Preparation and optical properties of silica@Ag-Cu alloy core-shell composite colloids

Article abstract of DOI:10.1016/j.jssc.2007.01.035

The silica@Ag-Cu alloy core-shell composite colloids have been successfully synthesized by an electroless plating approach to explore the possibility of modifying the plasmon resonance at the nanoshell surface by varying the metal nanoshell composition for the first time. The surface plasmon resonance of the composite colloids increases in intensity and shifts towards longer, then shorter wavelengths as the Cu/Ag ratio in the alloy shell is increased. The variations in intensity of the surface plasmon resonance with the Cu/Ag ratio obviously affect the Raman bands of the silica colloid core. The report here may supply a new technique to effectively modify the surface plasmon resonance.

Full text of DOI:10.1016/j.jssc.2007.01.035

ARTICLE IN PRESS
Journal of Solid State Chemistry 180 (2007) 1291–1297
www.elsevier.com/locate/jssc
Preparation and optical properties of silica@Ag–Cu
alloy core-shell composite colloids
Ã
Jianhui Zhang , Huaiyong Liu, Zhenlin Wang, Naiben Ming
National Laboratory of Solid State Microstructures, Department of Physics, Nanjing University, Nanjing 210093, China
Received 5 September 2006; received in revised form 22 January 2007; accepted 24 January 2007
Available online 15 February 2007
Abstract
The silica@Ag–Cu alloy core-shell composite colloids have been successfully synthesized by an electroless plating approach to explore
the possibility of modifying the plasmon resonance at the nanoshell surface by varying the metal nanoshell composition for the first time.
The surface plasmon resonance of the composite colloids increases in intensity and shifts towards longer, then shorter wavelengths as the
Cu/Ag ratio in the alloy shell is increased. The variations in intensity of the surface plasmon resonance with the Cu/Ag ratio obviously
affect the Raman bands of the silica colloid core. The report here may supply a new technique to effectively modify the surface plasmon
resonance.
r 2007 Elsevier Inc. All rights reserved.
Keywords: Ag–Cu alloy
1. Introduction
varying the core/shell ratio. For example, the surface
plasmon resonance of a silica core-gold shell nanoshell can
be varied over hundreds of nanometers in wavelength,
across the visible and into the infrared region of the
spectrum [9,10]. Recently, the studies of the plasmon
energies of the concentric metal nanoshell structure supply
us a new technique to manipulate the surface plasmon
resonance [11]. In the concentric metal nanoshell structure,
when the thickness of the interlayer spacing between the
inner and outer nanoshell is small enough, the coupling
strength and the energy between the plasmons on the inner
and outer shell are strong, and the resulting plasmon
energy differs dramatically from the isolated shell plas-
mons. It is also very interesting to reduce the interlay
spacing to atomic level and replace the inner and outer
shell of the concentric nanoshell with different kinds of
metal, namely, alloy, and control the resulting surface
plasmon by varying the alloying component.
When the wavelength of the incident light is tuned close
to the plasma resonance, conduction electrons in metal
surface are excited into an extended surface electronic
excited state called a surface plasmon resonance (SPR).
SPR has been widely investigated for its extensive
applications in phase-contrast microscopy, spatial light
modulation, enhancement of an electric field, absorption
measurement, radiation force to evanescent waves,
generation of second harmonic waves, energy transfer in
a solar cell, fluorescence spectroscopy, evanescent-wave
holography [1], refractive index or dielectric constant
sensor, biosensor [2], controlling nanoparticle growth [3],
and surface enhanced Raman scattering (SERS) effect
[4–7]. The tuning of SPR in intensity and wavelength is
desired for the applications of SPR [8–10].
The layered nanoparticle with a dielectric core and a
metallic nanoshell, known as a metal nanoshell, provides a
tunable geometry in which the surface plasmon wavelength
at the nanoshell surface can be precisely controlled by
In this report, the Ag–Cu alloy with controllable Cu/Ag
ratio has been successfully deposited on silica colloids in
order to explore the possibility of modifying the SPR by
varying the metal nanoshell composition. The Ag–Cu alloy
is chosen as a coating shell because Ag and Cu have a
strong plasmon absorption band which lies in the visible
Ã
Corresponding author. Fax: +86 25 8359 5535.
E-mail address: zhangjh@nju.edu.cn (J. Zhang).
0022-4596/$ - see front matter r 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jssc.2007.01.035

ARTICLE IN PRESS
J. Zhang et al. / Journal of Solid State Chemistry 180 (2007) 1291–1297
1292
region of the electromagnetic spectrum [12] that can be
easily detected. In addition, the plasmon resonance absorp-
tion band of Ag colloid is stronger and sharper in compa-
rison with Cu colloid, and occurs at shorter wavelengths.
The collective plasmon resonance absorption band of the
Ag–Cu alloy colloids will be easily distinguished from that
of Ag or Cu colloids. This would enable us to explore the
coupling between the plasmons of Ag and Cu, and find a
new technique to effectively modify the SPR.
composite particles were washed with ethanol by centrifu-
gation and ultrasonic dispersions.
Transmission electron microscopy (TEM) was per-
formed on a Hitachi Model H-800 transmission electron
microscope operated at 120 kV. Sample colloids were
placed on a carbon-coated copper grid. Scanning electron
microscopy (SEM) was performed on a LEO1530VP
scanning electron microscope. Energy-dispersive X-ray
analysis (EDX) was performed with a LEO153VP scanning
electron microscope. X-ray photoemission spectroscopy
(XPS) measurements were performed using MgKa radia-
tion (VG-ESCALAB-MK-II). The accuracy of the mea-
sured electron energies was 70.055 eV. Quantification of
the elemental concentrations was accomplished by correct-
ing photoelectron peak intensities for their cross sections.
X-ray diffraction (XRD) was carried out on a Rigaku
D/max-RA X-ray diffraction meter with CuKa radiation
2. Experimental details
The as-prepared silica colloids [13] with average diameter
of 500 nm were seeded with silver nanoparticles at first by
the electroless plating approach [9]. In brief, 20 ml freshly
prepared [Ag(NH3)2]+ ions (0.1 M) solution was added
into 80 ml as-prepared silica colloids suspension containing
1% PVP under stirring. The [Ag(NH3)2]+ ions were
absorbed onto the colloids surface by the negatively-
charged Si–OH groups through electrostatic attraction.
The negatively-charged Si–OH groups are the origin of the
surface charge to protect colloids from aggregation. After
30 min, the colloids were washed with ethanol by centri-
fugation and ultrasonic dispersion to remove the excessive
[Ag(NH3)2]+ ions, and dispersed in ethanol solution
(80 ml). Subsequently, 0.5% KBH4 solution (10 ml) was
added quickly to reduce the [Ag(NH3)2]+ ions absorbed on
the colloids surface to seed the silica colloids. The color of
the solution turned from white to light brown at once.
After 30 min, the colloids were washed with distilled water
to remove the excessive KBH4, and the silica colloids
seeded with Ag nanoparticles were obtained. Methanol
(10 ml), 0.5 g CuSO4 ꢀ 5H2O, 2.5 g NaKC4H4O6 ꢀ 4H2O, and
0.4 g NaOH were dissolved in turn in 20 ml distilled water,
then the solution was added into 30 ml ethanol solution
containing ꢁ0.25 g silica colloids after seeding under
stirring. After 2 min, 1 ml formaldehyde was added to
alloy Cu into Ag nanoparticles for one hour. During the
alloying reaction process, Cu2+ ions were reduced into Cu
through the following reaction:
˚
(l ¼ 1.5418 A). Raman spectra were measured with a Spex
1403 Raman spectrometer. All the spectra were recorded
under identical experimental conditions: 488 nm excitation
line of an argon laser, power 5 mW at the sample, three
accumulations of 120 s counting time each. For the
measurements of XRD, SEM, XPS, and Raman spectra,
the sample thin films were prepared as follows. A few drops
of suspension were spread onto a glass substrate, and dried
under the protection of nitrogen gas. Transmission spectra
were recorded on a U-3410 spectrophotometer. Sample
colloids were diluted with ethanol for the measurements.
Quartz cells with about 1-mm path length were used in the
experiment.
3. Results and discussion
Fig. 1 shows the typical TEM and SEM images of the
silica colloids at different stages of the coating process. As
seen in Fig. 1a, compared with the pure silica colloids (see
inset), the surface of the silica colloids after seeding process
was uniformly covered by Ag nanoparticles with average
diameter of ꢁ7 nm. After one times Cu alloying reaction,
both the coverage and size of the metal nanoparticles
CuSO4 þ 3NaOH þ HCHO Ag nanoparticles
Cu þ Na2SO4 þ HCOONa þ 3H2O:
ꢂ!
NaKC4H4O6
In the reaction, Ag nanoparticles work as the catalyst for
the formation of Cu, and the formation of Cu2O is
suppressed. Even if trace Cu2O appears in solution, most of
them will be reduced further into Cu due to the catalyst
effects of Ag nanoparticles. When the binary metallic
particles size reduces to nanoscale, the diffusion between
the two metals is enhanced greatly, and the spontaneous
alloying appears [14–16]. Here, the catalyst role and
nanoscale size of Ag particles make the diffusing of Cu in
Ag nanoparticles easy, and lead to the formation of Ag–Cu
alloy. The Cu alloying reaction step was repeated to
increase the Cu component in the alloy shell. Finally, the
(ꢁ11 nm) on the silica core increased, and the shell
thickness increased to ꢁ12 nm (Fig. 1b). The second Cu
alloying reaction further increased the size and shell
thickness of the metal nanoparticle to ꢁ15 and ꢁ16 nm
(Fig. 1c), respectively. After the third Cu alloying reaction,
the shell thickness increased to ꢁ 22 nm (Fig. 1d), but was
still too thin to sustain the dissolution of the silica core. As
seen in Fig. 1e, most of the hollow Ag–Cu alloy spheres
obtained by dissolving the silica core using HF acid were
broken and aggregated together. Due to the coalescence of
the seeds during the seeding growth process [9,10], the
metal nanoparticles tended to be heterodisperse. Some

ARTICLE IN PRESS
J. Zhang et al. / Journal of Solid State Chemistry 180 (2007) 1291–1297
1293
Fig. 1. TEM and SEM images of the composite particles synthesized at different stages: (a) silica colloids after Ag nanoparticles seeding, inset shows the
pure silica colloids before seeding; (b)–(d) silica colloids seeded with Ag after one, two, and three times Cu alloying reactions, respectively; (e) sample of (d)
after the dissolution of the silica core.
large particles with diameter of ꢁ30 nm were observed on
the composite colloids after three times Cu alloying
reactions (Fig. 1d).
are exactly the same as the values of zero valent Ag and Cu
[17,18]. This suggests that Ag and Cu in the coated shell of
the composite particles exist in zero valent state. It is worth
noting that the shoulder peaks at 2p3/2 and 2p1/2 of Cu may
arise from the small amount of Cu2O because the peaks at
2p3/2 and 2p1/2 arising from Cu in Cu2O are slightly larger
in binding energies than those originating from element Cu
[17,18]. The impurity such as Cu2O might arise from the
oxidation of Cu in air because that air exposure of the
surface of the samples was unavoidable while transferring
the samples from solution to glass substrate for XPS
measurements [18].
The element composition in the coated shell was further
investigated by EDX. As shown in Fig. 3a, Ag peaks
besides the Si and O peaks arising from the silica core were
observed in the silica colloids seeded with Ag nanoparti-
cles, indicating that Ag was successfully deposited on the
silica colloids. With increasing the shell thickness from
12 nm via 16 to 22 nm by the Cu alloying reactions, Cu
peaks appeared and increased in intensity in comparison
To ascertain the components in the composite particles,
XPS measurements were made with the result shown in
Fig. 2. It is seen that by increasing the Cu alloying reaction
times, both the Ag 3d and Cu 2p XPS peaks shift gradually
to low binding energies. The Cu/Ag atomic ratio calculated
from the ratio of the integral of the Cu 2p and Ag 3d XPS
peak spectra increases from 0 via 0.378, 0.599 to 0.977. The
maximum shift is about 0.3 eV and 0.5 eV for the Ag 3d and
Cu 2p peaks, respectively. The fact that both the Ag 3d and
Cu 2p peaks show a gradual shift rather than jumping to a
lower binding energy with the increase in Cu component
strongly suggests that Cu is alloyed into Ag nanoparticles
rather than simply deposited on the Ag nanoparticles
surface during the Cu alloying reaction. In addition, the
energy differences between the 3d5/2 and 3d3/2 for Ag
(6.0 eV) as well as between 2 p3/2 and 2p1/2 for Cu (19.8 eV)
remain unvaried during all the Cu alloying reactions, and

ARTICLE IN PRESS
J. Zhang et al. / Journal of Solid State Chemistry 180 (2007) 1291–1297
1294
Fig. 3. EDX spectra for the composite particles synthesized at different
stages: (a) silica colloids after Ag nanoparticles seeding; (b)–(d) silica
colloids seeded with Ag after one, two, and three times Cu alloying
reactions, respectively.
the metal crystallite size confirmed by SEM and TEM
images. In order to clearly show the influences of the Cu
alloying reactions on the XRD patterns, the expanded
region (37.7–38.61 2y) around the strongest (1 1 1) diffrac-
tion peak for all the samples are compiled in Fig. 4b.
As seen in the figure, with the increase of the Cu alloying
reaction times, the (1 1 1) peak gradually shifts to high
2y-degree, thus indicating the formation of the Ag–Cu
alloy. These analyses as well as the XPS results clearly
indicate that Ag–Cu alloy is successfully deposited on the
silica core.
Fig. 2. XPS spectra of Ag 3d (a) and Cu 2p (b) regions for the composite
particles synthesized at different stages: (1) silica colloids after Ag
nanoparticles seeding, (2)–(4) silica colloids seeded with Ag after one,
two, and three times Cu alloying reactions, respectively.
The UV–vis absorption spectra were recorded for the
composite particles to analysis the SPR character, which
are shown in Fig. 5. The silica colloids seeded with Ag
nanoparticles have a broad absorption peak around
456 nm (see curve a), which arises from the SPR of Ag
nanoparticles [9,10,12]. After performing one times Cu
alloying reaction, the SPR peak red shifts to 470 nm with a
great increase in intensity. A second peak around 585 nm
originating from the Cu SPR [12,19–21] appears as a
shoulder. By increasing the Cu alloying reaction times and
thus the Cu/Ag ratio, both the SPR absorption peaks of Ag
and Cu become increased in intensity, narrowed, blue
shifted, and tend to merge into a single absorption peak.
According to the previous reports [12,19,20], the formation
of one single absorption suggests that the Cu–Ag alloy is
formed during the Cu alloying reactions. Furthermore, if
Cu–Ag alloy is not formed and element Cu is simply
with the Ag peaks, and the Cu/Ag atomic ratio increases
from 0.38 via 0.65 to 1.07 (see Fig. 3c–d). These results are
similar to those obtained by the XPS spectra.
The XRD was performed to ascertain the crystal
structure of the coated shell on the composite particles.
As seen in Fig. 4a, the diffraction peaks of Ag for the silica
colloids after seeding are very weak due to the low mol
ratio of Ag/SiO2 (see curve 1). The diffraction of Ag is
barely discernable under the strong scattering background
from the amorphous silica colloids that exhibit a broad
peak in the small diffraction angle region. As shown in
curve 2–4, the Cu alloying reactions do not cause any
diffraction peaks of element or compound of Cu, but
enhance the Ag diffraction peaks and make these peaks
gradually discernable and strong. The increase in intensity
of the diffraction peaks should arise from the increase in

ARTICLE IN PRESS
J. Zhang et al. / Journal of Solid State Chemistry 180 (2007) 1291–1297
1295
Fig. 4. XRD patterns (a) of the composite particles synthesized at different stages: (1) silica colloids after Ag nanoparticles seeding, (2)–(4) silica colloids
seeded with Ag after one, two, and three times Cu alloying reactions, respectively. The corresponding expanded region (37.7–38.61 2y) around the (1 1 1)
diffraction peak for all the samples are compiled in (b).
deposited on the Ag nanoparticles in the coating process,
with increased Cu/Ag ratio, the Ag SPR band should be
depressed but not shift, and the Cu SPR band should
increase in intensity and red shift [12]. This is ruled out by
the experimental results.
Since most of the nanoparticles deposited on the silica
core are smaller than 20 nm in our samples, the absorption
is reasonably described by Mie scattering theory in the
electric dipole approximation [12,19,22] and is given by
than e2(Ag) beyond 200 nm in the wavelength of interest
[23]. However, as shown in Fig. 5, the SPR bands are
enhanced and narrowed with the alloying Cu, and begin to
blue shift and merge into a single band after two times of
Cu alloying reactions. Although the increases of particles’
size and coverage caused by the Cu alloying can account
for the increased intensity and the narrowing effect of the
SPR bands [9,10], the large blue shift of the SPR bands is
still an open question.
As reported by Magruder et al. [19] with increasing the
degree of alloying, the SPR band of the Ag–Cu composite
particles doped silica film red shifts and decreases in
intensity. However, with increasing Cu concentration in the
Ag–Cu alloy particles, the SPR blue shifts and decreases in
intensity [20]. In the Ag–Cu mixture nanoclusters doped
silica glass films [21], with the increase in Cu component,
the blue shift and decrease in intensity of both Cu and Ag
SPR bands have also been observed. All the above blue
shifts have been ascribed to an interaction between Ag and
Cu. Considering the report of Prodan et al. [11], the blue
shifts of the SPR bands induced by the second and third
times Cu alloying reactions can be ascribed to the SPR
coupling between Cu and Ag. Here the abnormal increase
in intensity of the SPR bands caused by the Cu alloying
reactions in comparison to the reported results [19–21]
should mainly originate from the increase of the shell
thickness [9,10]. According to the above results and
discussion, the goal coupling between the plasmons on
the inner and outer shell at atomic level in the concentric
metal nanoshell structure has been realized here to a
2
a ¼ 18pn3dpꢀ2=l½ðꢀ1 þ 2nd2Þ þ ꢀ22ꢃ,
(1)
where a is the absorption coefficient, e(l) ¼ e1+ie2 is the
dielectric constant of the metal, p is the volume fraction of
the metal particles and nd is the index of refraction of the
dielectric host. The absorption is expected to exhibit a peak
at the surface plasmon resonance frequency for which the
condition e1+2n2d ¼ 0 is met. Beyond 354 nm in wave-
length, the real part of the dielectric constants of both Ag
and Cu decreases with increasing wavelength, and e1(Cu) is
larger than e1(Ag) [23]. The alloying of Cu into Ag
nanoparticles will increase e1 and make the SPR peak red
shifted. The increase of particles size due to Cu alloying
reaction may also lead to a small red shift (o5%) of the
position of the SPR [22]. These, as well as the increase of
the shell coverage and thickness [9,10], collectively lead to
the red shift of the Ag SPR absorption peak caused by the
first times Cu alloying reaction as shown in Fig. 5b.
From Eq. (1), the intensity of the SPR is inversely related
to e2; the alloying Cu into Ag nanoparticles will increase e2
and thus decrease the SPR intensity because e2(Cu) is larger

ARTICLE IN PRESS
J. Zhang et al. / Journal of Solid State Chemistry 180 (2007) 1291–1297
1296
on the wavelength scale, at approximately 499.2 and
513.0 nm, respectively. So the SPR peaks located at 493.6
and 500.5 nm (calculated from the equation (lexcitation
line+lRaman scattering)/2 [6]) are expected to give maximum
contribution to the enhancement for the Raman bands at
460 and 1000 cmꢂ1, respectively. As shown in Fig. 5, the
SPR bands position in all the composite particles is closed
to the expected SPR bands for the maximum enhancement.
Furthermore, the SPR peaks become enhanced in intensity
by the Cu alloying reaction. These motivate us to
investigate whether the variations of the SPR intensity
with Cu component can affect the Raman bands of the
silica core, namely, whether the SERS effect can be found
in the composite colloids system itself.
As shown in Fig. 6a, two broad Raman bands around
460 and 1000 cmꢂ1 were observed for the thin film made of
the silica colloids seeded with Ag nanoparticles. The
alloying Cu into Ag nanoparticles by the first time
Cu alloying reaction enhances the SPR intensity, and
obviously enhances the two Raman bands (curve b).
Further increasing the SPR intensity by the second time
Cu alloying reaction not only further enhances the two
Raman bands, but also leads to a new Raman band around
790 cmꢂ1 (Fig. 6c). The new band arises from a Si–O
stretching vibration that contains considerable motion of
the silicon atom [26]. The third time Cu alloying reaction
continuously enhances the SPR band and the Raman
bands around 460 and 790 cmꢂ1, but greatly blue shifts the
SPR band to 431 nm and weakens the Raman band around
1000 cmꢂ1 (Fig. 6d), which can be ascribed to the large
deviation between the actual SPR position and the
expected SPR position (500.5 nm) for the maximum
enhancement. It should be noted that the above investiga-
tion about the SERS effect of the composite colloids
Fig. 5. Absorption spectra of the composite particles synthesized at
different stages: (a) silica colloids after Ag nanoparticles seeding; (b)–(d)
silica colloids seeded with Ag after one, two, and three times Cu alloying
reactions, respectively.
certain degree by coating the silica colloids with the Cu–Ag
alloy. The collective plasmon resonance of Cu–Ag alloy
differs dramatically from that of Ag or Cu, and can be
easily tuned in wavelength by changing the Cu/Ag ratio.
Furthermore, the unique core-shell structure provides us a
new technique to modify the SPR intensity of the Cu–Ag
alloy by varying the alloy shell thickness.
As seen from Fig. 4, the silica colloids core used here is
amorphous silica which usually has two broad Raman
bands [24–26]. One is located around 460 cmꢂ1, and is
attributed to the symmetric bending motion of the Si–O–Si
linkages. The other band is located around 1000 cmꢂ1, and
originates from the asymmetric stretching of the bridging
oxygen. Currently, local electromagnetic field enhanced by
the excitation of surface plasmons at the metal surface is
widely believed to be responsible for the surface enhanced
Raman scattering (SERS) effect [5–8]. An enhancement of
SERS effect can be maximized when the SPR wavelength is
equal to the average value of the excitation and Raman
scattering wavelengths [6]. With the 488 nm excitation line
of an argon laser used in our experiments, the Stokes bands
around 460 and 1000 cmꢂ1of amorphous silica are located,
Fig. 6. Raman spectra of the thin films made of the composite particles
synthesized at different stages: (a) silica colloids after Ag nanoparticles
seeding; (b)–(d) silica colloids seeded with Ag after one, two, and three
times Cu alloying reactions, respectively. The spectra are vertically shifted
for clarity.

ARTICLE IN PRESS
J. Zhang et al. / Journal of Solid State Chemistry 180 (2007) 1291–1297
1297
system itself is qualitative because there might be a high
proportion of re-absorption within the alloy shell sur-
rounding the silica core.
[7] J.B. Jackson, S.L. Westcott, L.R. Hirsch, J.L. West, N.J. Halas,
Appl. Phys. Lett. 82 (2003) 257.
[8] C. Jiang, S. Markutsys, V.V. Tsukruk, Langmuir 20 (2004) 882.
[9] J. Zhang, J. Liu, S. Wang, P. Zhan, Z. Wang, N. Ming, Adv. Funct.
Mater. 14 (2004) 1089.
4. Conclusion
[10] S.J. Oldenburg, R.D. Averitt, S.L. Westcott, N.J. Halas, Chem. Phys.
Lett. 288 (1998) 243.
In summary, the possibility of modifying the plasmon
resonance at the nanoshell surface through varying the
metal nanoshell composition has been investigated by
fabricating the silica@Ag–Cu alloy core-shell composite
colloids for the first time. It was found that varying the Cu/
Ag ratio of the alloy nanoshell has obvious influences on
the surface plasmon resonance of the composite colloids
and the Raman bands of the amorphous silica core.
[11] E. Prodan, C. Radloff, N.J. Halas, P. Nordlander, Science 302 (2003)
419.
[12] P. Mulvaney, Langmuir 12 (1996) 788.
[13] J.H. Zhang, P. Zhan, Z.L. Wang, W.Y. Zhang, N.B. Ming, J. Mater.
Res. 18 (2003) 649.
[14] H. Yasuda, H. Mori, M. Komatsu, K. Takeda, J. Appl. Phys. 73
(1993) 1100.
[15] H. Yasuda, H. Mori, Phys. Rev. Lett. 69 (1992) 3747.
[16] T. Shibata, B.A. Bunker, Z. Zhang, D. Meisel, C.F. Vardeman, J.D.
Gezelter, J. Am. Chem. Soc. 124 (2002) 11989.
[17] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, in: J. Chastain
Jr., R.C. King (Eds.), Handbook of X-ray Photoelectron Spectro-
scopy, Physical Electronics, Inc., Eden Prairie, MN, 1995, pp. 86–120
Acknowledgments
J.H. Zhang acknowledges Project 10404013 supported
by NSFC, and H. Dong, P. Zhan, and X.N. Zhao for their
help in the SEM measurements.
(Chapter II).
´
[18] A. Molnar, I. Bertoti, J. Szepvolgyi, G. Mulas, G. Cocco, J. Phys.
´ ´ ´
¨
Chem. B 102 (1998) 9258.
[19] R.H. Magruder III, J.E. Wittig, R.A. Zuhr, J. Non. Cryst. Solids 163
(1993) 162.
References
[20] R.H. Magruder III Jr., D.H. Osborne, R.A. Zuhr, J. Non. Cryst.
Solids 176 (1994) 299.
[1] S. Maruo, O. Nakamura, S. Kawata, Appl. Opt. 36 (1997) 2343.
[2] E. Fujii, T. Koike, K. Nakamura, S. Sasaki, K. Kurihara, D. Citterio,
Y. Iwasaki, O. Niwa, K. Suzuki, Anal. Chem. 74 (2002) 6106.
[21] G. De Tapfer L., M. Catalano, G. Battaglin, F. Caccavale, F.
Gonella, P. Mazzoldi, R.F. Haglud Jr., Appl. Phys. Lett. 68 (1996)
3820.
[3] R. Jin, Y.C. Gao, E. Hao, G.S. Me
Nature 425 (2003) 487.
´
traux, G.C. Schatz, C.A. Mirkin,
[22] U. Kreibig, L. Genzel, Surf. Sci. 156 (1985) 156.
[23] P.B. Johnson, R.W. Christy, Phys. Rev. B 6 (1972) 4370.
[24] P. Colomban, C. Truong, Raman Spectrosc. 35 (2004) 195.
[25] P. Gangopadhyay, R. Kesavamoorthy, K.G.M. Nair, R. Dhanda-
pani, J. Appl. Phys. 88 (2000) 4975.
[4] M. Moskovits, Rev. Mod. Phys. 57 (1985) 783.
[5] K. Kneipp, Y. Wang, H. Kneipp, L.T. Perelman, I. Itzkan, R.R.
Dasari, M.S. Feld, Phys. Rev. Lett. 78 (1997) 1668.
[6] N. Felidj, J. Aubard, G. Levi, J.R. Krenn, M. Salerno, G. Schider, B.
Lamprecht, A. Leitner, F.R. Aussenegg, Phys. Rev. B 65 (2002) 075419.
´
´
[26] C.H. Polsky, K.H. Smith, G.H. Wolf, J. Non. Cryst. Solids 248
(1999) 159.