eV were observed in most specimens, although
their relative intensity and temperature dependence were specific to an individual group
of samples vis-à-vis their growth history and morphology.
Another approach to the alteration of the surface properties in nanoscale ZnO
samples would be modification of the environment in which the nanocrystals reside.
Homogeneous ZnO/PMMA nanocomposites were prepared by incorporating ZnO
nanoparticles synthesized in various diols into a PMMA matrix by the free-radical bulk
polymerization. Room temperature PL spectra of the as-grown and PMMA-embedded
ZnO nanoparticles exhibited an excitonic band gap emission at 3.3 eV, a near band-gap
emission at ~ 3.1 eV and a broad defect band centered at ~ 2.4 eV. Relative intensity of
the defect vs. band gap luminescence depended on the parameters of ZnO preparation as
well as the average particle size. However, PMMA-embedded particles produced a much
stronger excitonic luminescence, whereas the ratio of the 3.1 eV to 2.4 eV remained
approximately constant. There was no indication of random lasing threshold pointing to
the ZnO/PMMA interfacial origin of the enhanced band-gap emission. The increase of
the excitonic emissions was most likely produced by the interface phenomena between
the ZnO and PMMA, such as, e.g., increased concentration of near-surface excitons.
109
Chapter V. Future Plans
There are several upshots from the results discussed here that may naturally lead
to an expansion of research on surface defect properties in nanocrystalline systems. The
most relevant issues to address in the nearest future are as follows.
1) Plasma-treatment protocols could be refined and optimized to controllably
amplify the effects of plasma species.
2) For the samples treated with variety of plasma species, a detailed spectral
analysis of the generated modifications of the BEx range could be performed,
and identification of surface-specific states attempted. Other spectral domains
(such as DAP, phonon replicas, visible bands, etc.) may be analyzed, to
correlate with the plasma-driven changes.
3) Effects of remote-plasma treatment of the larger quantity of precipitated ZnO
nanosytems of various morphologies should be addressed.
4) Effects of annealing of our samples, stand alone or simultaneous with plasma,
have not been investigated. Since temperature is one of the most effective
processing parameters, various annealing procedures will applied to the
specimens of interest.
5) The effects of polymer embedding (e.g., a fine spectral analysis) need further
elucidation for a variety of other nanocrystals as well as polymer matrices.
6) The optical train designed here could be customized to run Raman scattering
experiments, which has not been employed so far for experiments with our
110
samples. Raman spectroscopy may provide important information about
surface stresses, possible signatures of surface phonon modes as well as
confinement phenomena.
111
References
[1] P. Fons, K. Iwata, S. Niki, A. Yamada, K. Matsubara, J. Cryst. Growth, 627, 201
(1999).
[2] Y. Chen, D.M. Bagnall, H.J. Koh, K.T. Park, K. Hiraga, Z.Q. Zhu, T. Yao, J. Appl.
Phys., 84, 3912 (1998).
[3] R. D. Vispute, V. Talyansky, S. Choopun, Appl. Phys. Lett., 73, 348 (1998).
[4] Y. Liu, C. R. Gorla, S. Liang, N. Emanetoglu, Y. Lu, H. Shen, M. Wraback, J.
Electron. Mater., 29, 69 (2000).
[5] M. Kasuga, S. Ogawa, Jpn. J. Appl. Phys., 22, 794 (1983).
[6] N. Takahashi, K. Kaiya, T. Nakamura, Y. Momose, H. Yamamoto, Jpn. J. Appl.
Phys., 38, L454 (1999).
[7] P. X. Gao, J. Song, J. Liu, Z. L. Wang, Adv. Mater., 19, 67 (2007).
[8] H. Noh, M. Scharrer, M. A. Anderson, R. P. H. Chang, H. Cao, Phys. Rev. B, 77,
115136 (2008).
112
[9] Q. Zhang, T. P. Chou, B. Russo, S. A. Jenekhe, G. Cao, Int. Ed. Engl., 47, 2402
(2008).
[10] T. Dietl, H. Ohno, F. Matsukura, J. Cibert, D. Ferrand, Science, 287, 1019 (2000).
[11] S.J. Pearton, G.T. Thaler, M.E. Overberg, B. Gila, R. Frazier, C.R. Abernathy, J.
Phys. Condens. Matter, 16, R209 (2004).
[12] S.J. Pearton, W.H. Heo, M. Ivill, D.P. Norton, T. Steiner, Semicond. Sci. Technol,
19, R59 (2004).
[13] J. Liqiang, W. Baiqi, X. Baifu, L. Shudan, S. Keying, C. Weimin, and F. Honggang,
J. Solid State Chem., 177, 4221-4227 (2004).
[14] M. Bitenc, M. Marinšek, Z. Crnjak Orel, J. Eur. Ceram. Soc., 28, 2915 (2008).
[15] Z. Jing, J. Zhan, Adv. Mater., 20, 4547 (2008).
[16] C. G. van De Walle, Physica B, 308, 899 (2001).
[17] A. Janotti, C.G. van De Walle, Phys. Rev. B, 76, 165202 (2007).
[18] C.G. van De Walle, Phys. Rev. Lett., 85, 1012 (2000).
113
[19] Ü. Özgür, et. al. J. Appl. Phys., 98, 041301 (2005) and references therein.
[20] Q. Zhao, T.C. Cai, S. Wang, R. Zhu, Z. Liao, D. Yu, Appl. Phys. A – Mater.
Sci. Process., 100, 165 (2010).
[21] Z.W. Peng, G.Z. Dai, W.C. Zhou, Appl. Surf. Sci., 256, 6814 (2010).
[22] Z.W. Peng, G.Z. Dai, P. Chen, P. Cheng, Q. Zhangan, Q. Wan, B. Zou, Mater. Lett.,
64, 898 (2010).
[23] S. Maji, P. Bhattacharyya, A. Sengupta, H. Saha, Adv. Sci. Lett., 3, 154 (2010).
[24] T.K. Jia, W.M. Wang, F. Huang, Z.Y. Fu, X.H. Ma, W. Guo, .Rare Metal Mat. Eng.,
38, 979 (2009).
[25] D.H. Fan, R. Zhang, X.H. Wang, Phys. E- Low Dimens. Syst. Nanostruct., 42, 2081
(2010).
[26] H. Song, J.H. Kim, E.K. Kim, S.M. Wang, Thin Solid Films, 517, 3927 (2009).
[27] C. Majidi, M. Haataja, D.J. Srolovitz, Smart Mater. Struct., 19, 055027 (2010).
114
[28] R. Chowdhury, S. Adhikari, F. Scarpa, Phys. E- Low Dimens. Syst. Nanostruct., 42,
2036 (2010).
[29] H.Y. Yang, S.F. Yu, G.P. Li, T. Wu, Opt. Express, 18, 13647 (2010).
[30] H. Kalt, J. Fallert, R.J.B. Dietz, J. Sartor, D. Schneider, C. Klingshirn, Phys. Status
Solidi B-Basic Sol. State Phys., 247, 1448 (2010).
[31] S.M. Zhou, P. Wang, S. Li, B. Zhang, H. Gong, Z. Du, Chin. Phys. Letters, 25, 4446
(2008).
[32] Z.L. Wang, Annu. Rev. Phys, Chem., 55, 159 (2004).
[33] W.L. Hughes, Z.L. Wang, Appl. Phys. Lett., 82, 2886 (2003).
[34] M.M. Rahman, A.J.S. Ahammad, J.H. Jin, S.J. Ahn, J. Lee, Sensors, 10, 4855
(2010).
[35] U. Ozgur, D. Hofstetter, H. Morkoc, Proceedings of the IEEE, 98, 1255 (2009).
[36] R. Yakimova, G. Steinhoff, R.M. Petoral, G.R. Yazdi, C. Valhberg, V.
Khranovskyy, Biosens. Bioelectron., 22, 2780 (2007).
115
[37] Z.L. Wang, Nano Research, 1, 1 (2008).
[38] T. Hirai, Y. Harada, S. Hashimoto, N. Ohno, T. Itoh, J. Lumin., 113, 115 (2005).
[39] J. Grabowska, K. K. Nanda, E. McGlynn, J.P. Mosnier, M.O. Henry, A. Beaucamp,
A. Meaney, J. Mat. Sci.- Mater. Electron., 16, 397 (2005).
[40] J. Grabowska, A. Meaney, K.K. Nanda, J.P. Mosnier, M.O. Henry, J.R. Duclère, E.
McGlynn, Phys. Rev. B, 71, 115439 (2005).
[41] I. Shalish, H. Temkin, V. Narayanamurti, Phys. Rev. B, 69, 245401 (2004).
[42] L. Wischmeier, T. Voss, S. Börner, W. Schade, Appl. Phys. A, 84, 111 (2006).
[43] S. Sakohara, M. Ishida, M.A. Anderson, J. Phys. Chem. B, 102, 10169 (1998).
[44] L. Guo, S. Yang, C. Yang, P. Yu, J. Wang, W. Ge, G.K.L. Wong, Appl. Phys. Lett.,
76, 2901 (2000).
[45] J. Li, D. Zhao, X. Meng, Z. Zhang, J. Zhang, D. Shen, Y. Lu, X. Fan, J. Phys. Chem.
B, 110, 14685 (2006).
116
[46] Y. Yang, B.K. Tay, X.W. Sun, J.Y. Sze, Z.J. Han, J. X. Wang, X. H. Zhang, Y. B.
Li, S. Zhang, Appl. Phys. Lett., 91, 071921 (2007).
[47] Y. Yang, X.W. Sun, B.K. Tay, P.HT. Cao, J.X. Wang, X.H. Zhang, J. Appl. Phys.,
103, 064307 (2008).
[48] B.K. Meyer, H.D. Alves, H.M. Hofmann, W. Kriegseis, Phys. Stat. Sol., 241, 231
(2004) and references therein.
[49] A.F. Kohan, G. Ceder, D. Morgan, Phys. Rev. B., 61, 15019 (2000).
[50] D.W. Hamby, D.A. Lucca, J.K. Lee, M. Nastasi, H.S. Kang, S.Y. Lee, Nuc.
Instrum. Methods Phys. Res., 249, 196 (2006).
[51] J.A. Paramo, R.M. Peters, C.A. Quarles, H. Vallejo, Y.M. Strzhemechny, IOP
Conf. Ser.: Mater. Sci. Eng., 6, 012030 (2009).
[52] Y.P. Varshni, Physica, 34, 149 (1967).
[53] B.K. Meyer, J. Sann, D. Hofman, C. Neumann, A. Zeuner, Semicond. Sci. Technol.,
20, S62 (2005).
117
[54] K. Vanheusden, C.H. Seager, W.L. Warren, D.R. Tallant, Appl. Phys. Lett. 68, 403
(1996) and references therein.
[55] N. Garces, L. Wang, L. Bai, N. Gilles, L.E. Halliburton, Appl. Phys. Lett., 81, 622
(2002) and references therein.
[56] F. Su, Y. Liu, W. Chen, J. Appl. Phys., 100, 013107 (2006).
[57] A. Djurisic, Y.H. Leung, K.H. Tam, Appl. Phys. Lett., 88, 103107 (2006).
[58] D. Li, Y.H. Leung, A. Djurisic, Appl. Phys. Lett., 85, 1601 (2004).
[59] A. Djurisic, K. Tam, Y. Hsu. Nanotechnology, 18, 095702 (2007) and references
therein.
[60] P. Kasai, Phys. Rev. Lett., 130, 989 (1963).
[61] S.D. Kshirsagar, V.V. Nikesh, S. Mahamun, Appl. Phys. Lett., 89, 053120 (2006).
[62] R.M. Peters, J.A. Paramo, C.A. Quarles, Y.M. Strzhemechny, App. Accel. Res.
Indust., 965 (2009).
118
[63] M. Bitenc, P. Podbršček, Z. Crnjak Orel, M.A. Cleveland, J.A. Paramo, R.M. Peters,
Y.M. Strzhemechny, Cryst. Growth Des., 9, 997 (2009).
[64] M. Bitenc, Z. Crnjak Orel, Mater. Res. Bull., 5, 5 (2008)
[65] S.H. Ko, I. Park, H. Pan, N. Misra, M.S. Rogers, C.P. Grigoropoulos, A.P. Pisano,
Appl. Phys. Lett., 92, 154102 (2008).
[66] B.G. Wang, E.W. Shi, W.Z. Zhong, Cryst. Res. Technol., 33, 937 (1998).
[67] S.W. Liu, H.J. Zhou, A. Ricca, R. Tian, M. Xiao, Phys. Rev. B, 77, 113311 (2008).
[68] S. Hayashi, N. Nakamori, H. Kanamori, J. Phys. Soc. Japan, 46, 176 (1979).
[69] A.M. Verges, A. Mifsud, C.J. Serna, J. Chem. Soc. Faraday Trans., 86, 959 (1990)
[70] J. Fallert, R. Hauschild, F. Stelzl, A. Urban, M. Wissinger, H. Zhou, C. Klingshirn,
H. Kalt, J. Appl. Phys. 101, 073506 (2007).
[71] Y.M. Strzhemechny, J. Nemergut, P.A. Smith, J. Bae, D.C. Look, L.J. Brillson, J.
Appl. Phys., 94, 4256 (2003).
[72] J.A. Paramo, Y.M. Strzhemechny, M. Bitenc, Z. Crnjak Orel, in print.
119
[73] F.A. Sigoli, M.R. Davolos, M. Jafelicci, J. Alloys Compd., 262, 292 (1997).
[74] S.A.M. Lima, F.A. Sigoli, M. Jafelicci, M.R. Davolos, Int. J. Inorg. Mater., 3, 749
(2001).
[75] N. Kanari, D. Mishra, I. Gaballah, B. Dupre, M.P.E. Engn, Thermochim. Acta, 410,
93 (2004).
[76] C.L. Yan, D.F. Xue, J. Phys. Chem. B, 110, 11076 (2008).
[77] Y. Liu, J.E. Zhou, A. Larbot, M. Persin, J. Mater. Process. Technol., 189, 379
(2007).
[78] Z.H. Jing, J.H. Zhan, Adv. Mater., 20, 4547 (2008).
[79] R. Wahab, S.G. Ansari, Y.S. Kim, M.A. Dar, H.S. Shin, J. Alloys Compd., 461, 66
(2008).
[80] M. Bitenc, Z.Crnjak Orel, Materiali in Tehnologije, 45, 287 (2011).
[81] P.B. Himelfarb, G.W. Simmons, K. Klier, R.G. Herman, J. Catal., 93, 442 (1985).
120
[82] P. Porta, S. Derossi, G. Ferraris, M. Lojacono, G. Minelli, G. Moretti, J. Catal., 109,
367 (1988).
[83] G. De Giudici, F. Podda, R. Sanna, E. Musu, R. Tombolini, C. Cannas, A. Musinu,
M. Casu, Am. Mineral. 94, 1698 (2009).
[84] M. Gaft, R. Reisfeld, G. Panczer, Springer, Berlin, (2005).
[85] D. Japic, J.A. Paramo, M. Marinsek, Y. M. Strzhemechny, Z. Crnjak Orel, J. Lumin.,
132, 1589 (2012).
[86] X. Yang, C. Shao, H. Guan, X. Li, J. Gong, Inor. Chem. Commun., 7, 176 (2003).
[87] Y. Dong, Z.Q. Fang, D.C. Look, D.R. Doutt, G. Cantwell, J. Zhang, J.J. Song, L.J.
Brillson, J. Appl. Phys., 108, 103718 (2010).
[88] S.K. Chaudhuri, M. Ghosh, D. Das, A.K. Raychaudhuri, J. Appl. Phys., 108, 064319
(2010).
[89] B.X. Lin, Z.X. Fu, Y.B Jia,. Appl. Phys. Lett., 79, 943 (2010).
[90] D.C. Reynolds, D.C. Look, B. Jogai, J. Appl. Phys., 89, 6189 (2001) and references
therein.
121
[91] L. Wischmeier, T. Voss, I. Rückmann, J. Gutowski, Nanotech., 19, 135705 (2008)
and references therein.
[92] Y.M. Strzhemechny, J. Vac. Sci. Tech., 24, 1233 (2006).
[93] Y.M. Strzhemechny, H.L. Mosbacker, S.H. Goss, J. Elec. Mat., 34, 399 (2005).
[94] Y.M. Strzhemechny, H.L. Mosbacker, D.C. Look, D.C. Reynolds, C.W. Litton, N.Y.
Garces, N.C. Giles, L.E. Halliburton, S. Niki, L.J. Brillson, App. Phys. Let., 84, 2545
(2004).
[95] Y.M. Strzhemechny, J. Nemergut, J. Bae, D. C. Look, L. J. Brillson, Mat. Res. Soc.
Symp., 744, M.3.9.1 (2003).
[96] H.J. Ko, Y.F. Chen, S.K. Hong, H. Wenisch, T. Yao, Appl. Phys. Lett., 77, 3761
(2000).
[97] E. Tomzig, H. Helbig, J. Lumin., 14, 403 (1976).
[98] J.A. Paramo, Y.M. Strzhemechny, A. Anzlovar, M. Zigon, Z. Crnjak Orel, J. Appl.
Phys., 108, 023517 (2010).
122
[99] Q.H. Chen, S.Y. Shi, W.G. Zhang, Colloid Polym. Sci., 287, 533 (2009).
[100] Q.H. Chen, W.G. Zhang, J. Non-Cryst. Solids, 353, 374 (2007).
[101] M.M. Demir, P. Castignolles, Ü. Akbey, G. Wegner, Macromolecules, 40, 4190
(2007).
[102] M.M. Demir, K. Koynov, Ü. Akbey, C. Bubeck, I. Park, I. Lieberwirth, G. Wegner,
Macromolecules, 40, 1089 (2007).
[103] M.M. Demir, M. Memesa, P. Castignolles, G. Wegner, Macromol. Rapid
Commun., 27, 763 (2006).
[104] X.W. Du, Y.S. Fu, J. Sun, X. Han, J. Liu, Semicond. Sci. Technol., 21, 1202 (2006).
[105] A. Endruweit, A.D. Alobaidani, D. Furniss, A.B. Seddon, T. Benson, M.S.
Johnson, A.C. Long, Opt. Lasers Eng., 46, 97 (2008).
[106] S. Hess, M.M. Demir, V. Yakutkin, S. Baluschev, G. Wegner, Macromol. Rapid
Commun., 30, 394 (2009).
[107] R.Y. Hong, J.Z. Qian, J.X. Cao, Powder Technol., 163, 160 (2006).
123
[108] V. Khrenov, M. Klapper, M. Koch, K. Müllen, Macromol. Chem. Phys., 206, 95
(2005).
[109] S. Li, M.S. Toprak, Y.S. Jo, J. Dobson, D.K. Kim, M. Muhammed, Adv. Mater.,
19, 4347 (2007).
[110] K.W. Liu, R. Chen, G.Z. Xing, T. Wu, H.D. Sun, Appl. Phys. Lett., 96, 023111
(2010).
[111] P. Liu, Z. Su, J. Macromol. Sci., Phys., 45, 131 (2006).
[112] J.P. Richters, T. Voss, L. Wischmeier, I. Rückmann, J. Gutowski, Appl. Phys. Lett.,
92, 011103 (2008).
[113] D. Sun, N. Miyatake, H.J. Sue, Nanotechnology, 18, 215606 (2007).
[114] E. Tang, G. Cheng, X. Ma, Powder Technol., 161, 209 (2006).
[115] E. Tang, G. Cheng, X. Pang, X. Ma, F. Xing, Colloid Polym. Sci., 284, 422 (2006).
[116] D. Vollath, D.V. Szabó, S. Schlabach, J. Nanopart. Res., 6, 181 (2004).
[117] C. Vutha, S.K. Tiwari, R.K. Thareja, J. Appl. Phys., 99, 123509 (2006).
124
[118] Z.G. Wang, X.T. Zu, H.J. Yu, X. He, S. Zhu, Q.M. Wei, L.M. Wang, Nucl.
Instrum. Methods Phys. Res. B, 250, 196 (2006).
[119] Z.G. Wang, X.T. Zu, S. Zhu, X. Xiang, L.M. Fang, L.M. Wang, Phys. Lett. A, 350,
252 (2008).
[120] K.H. Park, J.H. Jung, F. Li, D.I. Son, T.W. Kim, Appl. Phys. Lett., 93, 132104
(2008).
[121] D.I. Son, D.H. Park, W. K. Choi, S.H. Cho, W.T. Kim, Nanotechnology, 20,
195203 (2009).
[122] D.I. Son, C.H. You, W.T. Kim, J.H. Jung, T.W. Kim, Appl. Phys. Lett., 94, 132103
(2009).
[123] A. Anžlovar, Z. Crnjak Orel, M. Žigon, unpublished.
[124] M. Bitenc, Z. Crnjak Orel, Mater. Res. Bull., 44, 381 (2009).
VITA
Personal Jorge Antonio Paramo
Background Son of Jorge Fabian Paramo and Blanca Esmeralda Villacis
Education Bachelor of Science, Physics & Mathematics,
Midwestern State University, Wichita Falls, Texas, 2004
Bachelor of Science, Manufacturing Engineering Technology,
Midwestern State University, Wichita Falls, Texas, 2005
Masters in Business Administration, Finance and Investments
Texas Christian University, Fort Worth, 2012
Doctor of Philosophy, Physics,
Texas Christian University, Fort Worth, 2012
Experience Market Operations Summer Associate, EDP Renewables NA
Houston, TX, 2012
Investment Management Intern, Cook Children’s Foundation
Fort Worth, TX, 2011 - 2012
Portfolio Manager/Analyst, Educational Investment Fund
Texas Christian University, Fort Worth, TX, 2011
Research/Teaching Assistantship, Texas Christian University
Fort Worth, TX, 2005- 2010
Teaching Assistant, Midwestern State University
Wichita Falls, TX, 2003-2005
Professional American Physical Society, Materials Research Society,
Society of Manufacturing Engineers,
National Society of Hispanic Physicists
Memberships Society of Physics Students
Sigma Pi Sigma National Physics Honor Society
1. “Growth-morphology-luminescence correlation in ZnO-containing nanostructures
synthesized from aqueous solutions”, D. Japic, J. A. Paramo, M. Marinsek, Y. M.
Strzhemechny, and Z. Crnjak Orel, Journal of Luminescence 132, 1589–1596
(2012).
2. “Effects of thermal annealing on the structural and optical properties of carbon-
implanted SiO2”, P. R. Poudel, J. A. Paramo, D. R. Diercks, Y. M. Strzhemechny,
and F. D. McDaniel, Journal of Nanoscience and Nanotechnology 12, 1835-1842
(2012).
3. “Effects of thermal annealing on the formation of buried β-SiC by ion
implantation”, P. R. Poudel, B. Rout, D. R. Diercks, J. A. Paramo, Y. M.
Strzhemechny, and F. D. McDaniel, Journal of Electronic Materials, 40, 1998 –
2003 (2011)
4. “Enhanced room temperature excitonic luminescence in ZnO/polymethyl
methacrylate nanocomposites prepared by bulk polymerization”, J. Antonio
Paramo, Yuri M. Strzhemechny, Alojz Anzlovar, Majda Zigon, and Zorica Crnjak
Orel, Journal of Applied Physics, 108, 023517 (2010).
5. “Growth of zinc oxide particles in the presence of silicon”, P. Podbršček, G.
Dražić, J. Paramo, Y. M. Strzhemechny, J. Maček, and Z. Crnjak Orel, Cryst Eng
Comm 12, 3071-3079 (2010).
6. “In situ surface photovoltage spectroscopy of ZnO nanopowders processed by
remote plasma”, R. M. Peters, S. P. Glancy, J. A. Paramo, and Y. M.
Strzhemechny, in Zinc Oxide and Related Materials - 2009, ed. S. Durbin, M.
Allen, and H. von Wenckstern (Mater. Res. Soc. Symp. Proc. Volume 1201,
Warrendale, PA, 2010), 1201-H03-03.
7. “Dual wavelength emissive ZnO tetrapods: effects of erbium/germanium doping”,
X. Huang, J. L. Coffer, J. A. Paramo, and Y. M. Strzhemechny, Crystal Growth &
Design 10, 32 (2010).
8. “Defect properties of ZnO nanopowders and their modifications induced by
remote plasma treatments”, J. A. Paramo, R. M. Peters, C. A. Quarles, H. Vallejo,
Y. M. Strzhemechny, IOP Conf. Series: Materials Science and Engineering 6,
012030 (2009).
9. “Correlation between optoelectronic and positron lifetime properties in as-
received and plasma-treated ZnO nanopowders”, R. M. Peters, J. A. Paramo, C.
A. Quarles, Y. M. Strzhemechny, in Application of Accelerators in Research and
Industry, ed. F. D. McDaniel and B. L. Doyle, AIP, pp. 965-969, (2009).
10. “Correlation between morphology and defect luminescence in precipitated ZnO
nanorod powders”, M. Bitenc, P. Podbršček, Z. Crnjak Orel, M. A. Cleveland, J.
A. Paramo, R. M. Peters, Y. M. Strzhemechny, Crystal Growth & Design 9, 997
(2009).
ABSTRACT
INFLUENCE OF MORPHOLOGY AND SURFACE CONDITIONS ON DEFECT
PROPERTIES OF NANOCRYSTALLINE ZINC OXIDE
by Jorge Antonio Paramo, Ph.D., 2012
Department of Physics and Astronomy
Texas Christian University
Dissertation Advisor:
Dr. Yuri M. Strzhemechny, Associate Professor of Physics
Performance of nanoscale ZnO-based systems depends on the nanomorphology
and surface conditions, in particular surface defect states. We investigated the impact of
the surface/interface phenomena on the defect-related properties for ZnO-containing
nanocrystalline systems. To probe these surface-related effects we employed
photoluminescence (PL) spectroscopy.
Among others, we studied ZnO-containing nanocrystalline specimens grown by
wet precipitation with size and morphology controlled by the synthesis parameters. We
observed a strong correlation between defect-related luminescence and morphological
sample variations. For example, there was a consistent relationship between the
surface/volume ratio and the relative intensity of the PL defect emission, indicating
strong influence of the optically-active surface states.
Commercially available ZnO nanopowders from several vendors were
investigated by PL. Observation of the size effects was somewhat overshadowed by the
sample-to-sample differences in quality, and thus defects’ content and abundance.
Temperature-dependent PL measurements in the bound-exciton (BEx) range were
performed to elucidate surface-related corrections to the excitons thermodynamics.
Specially, calculations for the excitonic activation energies indicated strong dependences
of the nanocrystal size on the predominant BEx dissociation channels. Also, we observed
nonlinear dependences of BEx peak broadening on temperature suggesting surface
phonon contributions. We used remote plasma treatments to tailor surface defect
properties of ZnO nanopowders. We report on the plasma-driven modifications of defect
optical signatures such as BEx and visible luminescence. Besides, plasma treatments
revealed size-dependent effects in the studied specimens.
PL was used to study ZnO nanoparticles embedded into a polymethyl
methacrylate (PMMA) matrix by bulk polymerization. We found that the polymer
encapsulation enhances room-temperature excitonic luminescence by several orders of
magnitude, similar to the effects of low temperatures on the as-received nanoparticles.
At the same time, relative intensities of the visible defect luminescence did not change
noticeably after the PMMA embedding. Intensity-dependent PL experiments showed no
indication of a random lasing threshold, thus we attributed the observed phenomena to
the influence of the PMMA/ZnO interfaces.
Top Related