Coupling effects of NaYF4Yb,Er upconversion nanoshells and au ag metallic nanoshells 4

55

Chapter 4 Synthesis and characterization of Au-Ag metallic nanoshells

4.1 Introduction
Hollow metallic structures have been of interests because of potential
applications in drug delivery,
87,88
photothermal therapy,
39,40
and fluorescence
enhancement.
54,89
Optical properties of hollow metallic nanoshells or
dielectric cores/metallic shells may be controlled by the size, shape, and shell
thickness, demonstrating higher sensitivity in localized surface plasmon
resonance (LSPR) as compared with their solid counterparts.
57
The LSPR of
metallic nanoshells may be tuned between the visible and the NIR
wavelengths by changing the dimension of interior cavity and shell
thickness.
90,91
Au nanoshells are one of the most studied due to its good
biocompatibility, thermal and chemical stability.
38,92,93
Since the media such
as water, blood, and tissue are relatively transparent in the NIR wavelengths,
94

the NIR absorbing Au nanoshells find potential biomedical applications.

43

Localized heating for selective destruction of cancer cells was demonstrated in
NIR absorbing Au nanoshells.
4,54,95

When the incident light is in resonance with the plasmon frequency of
the metallic particles, a strong enhancement of local field is produced at the
nanoparticle surface.
96
This local field enhancement around the metallic
particles could increase the fluorescence of nearby fluorophores such as
organic dyes, quantum dots, and UC nanoparticles.
97,98,99
The metallic
nanoshells may exhibit a larger local field enhancement around the particle
surface compared with their solid counterparts. Therefore, metallic nanoshells
56

may be a good candidate to enhance the fluorescence of nearby UC
nanoparticles.
In this thesis, Au-Ag nanoshells were synthesized via galvanic
replacement reaction between Ag templates and HAuCl
4
in toluene-ODE in
the presence of oleylamine.
56
The transformation from Ag templates to Au–
Ag nanoshells in the organic medium was studied. The microstructure,
surface morphology, shell thickness, composition, and optical properties of the

sample collected at different stages of the transformation were investigated.
Further, a size-dependent transformation was demonstrated.

4.2 Ag nanoparticles
Silver nanoparticles were used as a sacrificial template in the galvanic
replacement reaction with HAuCl
4
to form Au-Ag nanoshells. A TEM image
of as-synthesized Ag nanoparticles is shown in Fig. 4.1a. The Ag
nanoparticles consisted of ~85% in decahedral and ~15% in triangular prism
shapes, from counting the particles in 10 TEM images. The HRTEM image
showed pentagonal crossed lines on the particle, further confirming the
decahedral shape with pentagonal cyclic twinning (Fig. 4.1b).
100,101
The
decahedral Ag particles consisted of five rounded edges rather than five
straight edges. Therefore, the decahedral Ag particles lying down on the TEM
substrate appeared like equiaxed shape in the low magnification TEM image
(Fig. 4.1a) since the pentagonal crossed lines on the particle were not clearly
observed.
For non-truncated decahedral metal nanoparticles, the entire surface
was commonly covered by {111} facets,
101,102
which is the densest and lowest
57

surface energy fcc facet. The average size (estimated by random
measurements of ~100 particles from the TEM images) was found to be 43 ± 6
nm for the decahedral shape and 53 ± 9 nm in edge length for the triangular
prism shape. The XRD of the as-synthesized Ag nanoparticles matched well

with the fcc Ag reference [JCPDS file number PDF 4-783], (Fig. 4.1c). The
much higher (111) peak intensity indicated the (111) texture. Figure 4.1d
shows the UV-visible spectrum of Ag nanoparticles stabilized by oleylamine
dispersed in toluene. Their LSPR extinction peak was ~504-nm wavelength.


Fig. 4.1 (a) TEM images of as-synthesized Ag nanoparticles consisting of
decahedral (~43 nm in size) and triangular prism (~53 nm in edge length)
shapes. (b) HRTEM image of a decahedral Ag nanoparticle. (c) XRD of as-
synthesized Ag nanoparticles matched well with the fcc Ag reference (JCPDS
file number PDF 4-783). (d) UV-visible extinction spectrum of the
oleylamine-stabilized Ag nanoparticles in toluene showed an extinction peak
at ~504 nm wavelength.


58

4.3 Microstructure and surface morphology
Au-Ag nanoshells were synthesized via galvanic replacement reaction
between the as-synthesized Ag nanoparticles and HAuCl
4
solution. To study
the changes of microstructure and surface morphology, the samples were
collected at different stages of the reaction (Table 2.1) for TEM and SEM
analyses. Figure 4.2a shows the TEM image of the particles obtained at the
initial stage of the reaction, after the reaction between 15 mol of the HAuCl
4
and the Ag templates. Large voids (light contrasts) in each particle
(decahedral and triangular prism shapes) were observed, indicating Ag solids
(dark contrasts) at the center of the particles had been removed. At this stage,

the decahedral particles appeared like equiaxed particles.
The solids at the center gradually disappeared (increasing light
contrast) with increasing HAuCl
4
(20 mol), leaving behind a continuous dark
contrast at the particle periphery associated with a nanoshell (Fig. 4.2b). The
removal of solid Ag was attributed to Ag oxidation via reduction of HAuCl
4
.
The nanoshells were found in both equiaxed and triangular prism shapes.
Equiaxed nanoshells were likely transformed from decahedral Ag templates,
whereas the triangular prismatic nanoshells were from triangular Ag
nanoprisms. Note that the nanoshells did not collapse or fragment into small
solid particles with addition of excess HAuCl
4
(25 – 30 mol), as shown by
the TEM images in Fig. 4.2c, d. The contrast at the center of the nanoshells
appeared darker and the shells noticeably thicker when HAuCl
4
increased to
50 mol and 75 mol (Fig. 4.2e, f).
Figure 4.3 shows the HRTEM images of the nanoshells. The contrasts
in the HRTEM images indicated the structures could be either a nanoshell
59

(equiaxed or triangular prism shapes) with an interior cavity or a Ag core/Au
shell structure since Ag had a lower scattering contrast. This was further
investigated using EDX line scanning and elemental mapping, discussed in
detail in the following paragraphs.

Fig. 4.2 TEM images of the particles obtained from the reaction of the Ag
templates with (a) 15 mol, (b) 20 mol, (c) 25 mol, (d) 30 mol, (e) 50
mol, and (f) 75 mol of the HAuCl
4
. The scale bars for the images (a – f) are
50 nm.

60


Fig. 4.3 High-resolution TEM images of the nanoshells obtained from the
reaction between the Ag templates and 30 mol of the HAuCl
4
, (a) and (b)
equiaxed shape, (c) and (d) triangular prism shape.


Figure 4.4 shows the compositional line profile across a single particle
by the EDX line scanning analysis. The equiaxed- and triangular prism-
shaped particles (from left to right in Fig. 4.4) were obtained from increasing
the amount of HAuCl
4
. The TEM images (Fig. 4.4a, e) showed part of the
solids (dark contrasts) at the center of both the equiaxed- and triangular prism-
shaped particles were removed at the initial stage of the reaction. Their
compositional line profile showed the Ag intensity reached a maximum value
at the dark contrast regions (solids) and a minimum value at light contrast
regions. This confirmed the removal of Ag solids at the center of the
sacrificial templates. At this stage, the Au intensity was still low across the

single particle compared with that of the Ag (Fig. 4.4a, e), suggesting a low
relative concentration of Au. The relative concentration of the Au in the
61

single particle was 22.5% and 18.7% for the equiaxed- and triangular-shaped
particles, respectively.
A very thin, most likely incomplete Au shell was deposited on the
surface of Ag template when a small amount of HAuCl
4
was added at the
initial stage of the galvanic replacement reaction.
83
The oxidation of Ag likely
continued toward the interior of the particles through the sites at Ag template
surface that was not covered by Au, leading to subsequent large voids at the
center of each particle. When the oxidation was allowed to continue with
increasing amount of HAuCl
4
, most of the Ag solids at the center of the
particles were removed. This was indicated by light contrast at the center of
the particles with the minimum corresponding EDX intensity of Ag (Fig.
4.4b–d, f–h). At this stage, a continuous dark contrast at the particle periphery
with maximum intensities of both Au and Ag was observed. This confirmed
the formation of Au-Ag shells. A previous study reported the deposited Au
layer alloyed with the underlying Ag to form the Au-Ag shells.
83
The surface
energy of Au (1.50 J/m
2
) is higher than that of Ag (1.25 J/m

2
). Hence, the
deposited Au on the outer surface would therefore be more likely diffuse
inward and mix with Ag to decrease the surface energy of Au-Ag system.
103

In our study, the Au-Ag shells did not collapse when excess HAuCl
4

was added. Instead, the shells grew thicker as more Au was formed and
deposited on the surface to form Au-Ag shells, as shown by increased relative
concentration of Au in the shells (Fig. 4.4). The compositional line profile
confirmed the Au-Ag nanoshells did not consist of a solid Ag core for both
equiaxed and triangular prism shapes, since Ag intensity reached a minimum
at the center of the particle. The Au-Ag structure was unlikely a structure of
62

Ag shell/Au shell since the position of the EDX maximum intensities of both
Ag and Au of the shells overlapped each other.


Fig. 4.4 TEM images and compositional line profile of Au and Ag across a
single Au-Ag particle by EDX line scanning analysis. The particles (equiaxed
and triangular prism shapes) from left to right side were obtained with
increasing the amount of HAuCl
4
. The relative concentration of Au in the
equiaxed single particles was (a) 22.5%, (b) 44.1%, (c) 52.6%, (d) 65.5% and
in the triangular prism single particles was (e) 18.7%, (f) 26.2%, (g) 61.6%,
and (h) 68.1%. The scale bars in the TEM images (a – h) are 20 nm.

Further, EDX elemental mappings of Au and Ag of a nanoshell
structure showed that pure Au or Au-rich shell was not observed at the outer
63

surface of the nanoshells (Fig. 4.5). Instead, it was observed that Au mixed
with Ag in the shell to form Au-Ag shell, similar to the previous study.
82
No
EDX of Cl was detected, suggesting the surface of the single particle was
probably free from AgCl contamination.


Fig. 4.5 (a) TEM image of an equiaxed Au-Ag nanoshell and its elemental
mappings of (b) Au, (c) Ag, and (d) Au and Ag. (e) TEM image of a triangular
prismatic Au-Ag nanoshell and its elemental mappings of (f) Au, (g) Ag, and
(h) Au and Ag. The nanoshells were obtained from the reaction between the
Ag templates and excess HAuCl
4
. The scale bars in the images (a – h) are 20
nm.


The surface morphology and surface area were studied using SEM and
BET surface area analysis, respectively. Figure 4.6 shows the SEM images of
the particles obtained at different stages of the reaction. The arrows in Fig.
4.6a shows the pores at the surface of both equiaxed- and triangular-shaped
particles obtained at the initial stage of reaction, after the addition of 15 mol
of the HAuCl

4
. This indicated the oxidation of Ag templates was initiated at
the localized sites at the surface, resulting in pore formation at the surface.
64

Previous studies reported that the deposition of Au would initially
occur on the higher energy facets of cuboctahedral Ag templates such as
{100} and {110} facets, inhibiting oxidation of Ag initiated from these
facets.
104
Oxidation would preferentially start from the {111} facets. For non-
truncated decahedral Ag nanoparticles with entire surfaces enclosed by {111}
facets, the highest energy sites would be located at twin-boundaries, where
defects and lattice distortion commonly accumulated.
105,106
Thus, Au would
be initially deposited on the twin-boundary sites of the decahedral Ag
templates, whereas the oxidation of the Ag templates would locally start from
the {111} facets. In this work, the pores were observed at the {111} facets of
the decahedral Ag templates at the initial stage of the reaction (Appendix Fig.
D.1). This result indicated the oxidation initiated at {111} facets. For
triangular Ag nanoprisms, the two triangular surfaces were commonly
composed of almost {111} crystal facets, whereas their edges were typically
{100}, {110}, or {111} facets.
107
Our results showed the pores were found at
the triangular surface, indicating oxidation of triangular Ag nanoprisms started
at the triangular {111} facets (Fig. 4.6a).
Combining the TEM, EDX line scanning, and SEM results, the
localized oxidation likely continued toward the interior of the particles through

the pores, generating the large voids inside each particle, followed by
formation of interior cavity surrounded by a shell structure. In the galvanic
replacement reaction between Ag templates and HAuCl
4
, AgCl precipitates
would be formed as the interior of the Ag templates was oxidized. Oleylamine
would form a complex with AgCl, which is soluble in the organic medium.
108

The presence of abundant oleylamine in our work would facilitate removal of
65

AgCl precipitates from the interior cavity to the surrounding medium through
the pores.
With increasing the amount of HAuCl
4
(20 mol), more Au would be
deposited on the particles, leading to pore shrinkage and subsequent enclosed
shells as shown by the SEM images (Fig. 4.6b). The outer surface of the
equiaxed Au-Ag nanoshells was rough after formation of the enclosed shells.
On further increasing HAuCl
4
(25 – 75 mol), the outer surface (Fig. 4.6c–f)
became noticeable rougher than that of 20 mol of HAuCl
4
. The surface
roughness was not caused by AgCl precipitates on the particle surface as
discussed earlier. It was reported the presence of surface roughness
significantly increased near-field enhancements on the Au nanoshell surface
and resulted in the redshift of the LSPR peak as compared with the

corresponding smooth nanoshells.
109,110
The SEM images showed the
decahedral-shaped nanoshells were found (Fig. 4.6b). The decahedral-shaped
nanoshells gradually became equiaxed as the surface roughened (Fig. 4.6c–f).
The transformation of the Au-Ag nanoshells from Ag template in the organic
medium is shown in Fig. 4.7

66


Fig. 4.6 SEM images of the particles derived from the reaction between the
Ag templates and (a) 15 mol, (b) 20 mol, (c) 25 mol, (d) 30 mol, (e) 50
mol, and (f) 75 mol of HAuCl
4
. The pores were found at the surface of the
particle obtained at 15 mol of HAuCl
4
as indicated by arrows in image (a),
whereas the no pores were found after the addition of 20 – 75 mol of
HAuCl
4
. The scale bars for images (a – f) are 100 nm.


The BET surface area analysis of the particles with incomplete shells
was performed and compared with the particles with enclosed shells
(Appendix Fig. D.2). The measured BET surface areas were 28.0 and 14.0
m
2

/g for the particles with incomplete shells and the particles with enclosed
shells, respectively. This confirmed the formation of the interior cavity in the
Au–Ag particles. The average pore diameter of the Au–Ag particles with
incomplete shells by the BET measurement was ~17 nm.

67


Fig. 4.7 Illustration of the transformation from Ag templates to Au-Ag
nanoshells. (a) The transformation from the decahedral Ag templates to the
spherical-like or equiaxed Au-Ag nanoshells. (b) The transformation from the
triangular Ag nanoprisms to the triangular prismatic Au-Ag nanoshells. (1) Au
was deposited on the higher energy surface of Ag templates (e.g. {100}, {110}
facets or the twin-boundary sites) as very thin incomplete shells at the initial
stage of galvanic replacement reaction. Thus, the localized oxidation started
at {111} facets of Ag template surfaces those were not covered by Au,
resulting in the formation of pores at the (111) surfaces of the Ag templates
and small voids underneath the surface. The localized oxidation could start on
one to several (111) surfaces of the decahedral Ag templates at this stage. The
deposited Au formed an alloy with the underlying Ag surface to form Au-Ag
alloy shells. (2 – 3) The localized oxidation continued toward the interior of
the particles through the pores, generating larger voids inside the particles, and
(4) followed by the formation of interior cavity surrounded by a shell
structure. (5) The growth of the shells through the Au deposition led to the
pore shrinkage and subsequent enclosed shells.


4.4 Shell thickness
Both TEM and SEM results showed the Au-Ag nanoshells (equiaxed
and triangular prism shapes) did not collapse or transform into Au-rich solid

particles with addition of excess HAuCl
4
after the formation of the enclosed
shells. These results did not follow others’ observations of extensive Ag de-
alloying and pinhole formation with subsequent collapse of the shells.
108

Instead, the shells grew thicker with addition of excess HAuCl
4
. The average
shell thickness (estimated by random measurements from the TEM images)
68

increased with increasing amount of HAuCl
4
(Fig. 4.8). The shell thickness
increased from ~5 nm to ~10 nm and ~5 nm to ~8 nm for the equiaxed and
triangular prismatic Au-Ag nanoshells, respectively.


Fig. 4.8 The average shell thickness of the Au-Ag nanoshells collected after
the reaction between the Ag templates and different amounts of HAuCl
4
in
ODE-toluene medium in the presence of oleylamine at 60
o
C.


4.5 Composition

In galvanic replacement reaction, AgCl precipitates deposited on the
surface of the Ag template would inhibit further reaction between Ag and
HAuCl
4
, preventing the formation of Au-Ag nanoshells. Such a problem was
not observed in this thesis because of the presence of abundant oleylamine,
which facilitated AgCl to form a soluble complex as discussed earlier. This
soluble complex was removed from the samples by centrifugation and
washing using toluene in the presence of oleylamine.
69

Figure 4.9 shows two sets of experimental data of Au composition in
the samples collected at the different stages of reaction. The Au composition
was determined by EDX (Appendix Fig. D.3a) and XPS analysis (Appendix
Fig. D.3b, c), compared with those calculated using three theoretical
conditions (Appendix Table D.1): (i) complete galvanic replacement between
Ag nanoparticles and HAuCl
4
, (ii) no galvanic replacement, and (iii)
combination between complete galvanic replacement and no galvanic
replacement. The theoretical condition (i) (complete galvanic replacement)
assumed the Ag templates are stoichiometrically oxidized by HAuCl
4
when it
is reduced to Au. In galvanic replacement reaction, three moles of Ag are
consumed to form one mol of Au. The theoretical condition (ii) (no galvanic
replacement) assumed the Ag templates were not oxidized. This condition
assumed that all Ag templates underwent alloying with Au to form Au-Ag
alloys. The theoretical condition (iii) (the combination) consisted of the
complete galvanic replacement in a region of 0 15 mol of the HAuCl

4

(region I), followed by no galvanic replacement in a region of 20 75 mol of
the HAuCl
4
(region II) as shown in Fig. 4.9. The absence of Cl in the
experimental results indicated the removal of AgCl from the samples of Au-
Ag particles by centrifugation and washing.
The composition of the Au measured by the EDX was similar to the
results determined by XPS. Both the experimental results by the EDX and
XPS analysis showing a sharp increase of the Au % in the sample of Au-Ag
particles matched with the results calculated using the theoretical condition (i)
(the complete galvanic replacement) in the region I (Fig. 4.9). The sharp
increase of the Au % was attributed to the oxidation of Ag templates by
70

HAuCl
4
and Au formation. When HAuCl
4
increased from 20 mol to 75
mol (the region II), the experimental results of the Au % in the metallic
particles closely matched with the results calculated using the theoretical
condition (ii) (no galvanic replacement).


Fig. 4.9 Two experimental data of the percentage composition of Au in the
bimetallic particles synthesized using different amounts of HAuCl
4
was

determined by ( ) EDX and ( ) XPS analysis and compared with those
calculated using three theoretical conditions: (x) condition (i) (complete
galvanic replacement), (o) condition (ii) (no galvanic replacement), and (+)
condition (iii) (combination). The condition (iii) combined the complete
galvanic replacement in the region I (0 15 mol of the HAuCl
4
), followed by
no galvanic replacement in the region II (20 75 mol of the HAuCl
4
).


Overall, the experimental results in both the region I and II matched
with the theoretical condition (iii) combining the complete galvanic
replacements in the region I, followed by no galvanic replacement in the
region II. These results confirmed the oxidation of Ag templates via the
71

galvanic replacement took place at the initial stages of the reaction (region I),
whereas no extensive oxidation or Ag de-alloying in the Au-Ag shells
occurred in the region II. Thus, the increase of the Au % in the region II was
not as sharp as that in the region I. It may be expected that the Au % in the
sample in the region II would increase as sharply as that in the region I toward
a value of 100% if the extensive Ag de-alloying process in the Au-Ag shells
had continued in the region II. The experimental results in Fig. 4.9 were
consistent with the TEM and SEM results, showing the extensive Ag de-
alloying did not occur to the Au-Ag nanoshells, despite adding excess
HAuCl
4
. The increase of Au % in the region II may be due to Au deposition

from reduction of HAuCl
4
. A previous study showed oleylamine served as a
reducing agent in the synthesis of the Au nanoparticles from HAuCl
4
. In this
thesis, oleylamine could help reduce HAuCl
4
in the region II to form Au.
111
A
previous study reported the alloying of the deposited Au layer with the
underlying Ag surface occurred during the galvanic replacement reaction,
leading to the formation of Au–Ag alloy shells.
83
These homogeneous Au–Ag
alloys are thermodynamically more stable than either pure Ag or Au.
112,113
A
larger driving force would therefore be required to oxidize Ag in the Au–Ag
alloy shells than that for pure Ag template.
114


4.6 Optical properties
In this thesis, the optical properties of the samples collected at different
stages of the reaction of Ag templates and HAuCl
4
solution were investigated
using UV-visible-NIR extinction spectra. At initial stages of the reaction, the

extinction peak of the oleylamine-stabilized particles in toluene red shifted
72

from the visible (~670 nm) to NIR (~840 nm) after addition of 10  25 mol
of the HAuCl
4
(Fig. 4.10a, b). The formation of Au-Ag nanoshells was
observed at these stages of reaction as shown by TEM and SEM results.
When the HAuCl
4
increased from 25 mol to 75 mol, the extinction peak
blue shifted from ~840 nm to ~675 nm. At these stages, the shell thickness of
Au-Ag nanoshells increased with increasing HAuCl
4
(Fig. 4.8). The
progressive change of color of the samples collected at different stages of the
reaction is shown in Fig. 4.10c. The change in the extinction peaks and color
could be associated to the changes of microstructure, surface morphology,
composition, and shell thickness of the particles collected at different stages of
the reaction, which discussed earlier. The formation of nanoshell structures
from their solid templates, the increase of Au %, and surface roughness
contributed to the red shift of the LSPR extinction peak,
82,109,115,116
whereas the
increase of shell thickness would lead to the blue shift.
90

The LSPR extinction of metallic nanoshells may be described as
follows. It was reported LSPR extinction of metallic nanoshells results from
the coupling of the inner shell surface (cavity) and the outer shell surface

(sphere) plasmons over a separation distance essentially given by their shell
thickness.
58
The sphere and cavity plasmons coupled with each other, leading
to a splitting into two new plasmons, the lower energy symmetric or
“bonding” plasmon and the higher energy antisymmetric or “antibonding”
plasmon (Appendix B). The lower energy plasmon strongly interacted with
the incident optical field, whereas the higher energy mode showed weak
interactions. The strength of the coupling between the sphere and cavity
plasmons increased with decreasing the shell thickness (representing the
73

separation distance between the sphere and cavity plasmons), leading to a
larger fractional plasmon shift. This explains the optically active plasmon
resonance shifted to a longer wavelength with decreasing shell thickness.


Fig. 4.10 (a) UV-visible-NIR extinction spectra, (b) extinction peaks, and (c)
photographs of the oleylamine-stabilized particles in toluene were taken after
the reaction between the Ag templates and different amounts of HAuCl
4
in
ODE-toluene medium in the presence of oleylamine at 60
o
C. 


4.7 Size-dependent transformation
To investigate the size effects on the transformation of Ag templates to
Au-Ag nanoshells in the organic medium, Ag nanoparticles with smaller size

were synthesized. Figure 4.11a, b show TEM images of as-synthesized
decahedral Ag nanoparticles collected at different magnifications. The
HRTEM image (the inset of Fig. 4.11b) confirmed the decahedral structure of
as-synthesized Ag nanoparticles. The average size of the decahedral Ag
nanoparticles, estimated by random measurement of 200 particles from the
TEM images, was found to be 20 ± 1 nm. The XRD spectra confirmed the fcc
74

structure of Ag nanoparticles (Fig. 4.11c). UV-visible spectrum of the
oleylamine-stabilized decahedral Ag nanoparticles in toluene showed an
extinction peak at ~407 nm wavelength (Fig. 4.11d). As-synthesized
decahedral Ag nanoparticles (~20 nm) were used as sacrificial templates in
galvanic replacement reaction with HAuCl
4
. The nanostructural
transformation of such small Ag decahedrons via galvanic replacement
reaction with HAuCl
4
in the organic medium was compared with that of ~43-
nm Ag decahedrons which discussed earlier.


Fig. 4.11 (a) and (b) TEM images of as-synthesized decahedral Ag
nanoparticles (~20 nm in size) collected at magnifications of 30000x and
50000x, respectively. The inset of (b) shows HRTEM image of a decahedral
Ag nanoparticle. (c) XRD of as-synthesized decahedral Ag nanoparticles
matched well with the fcc Ag reference (JCPDS file number PDF 4-783). (d)
UV-visible extinction spectrum of the oleylamine-stabilized Ag nanoparticles
in toluene showed an extinction peak at ~407-nm wavelength.
75



Figure 4.12 shows the TEM images of samples obtained from galvanic
replacement reaction between the ~20-nm Ag decahedrons and different
amounts of HAuCl
4
in ODE-toluene medium in the presence of oleylamine at
60
o
C. The TEM images collected at lower magnifications are shown in
Appendix Fig. D.4. The results showed that voids (light contrast) in each
particle were found, indicating part of the Ag solids (dark contrasts) in each
~20-nm decahedrons had been oxidized and removed after addition of 10
mol of HAuCl
4
(Fig. 4.12a). At this stage of reaction, the surface openings
of the particles were found (arrows in Fig. 4.12a), indicating the formation of
the pores at the particle surface, which was similar to that of the ~43-nm Ag
decahedrons.
The Ag solids at the center of each particle were gradually oxidized
and removed with increasing amount of HAuCl
4
(Fig. 4.12b), followed by the
disappearance of the surface openings, subsequently forming an equiaxed shell
structure with interior cavity (Fig. 4.12c, d). The average interior cavity size
and shell thickness of these equiaxed nanoshells (estimated by random
measurement of 200 particles from the TEM images) were 17 ± 2 nm and 3 ±
1 nm, respectively. Atomic diffusion in smaller particles is faster than the
larger ones.
117

The TEM results showed the removal of Ag solids at the center
of the templates was faster for the small Ag decahedrons (~20 nm) compared
with the larger Ag decahedrons (~43 nm).
Addition of excess HAuCl
4
, the nanoshells gradually shrank (Fig.
4.12e–h) and subsequently transformed into solid nanoparticles (Fig. 4.12i).
The relative Au and Ag concentrations in the solid nanoparticles measured by
76

EDX analysis were ~90% and ~10 %, respectively. These results could
indicate extensive Ag de-alloying in small Au-Ag nanoshells (~17-nm interior
cavity/~3-nm shell), transforming Au-Ag nanoshells into Au-rich solid
nanoparticles. Once de-alloying of Ag occurred, vacancies would be
generated from the removal of Ag atoms in the Au-Ag shells, facilitating
further diffusion and de-alloying.
118
Incomplete shells were observed during
the shrinkage of the nanoshells as indicated by arrows in Fig. 4.12e– h. It has
been reported that pinholes in the wall of Au-Ag nanoboxes can be attributed
to the coalescence of the lattice vacancies due to the de-alloying of Ag,
leading to the formation of porous nanoboxes.
83

77


Fig. 4.12 TEM images of the metallic particles obtained from the reaction of
~20-nm decahedral Ag templates with (a) 10 mol, (b) 15 mol, (c) 20 mol,
(d) 25 mol, (e) 30 mol, (f) 40 mol, (g) 50 mol, (h) 60 mol, and (i) 75
mol of HAuCl
4
. The scale bars for the images (a – i) are 20 nm.


Figure 4.13 shows the size of metallic nanostructures measured during
the transformations of the ~20-nm and ~43-nm decahedral Ag templates via
galvanic replacement reaction with different amounts of HAuCl
4
in the
organic medium. At the initial stages (addition of 0 – 25 mol of HAuCl
4
),
the size of both the metallic nanostructures increased as more Au were formed
and deposited on the particles with increasing HAuCl
4
. At these stages, the

~20-nm and ~43-nm Ag templates transformed into ~17-nm interior
78

cavity/~3-nm shell and ~39-nm interior cavity/~6-nm shell Au-Ag particles,
respectively.
When the reaction continued with further increasing amount of
HAuCl
4
(25 – 75 mol), the size (outer diameter) of the small equiaxed Au-
Ag nanoshells (~17-nm interior cavity/~3-nm shell) obtained from the ~20-nm
decahedral Ag templates decreased. A sharp decrease in the size was
observed after addition of 60 – 75 mol of HAuCl
4
as the small nanoshells
were transforming into Au-rich solid nanoparticles as indicated by the TEM
images (Fig. 12h, i). However, the size (outer diameter) of larger equiaxed
Au-Ag nanoshells obtained from the transformation of the ~43-nm decahedral
Ag templates increased with further increasing amount of HAuCl
4
. These Au-
Ag nanoshells (~39-nm interior cavity/~6-nm shell) did not collapse or
transform into Au-rich solid nanoparticles when the reaction was continued
with addition of excess amount of HAuCl
4
. Instead, the size of the nanoshells
increased as the shells grew thicker outward due to the deposition of Au with
increasing HAuCl
4
as discussed earlier. These results indicated size-
dependent transformation from decahedral Ag templates to equiaxed Au-Ag

nanoshells in the organic medium. The size-dependent transformation for the
triangular Ag nanoprisms warrants further work.
The interior cavity size of Au-Ag nanoshells obtained from the
transformations of the ~20-nm and ~43-nm decahedral Ag templates were ~3
nm and ~4 nm smaller, respectively, than the size of their Ag templates. This
could be attributed to alloying between the deposited Au layer and the
underlying Ag in the templates, consistent with a previous report.
82


79


Fig. 4.13 Size of the metallic nanoparticles measured during the
transformations of ~20-nm and ~43-nm decahedral Ag templates via galvanic
replacement reaction with different amounts of HAuCl
4
. The size at zero
mol of HAuCl
4
was the Ag template size.


The theoretical and experimental studies showed that the standard
redox potential of the metal nanoparticles decreases with decreasing particle
size.
119
Further, the redox potential of small metal nanoparticles negatively
shifted in proportional to (1/particle radius) compared with the bulk metal
estimated from the difference in surface free energy between the metal

nanoparticles and their bulk counterparts.
120
In this work, the potential of the
metal nanoparticles (E
nano
) compared with that of the bulk counterparts (E
bulk
)
were expressed as
)1.4(
nF
G
EE bulknano



where ΔG is the surface free energy, n is the number of electrons involved in
the reaction (number of mole of electrons per mole of product), and F is