Size-controlled synthesis of nanoparticles. 1. silver-only aqueous suspensions via hydrogen reduction

Article abstract of DOI:10.1021/jp047565s

A method for size-controlled synthesis of chemically clean silver nanoparticles is proposed and described. The synthesis is based on the reduction of silver(I) oxide by hydrogen gas in water and offers the advantages of being easily scalable and able to produce essentially naked particles still exhibiting long-term stability with the metal surface readily accessible to various chemical modifications. The method allows synthesis of particles of any diameter between 15 and 200 nm simply by varying the reaction time. The addition of filtration and centrifugation techniques produces suspensions of narrow size distribution and relatively high, 10¹²10¹⁴ particles/mL, concentration. The addition of various complexating agents potentially allows control of nanoparticle shape. The nanoparticles were characterized by UV - vis spectroscopy, electron microscopy, and electron and X-ray diffraction techniques.

Full text of DOI:10.1021/jp047565s

13948
J. Phys. Chem. B 2004, 108, 13948-13956
Size-Controlled Synthesis of Nanoparticles. 1. “Silver-Only” Aqueous Suspensions via
Hydrogen Reduction
David D. Evanoff, Jr. and George Chumanov*
Department of Chemistry, Clemson UniVersity, Clemson, South Carolina 29634
ReceiVed: June 5, 2004; In Final Form: July 13, 2004
A method for size-controlled synthesis of chemically clean silver nanoparticles is proposed and described.
The synthesis is based on the reduction of silver(I) oxide by hydrogen gas in water and offers the advantages
of being easily scalable and able to produce essentially naked particles still exhibiting long-term stability
with the metal surface readily accessible to various chemical modifications. The method allows synthesis of
particles of any diameter between 15 and 200 nm simply by varying the reaction time. The addition of filtration
12
and centrifugation techniques produces suspensions of narrow size distribution and relatively high, 10 -
14
1
0 particles/mL, concentration. The addition of various complexating agents potentially allows control of
nanoparticle shape. The nanoparticles were characterized by UV-vis spectroscopy, electron microscopy, and
electron and X-ray diffraction techniques.
Introduction
as reducing agents, to name a few. The monovalent silver ion
1
8
has a large reduction potential, 0.7996 V, and can be easily
reduced by many different organic and inorganic compounds.
While many methods do not suggest the use of additional
protecting agent, those touting long-term stability often employ
In the past twenty years, the use of noble metal nanoparticles
in various fields of research has increased dramatically. This is
undoubtedly due to not only the bulk properties of noble metals
such as chemical stability, electrical conductivity, and in some
cases high catalytic activity, but also due to the unique optical,
electrical, catalytic, and magnetic properties that are a conse-
quence of nanometer dimensions. Silver nanoparticles, in
particular, are of great interest because of their ability to
efficiently interact with light by virtue of plasmon resonances,
which are the collective oscillations of the conduction electrons
in the metal. Silver nanoparticles certainly have the potential
to be the building blocks of future photonic and plasmonic
devices as the field of nanotechnology matures.
A requirement for this progress is the development of a
method that can reliably produce large quantities of silver
nanoparticles of narrow and controlled size distribution. Un-
doubtedly, different applications will require particles of various
dimensions and the ability to tailor surface chemistry. The ideal
method of synthesis should yield essentially naked particles so
that different surface functionalities can be readily introduced
for template-assisted or self-assembly based fabrication methods
for nanoscale devices.
5
,19
20
4
surfactants, such as NaAOT, CTAB, and Daxad 19, sodium
1
6
6,8
10
14
citrate, various thiols, gum Arabic, PEG, and poly(N-
to protect the particles from aggrega-
tion. Stabilizers such as NaAOT,
1
2,13
vinyl-2-pyrrolidone)
1
9,22
20
21
CTAB, and PVP have
also been effectively used to direct the aspect ratio of the
particles producing plates and rods and other shapes, although
a recent report indicates that such stabilizers are not necessary
to produce rod particles with very high aspect ratios.15 In
general, all of the above methods produce mainly small particles,
although very careful control of the reducing environment and
stabilizer concentration makes it possible to synthesize larger
particles as well.
Another category entails impinging high-intensity electro-
magnetic radiation on a system to either ablate particles from a
2
3
metal target surface or to stimulate the photoreduction of a
24,25
silver salt.
A recent report suggests that irradiation with light
allows further manipulation of particle size and shape through
the selective plasmon excitation after the initial reduction step.26
Silver particles can also be prepared by vacuum deposition of
the metal. Particle size and shape can be controlled by varying
Over the years, many methods for the synthesis of silver
nanoparticles have been implemented. These methods can be
divided into several categories. The first is various modifications
to either the Creighton method, which employs the reduction
of silver nitrate by sodium borohydride, or the Lee-Meisel
method, which is generally considered to be the reduction of
silver nitrate with sodium citrate, although several methods were
2
7
the geometry between the source and substrate or by surface
treatment of the substrate, either through chemical modifica-
2
8
29
1
tions or by patterning with polymeric spheres, followed by
thermal or solvent annealing. The latter method, known as
nanosphere lithography, has proven to be a versatile approach
to the fabrication of monodispersed periodic arrays of metallic
nanoparticles but seems impractical for large-scale production.
The electrolysis of silver ion solutions at an electrode surface
2
1-15
originally suggested in their paper. AgNO3
is the most
2
common source of silver ions although Ag2SO4, silver 2-ethyl-
1
6
17
hexonate, and silver perchlorate have also been used.
3
0
1
,2,5-8,17
1,2,15
4,10
16
followed by monolayer capping or stripping with ultrasonic
NaBH4,
sodium citrate,
ascorbic acid, DMSO,
potassium bitartate, ethanol,
31
9
,11
12
cavitation also results in the suspension of silver nanoparticles.
Other methods include the reduction of silver ion solutions by
hydrazine dihydrochloride,
pyridine, DMF,1 and poly(ethylene glycol)14 have all been used
3
32
a pulsed electron beam, solutions trapped in supercritical fluid-
3
3
induced, reverse micelles, or by the rapid expansion of a
*
Author to whom correspondence should be addressed. E-mail:
+
gchumak@clemson.edu.
supercritical fluid containing Ag solution into a reducing
1
0.1021/jp047565s CCC: $27.50 © 2004 American Chemical Society
Published on Web 08/24/2004

Size-Controlled Synthesis of Nanoparticles. 1
J. Phys. Chem. B, Vol. 108, No. 37, 2004 13949
environment.34 Silver particles can even be grown by the uptake
of silver ions by living alfalfa plants.35
None of the above methods appear suitable for general utility.
While most of the methods offer an acceptably narrow size
distribution, they do not offer much variability in the size that
can be produced. In the methods that can control particle size
over a wide range, scalability and “chemical cleanness” of both
the nanoparticle surface and the suspension can be a problem.
In this paper, we propose a general synthetic method that
produces silver nanoparticles in water without any extraneous
chemical species. The methods offer the advantages of being
easily scalable and able to produce essentially naked particles
still exhibiting long-term stability with the metal surface readily
accessible to various chemical modifications. Importantly, the
method allows synthesis of particles of any diameter between
15 and 200 nm simply by varying the reaction time. The addition
of filtration and centrifugation techniques produces suspensions
1
2
14
of narrow size distribution and relatively high, 10 -10
particles/mL, concentration.
Figure 1. Apparatus for silver nanoparticle synthesis.
The proposed synthetic method is based on the reduction of
silver(I) oxide by hydrogen gas in water. Hydrogen gas is a
common reducing agent. It was suggested by Lee and Meisel
X-ray diffraction patterns were collected using a Scintag 3100
PXRD with a Cu KR source.
2
to use hydrogen gas to reduce silver ions in a boiling solution
and is a very common reducing agent for the preparation of
silver nanoparticle composites.36-39 Hydrogen is an excellent
choice as a reducing agent because it imparts no residual
chemical impact on a system. Ag2O has been known since the
Results and Discussion
Synthetic Method. Silver nanoparticles can be synthesized
by the reduction of silver(I) oxide by H2(g) in ultrapure water.
Simple bubbling of the gas through the saturated silver oxide
solution at elevated temperature results in the formation of a
colloidal suspension. This procedure, however, is not very
efficient and does not provide required experimental control for
the synthesis of well-defined nanoparticles. A slow-growth
approach was developed, in which silver nanoparticles are
synthesized in saturated silver oxide solution at elevated
temperature in equilibrium with hydrogen gas at elevated
pressure. This method is quite efficient as hydrogen is consumed
only as needed for the metal reduction. Varying the temperature
controls the rate of the reaction and the overall time required
to grow particles to specific dimensions. Because hydrogen,
water, and silver oxide are the only components used in the
reaction, no other chemicals are present in the final colloidal
suspension. The method can be used for large-scale production
as the amount of nanoparticles produced is limited only by the
size of the reaction vessel.
A special reactor was constructed that consists of a 5000-
mL Pyrex round-bottomed flask with an attached reverse
condenser, thermometer, and spout (Figure 1). A hydrogen line
with two backflow preventers to ensure that no water vapor
can enter the gas cylinder and to limit a maximum allowable
gas pressure in the vessel is attached to the condenser. The gas
vent allows the vessel to be flushed prior to pressurization to
remove any residual air from the space above the liquid. The
vessel is filled with 3 L of ultrapure water and 2-3 g of Ag2O
is added. The mixture is shaken well before connecting to the
condenser and hydrogen line. A heating mantle is placed under
the vessel and its contents heated and maintained at 70 °C
without further agitation. Once the temperature has equilibrated,
the vessel is flushed with H2 and pressurized to 10 psi.
Hazard: overpressurizing or using a defective vessel may result
in explosion and cause severe injuries. Within 10-15 s of the
initial pressurization, the contents of the vessel become faint
yellow in color indicating the formation of metallic silver
particles 10-15 nm in diameter. As the reaction progresses,
aliquots can be taken through the spout. Removing the aliquots
from the reducing environment stops further growth of the
1
960s to undergo a slow decomposition to metallic silver in a
4
0,41
basic solution.
Although generally considered insoluble in
18
water, Ag2O has a low solubility of 0.053 g/L at 80 °C, which
is sufficient for nanoparticle synthesis.
Experimental Section
Chemicals. Ag2O (99.99%) was purchased from Alfa Aesar.
Hydrogen gas (99.9999%) was purchased from National Weld-
ers. Hazard: hydrogen gas mixed with air in concentrations
larger than 4% is explosive and must be handled in well-
ventilated areas. A 40 000 MW poly(4-vinylpyridine) and
AgNO3 were purchased from Aldrich. Ag2SO4 was purchased
from Fluka. NaBH4, HF (49%), and Reagent alcohol (HPLC
Grade) were purchased from Fisher Scientific. Hazard: HF
solutions and their fumes are highly toxic and corrosive and
must be handled in fume hoods with proper personal protection
equipment. (Tridecafluoro-1,1,2,2-tetrahydrooctyl)trichlorosilane
was purchased from Gelest, Inc. All chemicals were used
without further purification. Water with a resistivity greater than
1
8 MΩ‚cm was acquired from a four-bowl Millipure Milli-Q
system.
Methods. The vessel used for particle synthesis was a 5000-
mL Pyrex round-bottomed flask purchased from Ace Glass, Inc.
Filtration employed 90-cm Osmonics nylon filter membranes
purchased from Fisher Scientific. Centrifugation was done using
a Beckman J2-HS centrifuge with a JA-10 centrifuge head.
Extinction spectra were collected using a Shimadzu UV-2501PC
spectrometer. All spectra were processed and figures prepared
using Spectra-Solve for Windows (LasTek Pty. Ltd.). Particle
diameters were measured using a Hitachi S-4700 FE-SEM
operating between 8 and 12 keV. The substrates for SEM were
ITO glass slides (R ) 15 Ω) purchased from Delta Technolo-
gies. Electron diffraction patterns and corresponding images
were collected on a Hitachi HD-2000 STEM operating at 200
keV, with ED patterns being collected in a mode that uses a
quasi-parallel beam. Formvar-coated gold and copper grids with
type B carbon were purchased from Ted Pella, Inc. Powder

13950 J. Phys. Chem. B, Vol. 108, No. 37, 2004
Evanoff and Chumanov
Figure 2. Extinction spectra of silver nanoparticle suspensions taken
as aliquots from a single reaction in timed intervals.
particles. Likewise, the reaction can be stopped at any time
simply by releasing the gas from the vessel. In this way, control
of the particle diameter is purely a function of total reaction
time.
Figure 3. (A) Optical density of the silver nanoparticle suspension as
a function of the spectral position of the dipole component of the
plasmon resonance as observed in a typical reaction. (B) Dependence
of the spectral position of the dipole component of the plasmon
resonance on particle diameter.
Variations in the overall amount of silver oxide, hydrogen
pressure, volume of water, and temperature have little effect
on the final product but can significantly alter the rate of the
reaction. The reaction proceeds in a more consistent way if a
significant gas volume is left above the liquid in the vessel and
a large excess of silver oxide is used to maintain saturation.
An elevated temperature was selected to increase Ag2O solubil-
ity allowing rapid seed formation at the initial stage of the
reaction leading to high particle concentration in the final
suspension and relatively fast particle growth. Consistency in
temperature, H2 pressure, and the Ag2O/water ratio allows a
fairly accurate predication of the time required to grow particles
to a selected diameter.
factors: particle concentration and the extinction cross sections
of individual particles. Experimental determination of extinction,
scattering, and absorption cross sections for different size silver
nanoparticles synthesized by this method is presented in the
subsequent article. The measured extinction cross sections
support the statement that the concentration of particles in
suspension remains constant as the reaction progresses. Con-
sequently, Figure 2 clearly indicates that large particles interact
with light to a much greater extent than small ones by virtue of
a larger number of electrons participating in the plasmon
resonance. As long as the particles are not large enough to
exhibit strong retardation effects, the ability of particles to
interact with light scales as the particle volume (total number
of electrons). This trend can be observed in Figure 3A, in which
the optical density changes rapidly while peak position changes
slowly. The position of the kink on the curve (ca. 450 nm)
indicates the particle size at which phase retardation of incident
radiation becomes important resulting in a slower increase in
extinction during further particle growth. The effect of phase
retardation is also evident in Figure 3B at a particle diameter
of ca. 60 nm as the position of the dipole component of the
plasmon resonance begins to shift to longer wavelengths with
increasing particle size. The slower increase of extinction after
the plasmon resonance maximum begins to shift indicates that,
when retardation becomes significant, not all electrons in a
particle contribute equally to the resonance.
UV-Vis Spectroscopy. Extinction spectra of aliquots col-
lected in the course of a single reaction are shown in Figure 2.
Each curve corresponds to an aliquot taken in 1-10-min
intervals so that the total reaction time covered in the figure is
approximately 3 h representing 25 different particle sizes. The
spectra have not been normalized and the apparent consistent
evolution of the extinction maximum indicates the gradual
increase of particle size in a suspension of constant particle
concentration. In other words, as the reaction proceeds, no new
particles are formedsonly further growth of already existing
particles. The evolution of the maximum in the extinction spectra
starts from the initial peak position around 400 nm representing
the dipole component of the plasmon resonance of small silver
particles. As the particle size increases, the intensity of the
resonance increases and its position shifts to longer wavelengths.
Suspensions of small particles appear yellow and totally
transparent as the absorption contribution of the dipole resonance
dominates the extinction spectrum. For larger particles, increased
opalescence can be observed, as the resonant scattering contri-
bution of the dipole resonance becomes a dominating spectral
In addition to the dipole component of the plasmon resonance,
the extinction spectrum of small particles reveals two features:
a minimum around 320 nm caused by an interband transition
4
2
in the metal that causes damping of plasmon oscillations in
this spectral region and a shoulder at 350 nm. This shoulder,
although its exact nature is not fully understood, is a part of
plasmon resonance and its appearance is predicted by theoretical
4
4
component.
The dependence of optical density on the spectral position
of the dipole component of the plasmon resonance for a typical
reaction is given in Figure 3A. Optical density depends on two
4
3
calculations based on Mie formalism.

Size-Controlled Synthesis of Nanoparticles. 1
J. Phys. Chem. B, Vol. 108, No. 37, 2004 13951
Figure 4. SEM images of different sizes of silver nanoparticles synthesized by hydrogen reduction. The corresponding extinction, scattering, and
absorption spectra can be found in Figure 2 of the subsequent article. Mean particle diameters are (A) 29 nm, (B) 34 nm, (C) 37 nm, (D) 44 nm,
(E) 48 nm, (F) 52 nm, (G) 58 nm, (H) 61 nm, (I) 75 nm, (J) 78 nm, (K) 92 nm, (L) 97 nm, (M) 105 nm, (N) 113 nm, (O) 120 nm, (P) 136 nm.
The scale bar on each micrograph is 1 µm.
As the particles continue to grow and phase retardation begins
to play a significant role, the dipole maximum shifts to the red
revealing a new peak at ca. 420 nm. This peak represents the
quadrupole component of the plasmon resonance of large
particles.43 Another feature at ca. 380 nm in the spectra of large
particles is likely the result of two phenomena. A higher-order
multipole of the resonance such as the hexapole that is no
longer hidden under the envelope of the dipole resonance as
well as the transverse resonance of high aspect ratio particles
present in the suspension21 contribute to this shoulder. This
shoulder can be reduced but never completely eliminated by
centrifugation and filtration procedures, as described in subse-
quent discussion.
Electron Microscopy. SEM images of nanoparticles from
16 different suspensions are shown in Figure 4. The corre-
sponding extinction, scattering, and absorption spectra can be
found in Figure 2 of the subsequent article. Whereas the smallest
size (29 nm) shown in the figure is from a suspension that
43

13952 J. Phys. Chem. B, Vol. 108, No. 37, 2004
Evanoff and Chumanov
resulted after ca. 10 min reaction time, the largest size (136
nm) corresponds to a reaction time of ca. 4 h. Neither of these
sizes represents the limits to which silver particles can be
synthesized using this reaction: particles as small as 10 nm
and as large as 180 nm with a dipole plasmon resonance position
of 670 nm can readily be produced. Larger sizes can also be
synthesized although this requires that the reaction mixture be
stirred to keep the particles suspended.
An immobilization technique was developed to determine true
size and shape distributions, as well as aggregation state of
nanoparticles in suspension via electron microscopy (SEM).
Samples were prepared by first exposing a freshly plasma-
cleaned ITO glass slide to a 2 wt % poly(4-vinylpyridine) (PVP)
solution in ethanol for 3 h with gentle agitation. Afterward, the
slides were rinsed with copious amounts of ethanol and water
to remove any nonadsorbed PVP, dried with a stream of N2,
and treated in an oven at 110 °C for 1 h. This heat treatment
serves to relax the monolayer of polymer facilitating even
distribution across the surface. The lone pair of electrons on
the nitrogen of the pyridine ring binds strongly to silver.
Exposing the modified ITO slide to a silver suspension for ca.
Figure 5. SEM images of (A) silver nanoparticles as prepared and
(B) after filtration.
2
4 h will produce an evenly distributed, nonaggregated mono-
45
layer of silver particles on the slide. We believe that this
method of sample preparation for electron microscopy has a
distinct advantage over the conventional method of drying a
suspension of nanoparticles onto a carbon-coated grid for TEM.
Simple drying of the samples on a grid does not accurately
represent the true particle size and shape distribution of a
suspension. As the solvent evaporates, capillary forces tend to
segregate particles of different size and shape. We consistently
observe concentrating and stacking of high aspect ratio nano-
particles in selected areas of the grid which could, if taken out
of context, lead to the incorrect determination of particle
distribution. In addition, drying of suspensions induces aggrega-
tion of nanoparticles and does not allow the evaluation of the
suspension’s true aggregation state. In the immobilization
technique, the pyridine ring has the same affinity for silver metal
regardless of particle size, shape, or aggregation state, as long
as the substrates are exposed to the suspension for the amount
of time required to establish equilibrium between the surface
and the suspension. A single SEM image in this case provides
an accurate representation of the shape and size distribution of
a suspension as well as aggregation state. With proper experi-
mental conditions, this general approach can also be used for
imaging of other nanoparticles.
Figure 6. SEM image of a resuspended precipitate. Inset shows the
extinction spectrum of the suspension.
aspect ratio particles and, as long as the membrane is not dried,
the particles can be resuspended in water by gentle sonication.
An SEM image of such a suspension and the corresponding
extinction spectrum (inset) are shown in Figure 6. The extinction
spectrum exhibits a maximum around 380 nm typical of the
21
transverse resonance of such high aspect ratio particles.
As can be seen from Figure 7, the extinction spectrum of a
silver suspension synthesized via hydrogen reduction does not
change substantially after the filtration procedure indicating an
overall small percentage of high aspect ratio particles and
aggregates in the preparation. A noticeable reduction on the blue
side of the resonance indicates a decrease in the percentage of
high aspect ratio particles because the transverse resonance of
these particles is no longer contributing to the extinction.
Typically, filtration reduces the fraction of high aspect particles
to 3%.
Filtration and Centrifugation. An SEM image of an as-
prepared suspension of silver nanoparticles synthesized via
hydrogen reduction is shown Figure 5A. As can be seen, the
reaction produces a fairly monodispersed suspension that
contains, in addition to the major fraction, a small percentage
of very small particles and particles with high aspect ratio. The
latter is typically present as 10-15% of the total number of
particles whereas the contribution from the former does not
exceed a few percent and, in some preparations, was completely
absent. The high aspect ratio particles have a short axis of ca.
Centrifugation was employed to remove the fraction of
smaller particles, if present, and to concentrate the nanoparticle
suspension. The required time and speed depend on the diameter
of the major fraction but as a general rule, several hours at 150g
gently precipitates the main fraction without aggregation, leaving
the smaller particles in the supernatant. Once the supernatant
is removed the precipitate can be resuspended to a desired
particle concentration. The optical density, that is, concentration
of the silver suspension, can be readily increased 250 times
without affecting particle stability or aggregation state.
3
0 nm and a long axis between 0.1 and 10 microns.
To remove the high aspect ratio particles, any aggregated
clusters, and any silver oxide particulates that may remain in
the suspension, the mixture is gravity-filtered through hydro-
philic nylon membranes with mean pore sizes of 0.45, 0.22,
and 0.1 microns. The 0.1 micron membrane is omitted for
suspensions with mean particle diameters greater than 130 nm.
The method of filtration can also be used to collect the high

Size-Controlled Synthesis of Nanoparticles. 1
J. Phys. Chem. B, Vol. 108, No. 37, 2004 13953
[311], and [222] crystal planes of silver, respectively. The
background is most likely caused by Brownian motion of the
particles during the scan, solvent scattering, or the short-range
order of the solvation cage around the particles. Previous PXRD
studies of similarly prepared, dry silver nanoparticles coated
with titania did not exhibit any background further supporting
the crystalline nature of the particles synthesized by this
4
6
method.
Crystallinity of the nanoparticles was further studied with
electron diffraction. Diffraction patterns were captured using a
convergent beam TEM operating in a mode that approximates
a parallel beam. As can be seen in Figure 8B and 8C, the ED
patterns demonstrate the highly crystalline nature of the particles
although some patterns exhibit weaker, secondary spots that may
indicate twinned crystals or a submonolayer of an adsorbed
species, most likely silver oxide.
Particle Nucleation, Growth, and Stability. The reaction
of silver(I) oxide and hydrogen gas is represented by eq 1.
0
Ag O(aq) + H (g) f 2Ag (s) + H O(l)
(1)
2
2
2
While this equation highlights the simplicity of the reaction and
the “chemical cleanness” of the resulting nanoparticle suspen-
sion, it sheds little light on the mechanism by which the particles
are formed and stabilized. The solubility of silver(I) oxide in
Figure 7. Normalized extinction spectra of suspensions after filtration
through membranes with decreasing pore size. The inset depicts the
spectra prior to normalization. The spectra were normalized using two
points, the maximum of the dipole resonance and the minimum at 320
nm.
-4
-1 47
water (2.2 × 10 mol Ag kg ) is described by the following
4
8
reactions and solubility constants.
1
1
+
-
-8
/2
Ag O(s) + / H O T Ag + OH
K ) 1.9 × 10
2
2
2
s0
(2)
1
1
1
-6
/2
Ag O(s) + / H O T AgOH
K ) 2 × 10
(3)
2
2
2
s1
1
-
-
/2
Ag O(s) + / H O + OH T Ag(OH)
2 2 2 2
-4
K ) 1.9 × 10 (4)
s2
The pH of a 70 °C silver(I) oxide/water mixture in the Pyrex
reaction vessel has been measured to be ca. 10. Using the above
Ksp values, the major contributor of silver oxide solubility is
the monovalent silver ion, which roughly has a concentration
-
4
-
of 1 × 10 M. The concentration of the Ag(OH)2 would be
-8
around 3 × 10 M. Because of the low solubility product, these
values are only estimates as the composition of the saturated
solution can vary greatly depending on the presence of CO2-
4
8
(
g). Because of the basic nature of the saturated solution, the
composition may also be affected by the dissolution of the glass,
increasing the concentration of the silver cation.
The interior glass surface of the reaction vessel plays a major
role not only in silver(I) oxide solubility but also in the synthesis
mechanism. To study this effect, nanoparticle suspensions
synthesized in a Pyrex vessel were compared to those synthe-
sized in a quartz vessel and in a Pyrex vessel treated with a
fluorinated silane. The silanation of the vessel is similar to a
previously reported procedure. Briefly, a Pyrex vessel was
etched with a 3% hydrofluoric acid solution for 10 min and
then rinsed with copious amounts of water. The vessel was dried
for several hours in an oven at 110 °C. An aliquot of
(Tridecafluoro-1,1,2,2-tetrahydrooctyl)trichlorosilane was placed
in a small Teflon boat on a warm hot plate and the hot, dry
vessel was placed over the boat and allowed to sit for about 15
min. The vessel was then rinsed well with water to facilitate
hydrolysis of the precursor and to remove any unattached
fluorosilanes and was heated in an oven at 250 °C for an hour
Figure 8. (A) Powder X-ray diffraction pattern of a silver nanoparticle
suspension. (B-D) Typical electron diffraction patterns of silver
nanoparticles showing the [111] (B) and [011] (C) planes and a
corresponding STEM image (D). The scale bar is 150 nm.
2
8
Electron and X-ray Diffraction. The nanoparticles synthe-
sized by hydrogen reduction appear to be crystalline as crystal
faces are evident in SEM images. A PXRD pattern of an
aqueous suspension of silver nanoparticles with a diameter of
12
-1
1
30 nm and concentration of ca. 1 × 10 mL is shown in
Figure 8A. The pattern exhibits peaks at 2θ angles of 37.9, 44.0,
3.9, 76.7, and 80.9 that correspond to the [111], [200], [220],
6

13954 J. Phys. Chem. B, Vol. 108, No. 37, 2004
Evanoff and Chumanov
Figure 11. Extinction spectra of aliquots taken during the reduction
3 2
of AgNO by H in a Pyrex vessel.
Figure 9. Extinction spectra from the synthesis of silver nanoparticles
using a fluorocarbon-coated Pyrex vessel. Inset: an STEM image of
the final suspension. The scale bar is 600 nm.
vessels.48 To determine whether there is a direct relationship
+
between the concentration of Ag and the resulting nanoparticle
-4
concentration, hydrogen reduction was performed on a 3 × 10
M AgNO3 solution at 85 °C for 2 h. The extinction spectra of
aliquots taken during the reaction are shown in Figure 11
indicating a very low concentration of formed nanoparticles.
The result suggests that, although the monovalent silver cation
is a major component in the saturated silver(I) oxide solution,
+
Ag alone is not readily reduced to form silver nanoparticles.
It appears that the surface of the Pyrex glass plays a key role
in the formation of silver nanoparticles by catalytically inducing
nucleation through the interaction of the positive ions in Pyrex
and the negatively charged silver hydroxide complexes. These
negative hydroxide complexes anchor to the walls of the vessel
and are reduced by hydrogen molecules forming nucleation
centers that are weakly attached to the surface. The nucleation
centers undergo further growth and eventually leave the surface
as small silver particles. As the concentration of the particles
in the suspension increases, the reduction of silver shifts from
the surface of the vessel to the surface of small particles.
Consequently, the growth of the existing seeds becomes more
favorable than the formation of new ones. The surface of the
Pyrex vessel becomes coated with a silver mirror during the
nanoparticle synthesis indicating that not all seeds that are
formed on the walls of the vessel leave the surface. Some seeds
remain attached and grow into a continuous mirror film. The
same is not true for quartz and fluorocarbon-modified vessels;
however, the addition of Pyrex shot to these vessels greatly
increases the final concentration of nanoparticles and the shot
becomes coated with a mirror film. This observation additionally
supports the catalytic role that the surface of the vessel plays
in this reaction. The finite time required for the reaction to switch
from seed formation to particle growth determines the initial
size distribution of nanoparticles. During the remainder of the
reaction, all particles grow quite uniformly preserving the initial
size distribution and, as the particle become larger, they appear
more monodispersed.
Figure 10. Comparison of silver nanoparticle suspensions with the
same plasmon resonance maximum grown in a Pyrex vessel (solid line)
and a quartz vessel (dashed line).
to allow cross-linking of the silanes, making the interior of the
vessel hydrophobic. The extinction spectra from aliquots taken
during a 3-h synthesis are shown in Figure 9. The inset shows
an STEM image of the nanoparticles produced in the synthesis.
The particles are almost exclusively square plates that are not
stable under the 200 keV electron beam, “melting” into clusters
of small particles as was captured in the inset. The overall
concentration of particles in the suspension is quite low.
When silver suspensions are synthesized in a high-purity
quartz vessel, size and shape distributions were generally similar
to that of a suspension grown in Pyrex but the resulting concen-
tration of the particles was typically 10 times lower (Figure 10).
This result indicates that the chemical properties of the surface
play a crucial role in the formation of the initial seeds.
Particle stability is more than likely achieved by strong
coordination of water to the silver surface. The solvation cage
of the particles can be considered sufficiently “thick” to prevent
coalescence of the metal particles resulting in stable suspensions.
Comparing the extinction spectra of suspensions freshly prepared
+
Silver(I) oxide, specifically when it dissociates to yield Ag ,
has different solubility constants in glass and passivated glass

Size-Controlled Synthesis of Nanoparticles. 1
J. Phys. Chem. B, Vol. 108, No. 37, 2004 13955
surface passivation by uncontrolled contaminants, which may
come from the reactants or from the glass itself when exposed
to a pH 10 silver oxide solution.
Organic contaminants in the water supply are also of great
concern as they can act as a reducing agent for silver oxide.
Silver oxide/water mixture with trace organics will turn yellow
even at room temperature indicating the formation of small silver
nanoparticles that are passivated by reaction products. These
products are present in excess and are capable of passivating
more seeds as they form at the beginning of hydrogen reduction
that follows. As a result, the concentration of nanoparticles
increases as more particles must form since the reaction products
hinder the growth of the initial seeds. Large concentrations of
small particles lead to the collapse of the suspension before the
particles can grow to diameters greater than ca. 40 nm.
Conclusions
A method for size-controlled synthesis of chemically clean
silver nanoparticles is developed. The method is highly efficient
and easily scalable to large production. By conducting the
synthesis in the presence of various complexating species,
control is potentially achieved for synthesis of nanoparticles with
variety of shapes.
Figure 12. Extinction spectra of a “seeded” hydrogen reduced silver
suspension. Inset A: SEM image of the resultant suspension (scale
bar ) 1 µm). Inset B: Extinction spectrum of the seed suspension.
Acknowledgment. This research was supported by the
Environmental Protection Agency. The authors also wish to
thank Ames Laboratory for the equipment used in this work.
to those stored for more than a year revealed no change in the
position of the resonance and only a minimal change in optical
density. SEM imaging showed no change to the larger particles
and only a slight change to the smaller ones most likely caused
by photodegradation.
References and Notes
(1) Creighton, J. A.; Blatchford, C. G.; Albrecht, M. G. J. Chem. Soc.,
Faraday Trans. 2 1979, 75, 790.
(
(
(
2) Lee, P. C.; Meisel, D. J. Phys. Chem. 1982, 86, 3391.
3) Nie, S.; Emory, S. R. Science 1997, 275, 1102.
4) Sondi, I.; Goia, D. V.; Matijevi c´ , E. J. Colloid Interface Sci. 2003,
Effect of Extraneous Species. The purity of the reactants is
an absolute essential for the synthesis of the monodispersed,
nonaggregated silver nanoparticles by this reaction. Any inor-
ganic or organic contaminants in water or Ag2O can passivate
the surfaces of the growing particles leading possibly to
preferential growth along specific crystallographic axes. Using
this strategy, the growth of high aspect ratio particles was
achieved by using small, passivated silver particles to “seed”
the synthesis reaction. The small silver seeds were synthesized
260, 75.
(5) Cason, J. P.; Khambaswadkar, K.; Roberts, C. B. Ind. Eng. Chem.
Res. 2000, 39, 4749.
(6) Li, X.; Zhang, J.; Xu, W.; Jia, H.; Wang, X.; Yang, B.; Zhao, B.;
Li, B.; Ozaki, Y. Langmuir 2003, 19, 4285.
(7) Shirtcliffe, N.; Nickel, U.; Schneider, S. J. Colloid Interface Sci.
999, 211, 122.
1
(8) He, S.; Yao, J.; Jiang, P.; Shi, D.; Zhang, H.; Xie, S.; Pang, S.;
Gao, H. Langmuir 2001, 17, 1571.
(9) Leopold, N.; Lendl, B. J. Phys. Chem. B 2003, 107, 5723.
(10) Velikov, K. P.; Zegeres, G. E.; van Blaaderen, A. Langmuir 2003,
-
5
by the reduction of a 5 × 10 M Ag2SO4 solution with 2 ×
M NaBH4 under rapid stirring and sonication in an ice
cold bath. The suspension was then heated to near boiling for
h to decompose any residual borohydride, break up the
1
9, 1384.
11) Nickel, U.; Castell, A.; P o¨ ppl, K.; Shirtcliffe, N. Langmuir 2000,
16, 9087.
(12) Tan, Y.; Dai, X.; Li, Y.; Zhu, D. J. Mater. Chem. 2003, 13, 1069.
-
4
1
0
(
5
(
(
13) He, R.; Qian, X.; Yin, J.; Zhu, Z. J. Mater. Chem. 2002, 12, 3783.
14) Chen, D.-H.; Huang, Y.-W. J. Colloid Interface Sci. 2002, 255,
aggregates, and concentrate the suspension. Following two-step
centrifugation to narrow size distribution (Figure 12, inset B),
the seeds were added to the reactor containing silver(I) oxide
and the mixture was adjusted to an optical density of 0.3 at the
plasmon resonance maximum of the seeds.
2
99.
(15) Caswell, K. K.; Bender, C. M.; Murphy, C. J. Nano Lett. 2003, 3,
67.
16) Rodr ´ı guez-Gattorno, G.; D ´ı az, D.; Rend o´ n, L.; Hern a´ ndez-Segura,
6
(
G. O. J. Phys. Chem. B 2002, 106, 2482.
The mixture was then placed in a hydrogen environment and
the reaction proceeded as before. Figure 12 shows the extinction
spectra of aliquots taken at 30-min intervals and a typical SEM
image of the resultant suspension (inset A). The suspension
appears to be a very complex mixture of rods, prisms, and plates
with only a few spherical particles.
(17) Van Hyning, D. L.; Zukoski, C. F. Langmuir 1998, 14, 7034.
(
18) CRC Handbook of Chemistry and Physics, 60th ed.; Weast, R. C.,
Ed.; CRC Press: Boca Raton, FL, 1979.
19) Maillard, M.; Giorgio, S.; Pileni, M.-P. J. Phys. Chem. B 2003,
107, 2466.
(
(
(
20) Chen, S.; Carroll, D. L. Nano Lett. 2002, 2, 1003.
21) Sun, Y.; Yin, Y.; Mayers, B. T.; Herricks, T.; Xia, Y. Chem. Mater.
2
002, 14, 4736.
Because of the 10-fold excess of NaBH4 used in the synthesis
of the seeds, the surface of the particles is passivated by borates,
which hinder silver deposition at certain crystal faces more than
at others leading to preferential growth along certain crystal-
lographic axes. It seems possible that with proper control of
the seed particles’ surface chemistry, this reaction could yield
particles of many different shapes. Conceivably, high aspect
ratio particles that grow in the unseeded reaction result from
(22) Zhang, J.; Han, B.; Liu, M.; Liu, D.; Dong, Z.; Liu, J.; Li, D.;
Wang, J.; Dong, B.; Zhao, H.; Rong, L. J. Phys. Chem. B 2003, 107, 3679.
23) Mafun e´ , F.; Kohno, J.-Y.; Takeda, Y.; Kondow, T.; Sawabe, H. J.
(
Phys. Chem. B 2000, 104, 8333.
(24) Abid, J. P.; Wark, A. W.; Brevet, P. F.; Girault, H. H. Chem.
Commun. 2002, 7, 792.
(
25) Huang, H. H.; Ni, X. P.; Loy, G. L.; Chew, C. H.; Tan, K. L.;
Loh, F. C.; Deng, J. F.; Xu, G. Q. Langmuir 1996, 12, 909.
26) Jin, R.; Cao, Y. C.; Hao, E.; M o¨ traux, G. S.; Schatz, G. C.; Mirkin,
C. A. Nature 2003, 425, 487.
(

13956 J. Phys. Chem. B, Vol. 108, No. 37, 2004
Evanoff and Chumanov
(
27) Zhao, Y.-P.; Ye, D.-X.; Wang, G.-C.; Lu, T.-M. Nano Lett. 2002,
(38) Plyuto, Y.; Berquier, J.-M.; Jacquiod, C.; Ricolleau, C. Chem.
Commun. 1999, 17, 1653.
2
, 351.
(
23.
28) Malynych, S. Z.; Chumanov, G. J. Vac. Sci. Technol., A 2003, 21,
(39) Wang, T. C.; Rubner, M. F.; Cohen, R. E. Langmuir 2002, 18,
3370.
7
(
29) Jensen, T. R.; Malinsky, M. D.; Haynes, C. L.; Van Duyne, R. P.
J. Phys. Chem. B 2000, 104, 10549.
30) Kotov, N. A.; Zaniquelli, M. E. D.; Meldrum, F. C.; Fendler, J. H.
Langmuir 1993, 9, 3710.
31) Zhu, J.; Liu, S.; Palchik, O.; Koltypin, Y.; Gedanken, A. Langmuir
000, 16, 6396.
32) Rele, M.; Kapoor, S.; Mukherjee, T. Mater. Res. Bull. 2002, 37,
275.
33) McLeod, M. C.; McHenry, R. S.; Beckman, E. J.; Roberts, C. B.
J. Phys. Chem. B 2003, 107, 2693.
34) Meziani, M. J.; Rollins, H. W.; Allard, L. F.; Sun, Y.-P. J. Phys.
Chem. B 2002, 106, 11178.
35) Gardea-Torresdey, J. L.; Gomez, E.; Peralta-Videa, J. R.; Parsons,
J. G.; Troiani, H.; Jose-Yacaman, M. Langmuir 2003, 19, 1357.
36) Mitrikas, G.; Trapalis, C. C.; Kordas, G. J. Non-Cryst. Solids 2001,
86, 41.
37) Pol, V. G.; Srivastava, D. N.; Palchik, O.; Palchik, V.; Slifkin, M.
A.; Weiss, A. M.; Gedanken, A. Langmuir 2002, 18, 3352.
(
40) Dirkse, T. P.; Vander Lugt, L. A.; Schyders, H. J. Inorg. Nucl.
Chem. 1963, 25, 859.
41) Dirkse, T. P.; Vander Hart, D.; Vriesenga, J. J. Inorg. Nucl. Chem.
965, 27, 1779.
42) Kreibig, U.; Vollmer, M. Optical Properties of Metal Clusters;
(
(
1
(
(
2
Springer Series in Materials Science 25; Springer-Verlag: New York, 1995;
p 18.
(
2
(43) Kreibig, U.; Zacharias, P. Z. Phys. 1970, 231, 128.
(
(44) Kawata, S. Near-Field Optics and Surface Plasmon Polaritons;
Topics in Applied Physics; Springer-Verlag: New York, 2001; pp 107-8.
45) Malynych, S.; Luzinov, I.; Chumanov, G. J. Phys. Chem. B 2002,
06, 1280.
(
(
1
(
(46) Kumbhar, A.; Chumanov, G. J. Nanosci. Nanotech. 2004, 4, 299.
(
(47) Dirkse, T. P. Solubility Data. SilVer (I) Oxide in Water and Aqueous
2
SolVents; Solubility Data Series 23; Pergamon Press: New York, 1986; p
(
95.
(48) Biedermann, G.; Sill e´ n, L. G. Acta Chem. Scand. 1960, 14, 717.