We report on the detection mechanism of spontaneous polarization using electrostatic force microscopy in zinc oxide nanowires and nanostars grown by vapor-liquid-solid technique. Optical and structural properties are investigated in detail to understand the complex ZnO nanostructures comprehensively. Calculations are carried out to estimate the electric field from the change in interleave amplitude induced by the electrostatic force due to the spontaneous polarization effects. Attraction of the probe between the tip and the sample varies for different structures with a stronger attraction for nanostars as compared to nanowires. Strength of electric field is dependent on the orientation of nanowires and nanostars c-axis with measured magnitude of electric field to be ∼107 V/m and 108 V/m respectively. This technique presents a unique detection mechanism of built-in spontaneous polarization and electric field from polar ZnO nanowires with applications in voltage gated ion channels, nano-bio interfaces, optoelectronic and photonic devices.

Zinc oxide (ZnO) is an interesting and useful semiconductor material because of its various uses in both practical and fundamental studies.1,2 ZnO crystals are generally of three types: wurtzite, zinc blende, and rock salt. Wurtzite ZnO is the most stable one when considered at the ambient conditions and represents one of the important class of crystal structures that possess versatile properties that are important for applications in optoelectronics, photonics, sensors, photovoltaics, and electrical devices.3–6 Wurtzite ZnO crystals have electrical polarity and their polarization can be categorized as either induced polarization such as piezoelectric interaction or built-in spontaneous polarization. The piezoelectric polarization occurs due to an external strain that causes the piezoelectric crystal to have a macroscopic polarization as a result of displacement of ions, which has been well researched.7,8 Unlike piezoelectricity, ZnO also tends to have a built-in spontaneous polarization along the growth axis (c-axis). This inbuilt spontaneous polarization induces a natural electric field in these crystals that affects the device properties significantly.9 Such an effect can be utilized in many device structures such as bio-nano interfaces or high electron mobility transistor (HEMT) structures, opening possibilities in various applications such as nanobiosciences and optoelectronics.10,11

Extensive research has been done in the past decade on evaluating the properties of quasi-one dimensional ZnO nanostructures12,13 yet the reports on the effects of spontaneous polarization in nanoscale structures have not been investigated extensively yet. In this work, we have prepared two different kinds of zinc oxide nanostructures using vapor-liquid-solid (VLS) growth mechanism under different growth conditions, and report nanowires (NW) and nanostars (NS)-like formation. We have then investigated their luminescence and structural properties in detail. In our previous study, we proposed a method to estimate the electric field of a spherical zinc oxide quantum dot.14 Herein, we estimate the magnitude of electric field for complex structures such as ZnO nanowire and nanostars from the electrostatic force analysis by measuring the amplitude change in the atomic force microscopy (AFM) probe.

ZnO nanowires and nanostars have been grown using VLS growth mechanism. The growth is carried out in a 2″ diameter quartz tube at atmospheric pressure with the tube inserted inside an isothermal furnace. The source mixture was placed at the center of the quartz tube with the source consisting of zinc oxide (99.99%, Sigma Aldrich) mixed with graphite powder (Aldrich Company) in a weight ratio of 1:1. One end of the tube was connected to an Argon gas flow controller while the other end was outflowing gas through the exhaust. A 4 nm gold layer is deposited on the Si substrate using an e-beam evaporation tool (Varian) at a deposition rate of 1 Å/s. The system was heated at 910 °C with a constant gas flow of Ar at ∼150 sccm for ∼40–50 min after the system was stabilized at desired temperature. The silicon substrate (n-type, 100 plane) with a 4 nm gold catalyst is placed downstream from the quartz tube 2 cm away from the source powder for NWs growth. For NSs growth, the gas flow, temperature, and duration is kept exactly the same, but the distance of the substrate from the source is increased to ∼6 cm which resulted in a decreased substrate temperature to ∼650 °C. The carbothermal reaction at the growth temperature can be explained as the following reaction equations:

Carbothermalreaction:ZnO(s)+C(s)=Zn(v)+CO(g),
(1)
Catalystalloyformationreaction:Zn(v)+Au(s)=AuZn(l),
(2)
Overallformationreaction:AuZn(l)+Zn(v)+12O2=AuZn(l)+ZnO(s).
(3)

In order to confirm the growth of ZnO NWs and NSs, variable-pressure Hitachi S-3000N scanning electron microscopy (SEM) measurements have been carried out. Chemical composition is measured using energy dispersive spectroscopy (EDS), whereas optical properties are investigated using photoluminescence spectroscopy at room temperature with a 325 nm He-Cd laser using an Acton 2500i spectrometer. Transmission electron microscopy (TEM) of the nanostructures is carried out using a JEOL JEM 3010 electron microscope. For TEM measurements, samples are prepared using ZnO NWs and NSs samples and removing them from the surface. The detached NWs and NSs are transferred using isopropanol by placing a drop of them on a 200 mesh copper grid having a carbon layer with holes on them (SPI Supplies, West Chester, PA).

Finally, AFM and electrostatic force microscopy (EFM) measurements are performed on a Bruker Dimension Icon AFM (Billerica, MA). The AFM samples are prepared by dropping detached NWs and NSs on a gold coated silicon wafer. A conductive SCM-PIT probe is used for EFM measurements to scan the surface in a lift mode profile.

Figures 1(a) and 1(b) show the general morphologies of ZnO NWs and NSs deposited on the Si ⟨100⟩ substrates. The typical length of a nanowire is 50–100 μm while the average diameter is 50–100 nm. The nanostars exhibit a height of 2–2.5 μm whereas each arm of the stars has sharper tips and wider bases. Diameters at the bases and the tips are in the range of 100–120 nm and 50–70 nm, respectively, whereas the length of each arm ranges from 700 nm to 1 μm. The chemical composition of the deposited star shaped products is obtained from the EDS spectra as shown in Fig. 1(c), confirming the grown structure from zinc and oxygen only. The carbon signature is from the carbon film coating on the SEM grid. Figs. 1(d) and 1(e) show a single nanowire and star after removing them from the substrate and dropping the product on the gold coated Si substrates for electrostatic force analysis.

FIG. 1.

SEM image of (a) ZnO NWs; (b) ZnO NSs; (c) EDS profile of NSs; (d) single ZnO nanowire after dropcasting on another Si substrate; and (e) single NS structure.

FIG. 1.

SEM image of (a) ZnO NWs; (b) ZnO NSs; (c) EDS profile of NSs; (d) single ZnO nanowire after dropcasting on another Si substrate; and (e) single NS structure.

Close modal

Growth morphology depends on a number of factors that include temperature, surface diffusion rates, and availability of zinc and oxygen vapors.15 Effects of polar surfaces and temperature gradient seem to change the growth morphology significantly. The c-plane for ZnO is a polar surface with positively charged zinc terminated ⟨0001⟩ and negatively charged oxygen terminated ⟨0001¯⟩ surfaces. The nanostars array grew in a temperature zone in the range of ∼650–700 °C while nanowires were grown at 950 °C. After the substrate is loaded in the reaction chamber, the zinc source reacts with oxygen and forms ZnO droplet on the substrate after being transported by the carrier gas. When the supersaturation increases to a level of formation of ZnO nuclei, continuous vapor supply results in the formation of ZnO nanowires with c-axis as its preferred orientation direction as indicated in Fig. 2(a). As the growth time continues, the gold catalyst sheds off the nanowire. When we increased the distance of the source and the substrate in the reaction chamber, this resulted in lower vapor supply as well as reduced temperature zone. It is also noted that the surface diffusion is relatively lower in the lower temperature region.16 We also see sharpness of tips as compared to their bases which may be due to the lower surface diffusion and limited reactant supply. This resulted in formation of star-shaped structures extending along the six radial directions along the positive c-axis rather than the negative c-axis as illustrated in Fig. 2(b).

FIG. 2.

Schematic growth process illustrating growth of (a) ZnO NWs and (b) NSs.

FIG. 2.

Schematic growth process illustrating growth of (a) ZnO NWs and (b) NSs.

Close modal

Further structural characterization of ZnO nanostructures was performed using TEM measurements. Figure 3(a) shows the selected area electron diffraction pattern (SAED) from a single ZnO nanowire. The SAED pattern reveals that the structures are obtained from a single nanowire and grew in the preferred c-axis. The single crystalline features are also confirmed using TEM images, whereas the distance between the lattice fringes matches well with the ZnO lattice constants as shown in Fig. 3(b).

FIG. 3.

TEM image of a single nanowire grown on a Si (100) substrate: (a) SAED pattern from a ZnO NW, the inset shows a NW of ∼50 nm diameter and (b) lattice fringing pattern from ZnO NW.

FIG. 3.

TEM image of a single nanowire grown on a Si (100) substrate: (a) SAED pattern from a ZnO NW, the inset shows a NW of ∼50 nm diameter and (b) lattice fringing pattern from ZnO NW.

Close modal

A detailed TEM analysis is done for the ZnO stars as shown in Fig. 4. Studies so far report a single arm of various shaped ZnO nanostructures such as flower shaped or tetrapod instead of complete analysis.16,17 We performed a detailed TEM analysis for all arms of a zinc oxide star. Fig. 4(a) shows the TEM image of a single ZnO star; however, Figs. 4(b) and 4(c) show the lattice fringes pattern of a lower and top portion of arm 1 of ZnO NS, respectively. If we see the lower part of arm 1, lattice fringes clearly reveal the single crystallinity of arm 1 and also indicate that the direction of the lattice fringes is perpendicular to the growth direction of the arm (as indicated by the arrow direction in Fig. 4(b)). However, if we see the junction of the two arms of a star in Fig. 4(c) (point 7), we see a mixed fringes pattern in two or more directions as indicated by the two arrow pointers in Fig. 4(c). We also see some ripple like contrasts at the junction most probably coming from the strain occurring at the intersection of the two arms. Figs. 4(d)–4(i) show the SAED pattern of all the arms of a nanostar. Individually all the arms indicate single crystallinity but if we see the overall star, the c-axis for all arms is pointing in different directions depending on the growth direction of the individual arm. These random orientations of the preferred c-axis of nanostar will make it harder to evaluate the electric field of a nanostar as compared to nanowire that will be presented later in this study.

FIG. 4.

TEM images of (a) ZnO star showing corners of six arms marked from (1) to (6) and base of arm 1 marked as (7); (b) HRTEM image showing lattice fringes of the tip of arm 1 (arrow indicates lattice fringes perpendicular to the growth direction of arm 1); (c) HRTEM image showing lattice fringes of the base of arm 1 (arrows indicates mixed lattice fringing pattern at the intersection of arm 1); and (d)–(i) SAED patterns of arms (1)–(6), respectively.

FIG. 4.

TEM images of (a) ZnO star showing corners of six arms marked from (1) to (6) and base of arm 1 marked as (7); (b) HRTEM image showing lattice fringes of the tip of arm 1 (arrow indicates lattice fringes perpendicular to the growth direction of arm 1); (c) HRTEM image showing lattice fringes of the base of arm 1 (arrows indicates mixed lattice fringing pattern at the intersection of arm 1); and (d)–(i) SAED patterns of arms (1)–(6), respectively.

Close modal

In order to investigate the optical properties of ZnO NWs and NSs, photoluminescence measurement is done and a comparison is shown in Fig. 5. As-grown PL spectra show two emission bands, a near band edge (NBE) emission at 381 nm (3.25 eV) and broad green emission at ∼506 nm (2.45 eV) often related to oxygen vacancies. However, the star-shaped structure shows a sharp and strong UV emission whereas visible emission is quenched significantly. It is well known that the NBE peak is due to the recombination of free excitons, whereas impurities and structural defects relating to oxygen vacancies or zinc interstitials are responsible for the deep level defect state.2 Intensity of this visible emission is also dependent on the intrinsic defects variations; improvement in crystal quality often results in the reduction of deep level emission as compared to the NBE peak.18 Our results show that ZnO NSs have lower impurity levels and defect states as compared to the as-grown wires indicating improved quality. It is also reported that the ratio of UV to green emission is dependent on the size.19 We can also draw the conclusion based on our photoluminescence results with thin nanowires showing reduced NBE peak and large defect states as compared to larger diameter nanostars that increasing the diameter of nanowire results in reduced surface states which in turn results in enhanced PL intensity and decreased defect states.

FIG. 5.

Room temperature PL spectra for the ZnO NWs and NSs.

FIG. 5.

Room temperature PL spectra for the ZnO NWs and NSs.

Close modal

Semiconductor nanoparticles such as ZnO with built in spontaneous polarization can generate sufficient electric fields that can be utilized for various applications such as voltage gated ion channels, nanobiosciences and optoelectronic devices.10,11 Considering that wurtzite ZnO has an electrical polarity and normally grows along its c-axis, only the c-component of the spontaneous polarization exists.14 In the case of nanowires, we know that the nanowire consists of a unit cell; however, this is not always the case. If the ZnO crystal consists of a unit cell, its layers along the c-axis should start with Zn atoms and terminate with O or vice versa.20 However in the case when crystal is terminated by both Zn and O layers, this implies that we have extra ions in addition to the unit cells. These extra ions or additional charges will change the spontaneous polarization in the crystal. Effects of these charges become more dominant when the size of crystal becomes smaller, for example, in the case of a quantum dot with nanosized dimensions. Therefore, it is very important to estimate the magnitude of electric field due to spontaneous polarization in nanoscale structures. We have observed from our TEM measurements that ZnO nanowire consists of a single crystal whereas c-axis of a nanostar is oriented in random directions and so are their dipole moments (DM). This is depicted in schematic as shown in Figures 6(a) and 6(b) for the case of NW and NS, respectively. We are utilizing the effects produced due to the spontaneous polarization from ZnO nanowires and nanostars using EFM measurements and mathematical modeling.

FIG. 6.

Schematic illustration of (a) growth direction and c-axis for ZnO NW. (b) Different scenarios for the net dipole moments (DM) of random orientations of a ZnO NS; blue arrows indicates the effect of dipole moment that cancels out each other, whereas red arrows indicates the effective dipole moments that do not cancel each other and contribute to the built-in spontaneous polarization realized from a zinc oxide nanostar.

FIG. 6.

Schematic illustration of (a) growth direction and c-axis for ZnO NW. (b) Different scenarios for the net dipole moments (DM) of random orientations of a ZnO NS; blue arrows indicates the effect of dipole moment that cancels out each other, whereas red arrows indicates the effective dipole moments that do not cancel each other and contribute to the built-in spontaneous polarization realized from a zinc oxide nanostar.

Close modal

Atomic force microscopy was performed on a nanowire and nanostar after the sample was prepared on a gold coated substrate as described earlier for the topology measurement. Figure 7 shows the typical pattern of a ZnO NW and NS with corresponding height profiles for selected (dash lines) regions. Figure 7(a) shows the height profile for gold coated Silicon substrate for reference purposes. It appears that the gold coated Si substrate shown in a 1 μm × 1 μm area scan has a height profile ranging from 4 to 5 nm. ZnO NWs as shown in Fig. 7(b) show the height of NWs lying on the substrate to be 50–100 nm which is close to the diameter of a nanowire whereas NS shows ∼1 μm of the NS height.

FIG. 7.

Typical AFM patterns of (a) a gold coated Si substrate, (b) ZnO NW, and (c) ZnO NS. Height profiles for selected (solid and dashed lined) regions of (d) a gold coated Si substrate, (e) ZnO NW, and (f) ZnO NS.

FIG. 7.

Typical AFM patterns of (a) a gold coated Si substrate, (b) ZnO NW, and (c) ZnO NS. Height profiles for selected (solid and dashed lined) regions of (d) a gold coated Si substrate, (e) ZnO NW, and (f) ZnO NS.

Close modal

The EFM measurements are performed to study the electric field produced by ZnO nanostructures due to its built-in spontaneous polarization effects. To quantify the electrostatic force attraction due to only built-in spontaneous polarization effects, we have used “lift mode” in our study. In this mode, the probe is lifted up to the level that the atomic interaction is significantly smaller than the electrostatic force, where the lift height was found empirically. The time it takes to lift the probe is in the order of microseconds, which is several orders of magnitude longer than the lifetime of the piezo induced charge for ZnO NWs (∼80 ps).21 As a result, the contribution from the strain-induced piezo polarization is ignored. Using the AFM in the lift mode profile, probe displacement is found that is used to indirectly measure the electrostatic force. The resulting electric field from the spontaneous polarization is calculated using the net surface charge of the ZnO NWs and NSs. Figure 8 shows the EFM pattern with the corresponding probe displacement profiles of the NW and NSs. Lift scan is done for all three samples; it is seen that the sample for the Si+Au substrate (Fig. 8(a)) does not attract the probe and their interleave amplitude is in the 50–100 picometer range (Fig. 8(d)) as indicated by the two marked regions in (Fig. 8(a)) which means that the regions are without any electrostatic force. EFM pattern for ZnO NWs is shown in Fig. 8(b) with multiple regions scanned for estimating the interleave amplitude profile for comparison. We first scanned a ZnO NW region 1 (red circled areas) that shows strong attraction for the probe on the order of 10–15 nm range as seen by the large change in interleave amplitude scan (Fig. 8(e)). For comparison, we scanned region 2 with no nanowire that does not show any change in interleave amplitude confirming that the attraction of the probe is from the spontaneous polarization induced electric field from ZnO nanowire. Fig. 8(c) shows the EFM pattern from NSs. We saw attractions from two different places of region 3 of the magnitude of ∼25–40 nm (Fig. 8(f)). We performed the SEM measurements on them and figured out the larger lift height areas are the ones with multiple nanostars clustered together whereas smaller areas indicate a single NW, however both cases clearly exhibit strong electrostatic force attractions.

FIG. 8.

(a) Interleave amplitude scan for Si and gold coated substrates for reference. (b) Interleave amplitude for ZnO NW with red solid circles indicating areas of NWs, subset indicates SEM image of a single NW. (c) Interleave amplitude scan for ZnO NS with small red solid circles indicating single NS areas while bigger red circle indicating multiple stars as shown by the SEM images in subsets. (d)–(f) Displacement profile images for the Si+Au substrate, NW, and NSs, respectively, whereas dashed and solid lines in (a)–(c) indicate the areas scanned for respective displacement profiles.

FIG. 8.

(a) Interleave amplitude scan for Si and gold coated substrates for reference. (b) Interleave amplitude for ZnO NW with red solid circles indicating areas of NWs, subset indicates SEM image of a single NW. (c) Interleave amplitude scan for ZnO NS with small red solid circles indicating single NS areas while bigger red circle indicating multiple stars as shown by the SEM images in subsets. (d)–(f) Displacement profile images for the Si+Au substrate, NW, and NSs, respectively, whereas dashed and solid lines in (a)–(c) indicate the areas scanned for respective displacement profiles.

Close modal

Theoretical calculations on the electrostatic force attraction between the tip and the sample that is used to detect the spontaneous polarization are performed. The electrostatic force (F) acting on the tip is expressed by the electric force law as follows:

F=qE,

where q is the charge and E is the electric field.

Our AFM probe has the spring constant of 2.8 N/m, the corresponding electric force calculated to the 10 nm and 25 nm displacements of nanowires and nanostars are 28 and 70 nN, respectively. Further exposed surface areas for the NW and NS are calculated considering nanowires as a cylinder with surface area of πr2+2πrL. The length of a nanowire is taken as 5 μm whereas the length of a star arm is taken as 1 μm. However for the case of nanostar with six arms, effective surface area with three different cases of orientation dependence is considered. As seen in the TEM image of ZnO stars in Fig. 4, we see random orientations of the preferred c-axis of nanostar. We have simplified the case and assumed three different cases of overall polarization effect. This includes effective dipole moments due to two, three, and four arms of the stars while considering the rest of the dipoles that cancel the effects of each other as depicted schematically in Fig. 6. In that case, effective area for all three different cases of star is considered separately and net electric field is calculated.

In order to estimate the charge, it is important to consider the capacitance between the tip and the surface as well. Using a simplified parallel plate capacitor model, capacitance (C) can be expressed as C=ϵoϵrAd, where ϵo = 8.85 × 10−12 F/m as the vacuum permittivity and ϵr = 8.91 as the permittivity of ZnO and is calculated to be on the order of 6 × 10−15 F for the nanowire. With a 0.3 V electric potential applied during EFM measurements, charge is estimated using Q=CV. The amplitude of the electric field is thus calculated by E=Fq and turns out to be on the order of 107 V/m for nanowire and in the range of 1.5 × 108–4.6 × 108 V/m for nanostars.

Thus, the technique presented using EFM measurements can be utilized efficiently for the quantitative measurement of electric field instead of employing an approximations and assumptions based approach. Since many different ZnO structures have been proposed by the researchers,22,23 knowing the built-in spontaneous polarization and the magnitude of electric field will give important insight into its interaction with other materials and devices and can be employed for applications such as nano-bio interfaces, photonics and optoelectronics.10,11

In summary, we have reported optical, structural, and electrostatic force analysis for complex ZnO structures such as nanowires and nanostars. A route to estimate the built-in spontaneous polarization effect from wurtzite ZnO nanowires and nanostars polar surfaces is presented by measuring the variation in interleave amplitude, attraction of AFM probe, and orientation dependence of ZnO. Photoluminescence measurements indicate enhanced UV emission and quenching of visible peak for ZnO star-like structures as compared to nanowires. TEM results of nanowire show single crystallinity and nanowires growth along the c-axis, whereas the stars indicate c-axis orientation along the length of its arms. The technique presented to estimate spontaneous polarization and electric field provides an important insight for future research and applications in nano-bio interfaces, photonics, and optoelectronics.

This work was partially supported by Grant No. FA9550-15-1-0493 from the Air Force Office of Scientific Research. We would also like to thank Dr. Alan Nicholls from Research Resource Center, University of Illinois at Chicago for his insight and useful discussions on our TEM results.

1.
Z. L.
Wang
,
J. Phys.: Condens. Matter
16
,
R829
(
2004
).
2.
S.
Farid
,
S.
Mukherjee
,
K.
Sarkar
,
M.
Mazouchi
,
M. A.
Stroscio
, and
M.
Dutta
,
Appl. Phys. Lett.
108
,
021106
(
2016
).
3.
A.
Schleife
,
F.
Fuchs
,
J.
Furthmüller
, and
F.
Bechstedt
,
Phys. Rev. B
73
,
245212
(
2006
).
4.
G. C.
Yi
,
C.
Wang
, and
W. I.
Park
,
Semicond. Sci. Technol.
20
,
S22
(
2005
).
5.
K.
Keis
,
L.
Vayssieres
,
S. E.
Lindquist
, and
A.
Hagfeldt
,
Nanostruct. Mater.
12
,
487
(
1999
).
6.
K.
Xu
,
M.
Purahmad
,
K.
Brenneman
,
X.
Meshik
,
S.
Farid
,
S.
Poduri
,
P.
Pratap
 et al, “
Design and applications of nanomaterial-based and biomolecule-based nanodevices and nanosensors
,” in
Design and Applications of Nanomaterials for Sensors
(
Springer
,
Netherlands
,
2014
), pp.
61
97
.
7.
Z. L.
Wang
and
J.
Song
,
Science
312
,
242
(
2006
).
8.
S.
Xu
,
B. J.
Hansen
, and
Z. L.
Wang
,
Nat. Commun.
1
,
93
(
2010
).
9.
X. Y.
Kong
and
Z. L.
Wang
,
Nano Lett.
3
,
1625
(
2003
).
10.
S.
Farid
,
X.
Meshik
,
M.
Choi
,
S.
Mukherjee
,
Y.
Lan
,
D.
Parikh
,
S.
Poduri
,
U.
Baterdene
,
C. E.
Huang
,
Y.
Wang
, and
P.
Burke
,
Biosens. Bioelectron.
71
,
294
(
2015
).
11.
S.
Farid
,
S.
Mukherjee
,
H.
Jung
,
M. A.
Stroscio
, and
M.
Dutta
,
Mater. Res. Express
2
,
025007
(
2015
).
12.
M. H.
Huang
,
Y.
Wu
,
H.
Feick
,
N.
Tran
,
E.
Weber
, and
P.
Yang
,
Adv. Mater.
13
,
113
(
2001
).
14.
M. S.
Choi
,
X.
Meshik
,
S.
Mukherjee
,
S.
Farid
,
S.
Doan
,
L.
Covnot
,
M.
Dutta
, and
M. A.
Stroscio
,
J. Appl. Phys.
118
,
194304
(
2015
).
15.
P. X.
Gao
and
Z. L.
Wang
.
Appl. Phys. Lett.
84
,
2883
(
2004
).
16.
A.
Umar
,
S.
Lee
,
Y. H.
Im
, and
Y. B.
Hahn
,
Nanotechnology
16
,
2462
(
2005
).
17.
V. A. L.
Roy
,
A. B.
Djurišić
,
W. K.
Chan
,
J.
Gao
,
H. F.
Lui
, and
C.
Surya
,
Appl. Phys. Lett.
83
,
141
(
2003
).
18.
I.
Shalish
,
H.
Temkin
, and
V.
Narayanamurti
,
Phys. Rev. B
69
,
245401
(
2004
).
19.
X. L.
Wu
,
G. G.
Siu
,
C. L.
Fu
, and
H. C.
Ong
,
Appl. Phys. Lett.
78
,
2285
(
2001
).
20.
F.
Claeyssens
,
C. L.
Freeman
,
N. L.
Allan
,
Y.
Sun
,
M. N.
Ashfold
, and
J. H.
Harding
,
J. Mater. Chem.
15
,
139
(
2005
).
21.
S. M. A.
Schubert
,
S.
Senz
,
M.
Alexe
,
D.
Hesse
, and
U.
Gösele
,
Appl. Phys. Lett.
92
,
122904
(
2008
).
22.
Y.
Dai
,
Y.
Zhang
,
Q. K.
Li
, and
C. W.
Nan
,
Chem. Phys. Lett.
358
,
83
(
2002
).
23.
Z.
Wang
,
X. F.
Qian
,
J.
Yin
, and
Z. K.
Zhu
,
Langmuir
20
,
3441
(
2004
).