Langmuir 2003, 19, 6693-6700 6693

A General Method To Coat Colloidal Particles with Silica

Christina Graf,*,†,§ Dirk L. J. Vossen,†,‡ Arnout Imhof,† and
Alfons van Blaaderen*,†,‡
Soft Condensed Matter and Biophysics, Debye Institute, Utrecht University,
Ornstein Laboratory, Princetonplein 5, 3584 CC Utrecht, The Netherlands, and FOM Institute
for Atomic and Molecular Physics, Kruislaan 407, 1098 SJ Amsterdam, The Netherlands

Received May 8, 2003

A general method to coat colloids with silica is described. The amphiphilic, nonionic polymer poly-
(vinylpyrrolidone) (PVP) was adsorbed to various colloidal particles such as small gold colloids, gold-shell
silica-core particles, small and large silver colloids, boehmite rods, gibbsite platelets, and positively or
negatively charged polystyrene. After this functionalization the stabilized particles could be transferred
to a solution of ammonia in ethanol and directly coated with smooth and homogeneous silica shells of
variable thickness by addition of tetraethoxysilane in a seeded growth process. The length of the polymer
used strongly influences the stability of the colloids and the homogeneity and smoothness of the initial
silica coating. This method is especially useful for colloidal particles that cannot be covered directly with
SiO2 by a Stöber-like growth process. Compared to methods known from the literature for the coating of
such particles, this new method is faster and requires neither the use of silane coupling agents nor a
precoating step with sodium silicate, which is poorly reproducible.

Introduction fluorescence microscopy9 or as tracer particles in, for

example, fluorescence recovery after photobleaching
Silica-coated colloidal particles are a class of materials (FRAP)10 or time-resolved phosphorescence anisotropy11
widely used in many fields of colloid and materials science. measurements. Moreover, the possibility of the placement
A wide variety of coating procedures has been developed of dye molecules with nanometer precision allows the study
for these samples. These surface coatings allow manipu- of the local density of states in photonic applications.12
lation of the interaction potential and make it possible to Interest in the use of silica-coated particles as building
disperse colloids in a wide range of solvents from very blocks for photonic crystals is increasing.13-16 Here the
polar to apolar. The so-called Stöber growth1 of silica shells outer silica shell allows tuning of not only the interaction
by addition of tetraethoxysilane to solutions of seed potential of the particles but also the optical properties
particles in an ethanol/ammonia mixture yields smooth of the crystal.
surfaces since the growth takes place on a molecular An outer silica coating also offers new possibilities for
scale.2-4 If silica-coated particles are grown further by the shape control of a particle. Silica particles can be
this procedure, the polydispersity of the particles decreases anisotropically deformed in a controlled way by ion beam
with R-1, where R is the particle radius.4 This makes it irradiation.17 This method was successfully applied to
possible to grow crystals of core shell particles even when silica-coated gold particles.17 Due to deformation of the
the cores are polydisperse. silica shell it was possible to deform spherical gold particles
Silica colloids and silica-coated particles are often used into prolate colloids in a controlled way, while normally
as model particles to study phase behavior, rheology, and gold particles remain unchanged under these conditions.
diffusion. Surface modifications of silica spheres have There are many surfaces that can be directly coated
yielded systems with hard-core potentials,5 short-range with silica because of the significant chemical affinity of
attractive potentials,6 and Yukawa potentials.7 these materials, like clay minerals,18 hematite,19 zirconia,
In addition, the silica layer allows for controlled and titania.20 However, many other surfaces can only be
placement of various dyes.8 Such dye-labeled particles can coated with the help of stabilizers, surfactants, silane
be used in quantitative real-space studies with confocal coupling agents, or a fast precipitation from a water glass
(9) van Blaaderen, A.; Wiltzius, P. Science 1995, 270, 1177.
* To whom correspondence should be addressed. E-mail: (10) van Blaaderen, A.; Peetermans, J.; Maret, G.; Dhont, J. K. G.
cgraf@phys-chemie.uni-wuerzburg.de (C.G.); a.vanblaaderen@ J. Chem. Phys. 1992, 96, 4591.
phys.uu.nl (A.B.). (11) Lettinga, M. P.; van Zandvoort, M. A. M. J.; van Kats, C. M.;
† Utrecht University. Philipse, A. P. Langmuir 2000, 16, 6156.
‡ FOM Institute for Atomic and Molecular Physics.
(12) Photonic Crystals and Light Localization in the 21st Century;
Kluwer Academic Publishers: Dordrecht, Boston, London, 2001; Vol.
§ Present address: Institut für Physikalische Chemie, Univer-
563.
sität Würzburg, Am Hubland, D-97074 Würzburg, Germany. (13) Garcia-Santamaria, F.; Salgueirino-Maceira, V.; Lopez, C.; Liz-
(1) Stöber, W.; Fink, A.; Bohn, E. J. Colloid Interface Sci. 1968, 26, Marzan, L. M. Langmuir 2002, 18, 4519.
62. (14) Graf, C.; van Blaaderen, A. Langmuir 2002, 18, 524.
(2) Philipse, A. P. Colloid Polym. Sci. 1988, 266, 1174. (15) Velikov, K. P.; Blaaderen, A. v. Langmuir 2001, 17, 4779.
(3) Bogush, G. H.; Zukoski, C. F. J. Colloid Interface Sci. 1991, 142, (16) Moroz, A. Phys. Rev. B 2002, 66, 115109.
19. (17) Snoeks, E.; van Blaaderen, A.; van Dillen, T.; van Kats, C. M.;
(4) van Blaaderen, A.; van Geest, J.; Vrij, A. J. Colloid Interface Sci. Brongersma, M. L.; Polman, A. Adv. Mater. 2000, 12, 1511. Roorda, S.;
1992, 154, 481. van Dillen, T.; Kooi, B.; de Hosson, J.; Graf, C.; van Blaaderen, A.;
(5) van Helden, A. K.; Jansen, J. W.; Vrij, A. J. Colloid Interface Sci. Polman, A. Submitted for publication in Adv. Mater.
1980, 8, 312. (18) Iler, R. K. U. S. Patent No. 2,885,366, 1959.
(6) Jansen, J. W.; de Kruif, C. G.; Vrij, A. Chem. Phys. Lett. 1984, (19) Ohmori, M.; Matijevic, E. J. Colloid Interface Sci. 1992, 150,
107, 450. 594.
(7) Philipse, A. P.; Vrij, A. J. Colloid Interface Sci. 1989, 128, 121. (20) Ryan, J. N.; Elimelech, M.; Baeseman, J. L.; Magelky, R. D.
(8) van Blaaderen, A.; Vrij, A. Langmuir 1992, 8, 2921. Environ. Sci. Technol. 2000, 34, 2000.

10.1021/la0347859 CCC: $25.00 © 2003 American Chemical Society

Published on Web 07/11/2003
6694 Langmuir, Vol. 19, No. 17, 2003 Graf et al.

solution, and most of these coating methods are multistep

processes. For instance, in procedures using mercapto- or
aminosilanes as coupling agent gold colloids,21 silver
colloids,22 and semiconductor nanocrystals23,24 could be
coated with silica.
A critical step in many silica-coating procedures is the
transfer of colloids that are only stable in aqueous solution
to ethanol where the classical Stöber process is performed.
A widely used technique to achieve this transfer is the
water glass process.18,21,25 First, colloidal particles are
covered with a thin layer of sodium silicate in aqueous
solution. If the particles are then sufficiently stabilized,
they can be transferred into ethanol and be further coated
with a thicker silica layer using seeded growth. In this
way boehmite rods could be coated with silica via a three-
step coating procedure.25 Such a step was also necessary
for the coating of gold and silver colloids via the use of
silane coupling agents. A disadvantage of this method is
that the growth of the initial shell with sodium silicate is
strongly pH dependent and not very well controllable.
The formation of such an initial silica coating on gold
colloids requires reaction times between 24 h and several
weeks14,21,26,27 before a sufficiently thick shell for transfer
of the particles into ethanol is achieved.
In this paper we present a general, simple and fast Figure 1. Diagram of the general procedure for the coating
method to coat colloids with silica. This method is based of colloids with silica. In the first step poly(vinylpyrrolidone)
on the use of poly(vinylpyrrolidone) (PVP) as a coupling is adsorbed onto the colloidal particles. Then these stabilized
agent. This amphiphilic, nonionic polymer is widely used particles are transferred into a solution of ammonia in ethanol.
in science and technology and adsorbs onto a broad range A silica shell is grown by consecutive additions of tetraethoxy-
silane.
of different materials such as metals (e.g., gold, silver,
iron), metal oxides (kaolinite, TiO2, iron oxide, alumina),28
polystyrene,29 silica,30 graphite,31 and cellulose.32 It sta- of the glassware was obtained from a water purification system,
WATER PRO PS from Labcono, and had a measured resistivity
bilizes colloidal particles in water and many nonaqueous
of 2 MΩ.
solvents. In this article we show how PVP can be adsorbed
Syntheses. The general procedure to coat colloids with silica
onto various colloids which can then be directly transferred consist of two steps: adsorption of PVP and growth of the silica
into an ammonia/ethanol mixture where smooth and shell after transfer of the particles to ethanol. An outline of the
homogeneous silica coatings of variable thickness can be synthesis is shown in Figure 1.Synthesis of the Colloids. Small
grown by addition of tetraethoxysilane (TES). Further, gold colloids of 7 and 19 nm radius were synthesized according
we show that the length of the PVP used plays an to the standard sodium citrate reduction method33,34 and were
important role in the stability of the particles during the not further purified. Silica particles (228 nm radius) with a 38
growth process and the smoothness and homogeneity of nm thick gold shell were prepared as described in ref 14 and
the silica shells obtained. purified by repeated sedimentation and redispersion in water.
Small silver particles of 13 nm radius were synthesized by a
modified polyol process:35 Silver nitrate was reduced in ethylene
Experimental Section
glycol in the presence of PVP-10. After the synthesis, the particles
Materials. Tetraethoxysilane (TES, g98.0%) was obtained were separated from ethylene glycol by addition of acetone (500
from Fluka. Poly(vinylpyrrolidone) with average molar masses mL of acetone/75 mL of reaction mixture) and subsequent
of 360 kg/mol (PVP-360), of 40 kg/mol (PVP-40), and of 10 kg/mol centrifugation at 600g as described in ref 35. Next, the super-
(PVP-10) was purchased from Sigma-Aldrich. Poly(vinylpyrroli- natant was removed and the particles were redispersed in
done) with an average molar mass of 3.5 kg/mol (PVP-3.5) was ethanol, again centrifuged at 600g, and redispersed in a solution
obtained from Acros Organics. Ethanol (p.a.), acetone (p.a.), of ammonia in ethanol (4.2 vol % ammonia (29.3 wt % NH3 in
ammonia (29.3 wt % NH3 in water), and hydrofluoric acid (38- water) in ethanol). This solution (c ) 0.31 g/L) could be directly
40%) were purchased from Merck. All chemicals were used as used in the silica-coating step (see below). Large silver colloids
received. Water used in the described reactions and for cleaning of 320 nm radius were synthesized by reducing silver nitrate
with ascorbic acid in the presence of the polymeric stabilizer
(21) Liz-Marzan, L. M.; Giersig, M.; Mulvaney, P. Langmuir 1996, gum arabic36 and purified by repeated sedimentation and
12, 4329. redispersion in water. Boehmite rods and gibbsite platelets were
(22) Ung, T.; Liz-Marzan, L. M.; Mulvaney, P. Langmuir 1998, 14, prepared from aqueous aluminum oxide solutions by hydro-
3740. thermal treatment at 85 °C (gibbsite) and at 150 °C (boehmite)
(23) Correa-Duarte, M. A.; Giersig, M.; Liz-Marzan, L. M. Chem.
Phys. Lett. 1998, 286, 497. and purified by dialysis against demineralized water as described
(24) Bruchez, M.; Moronne, M.; Gin, P.; Weiss, S.; Alivisatos, A. P. in refs 37 and 38.
Science 1998, 281, 2013. Cationic polystyrene spheres were prepared by surfactant-
(25) van Bruggen, M. P. B. Langmuir 1998, 14, 2245. free emulsion polymerization as described in ref 39 and anionic
(26) Makarova, O. V.; Ostafin, A. E.; Miyoshi, H.; Norris, J. R.; Meisel, sulfate stabilized polystyrene spheres as described in ref 40.
D. J. Phys. Chem. B 1999, 103, 9080.
(27) Hall, S. R.; Davis, S. A.; Mann, S. Langmuir 2000, 16, 1454.
(28) Pattanaik, M.; Bhaumik, S. K. Mater. Lett. 2000, 44, 352. (33) Enüstün, B. V.; Turkevich, J. J. Am. Chem. Soc. 1963, 85, 3317.
(29) Smith, J. N.; Meadows, J.; Williams, P. A. Langmuir 1996, 12, (34) Frens, G. Nature (London), Phys. Sci. 1973, 241, 20.
3773. (35) Silvert, P. Y.; Herrera-Urbina, R.; Duvauchelle, N.; Vijayakrish-
(30) Esumi, K.; Matsui, H. Colloid Surf., A: Physicochem. Eng. Asp. nan, V.; Elhsissen, K. T. J. Mater. Chem. 1996, 6, 573.
1993, 80, 273. (36) Velikov, K. P.; Zegers, G. E.; van Blaaderen, A. Langmuir 2003,
(31) Otsuka, H.; Esumii, K. J. Colloid Interface Sci. 1995, 170, 113. 19, 1384.
(32) Kotelnikova, N. E.; Panarin, E. F.; Kudina, N. P. Zh. Obshch. (37) Wierenga, A. M.; Lenstra, T. A. J.; Philipse, A. P. Colloid Surf.,
Khim. 1997, 67, 335. A: Physicochem. Eng. Asp. 1998, 134, 359.
Method for Coating Colloidal Particles with Silica Langmuir, Vol. 19, No. 17, 2003 6695

Table 1. Coating with Silica

colloid solution PVP solution
radius cc Vd Mw e c V centrifugation (EtOH)/NH3a TESb
sample [nm] [g/L] [mL] [kg/mol] [g/L] [mL] sf [g] V [mL] V [µL]
gold colloids 7 0.098 40.00 10 25.6 3.45 600 6.57 66.3
gold colloids 19 0.058 34.70 10 25.6 0.65 60 1.97 19.9
gold-shell silica-core particles 266 1.43 30.00 40 110.0 2.11 1 58.50 261.0
small silver colloids 13 10 9.90 100.0
large silver colloids 320 4.72 5.00 40 9.5 8.7 1 8.63 87.2
boehmite rods 5g 9.08 0.55 10 10 99.45 500 40.59 410.0h
gibbsite platelets 167i 8.00 2.00 40 34 198.00 200 138.60 1400.0
cationic polystyrene colloids 186 0.41 4.75 360 7.0 4.75 0.25 500
386 18.97 4 7.6 4 0.33 1500
anionic polystyrene colloids 170 5.40 4 360 6.5 4 0.33 500
a Ammonia (29.3 wt % NH3 in water; polystyrene spheres) or solution of 4.2 vol % ammonia (29.3 wt % NH3 in water) in ethanol. b TES
solution (10 vol % in ethanol). c Concentration. d Total volume. e Molecular mass of poly(vinylpyrrolidone). f Acceleration. g Length 121 nm.
h Stepwise addition over 6 h (see Experimental Section). i Height 10 nm.

Coating with Silica. The reaction vessels used in the silica- The size of the colloids was determined with the measurement
coating step were cleaned with hydrofluoric acid (8 vol %) and tools of the Analysis software provided with the Philips Tecnai
after that rinsed several times with water. All reaction steps microscopes by binarization of the images, separation of the
with gold or silver colloids were carried out under exclusion of particles by standard image analysis techniques, and estimation
light. of the mean particle radius from the measured surface area. A
Functionalizion with PVP. The molar mass of the PVP used diffraction grating was used to calibrate the magnification.
in this step depends on the size of the particles used and is given SEM. Scanning electron microscopy (SEM) was used to study
in Table 1. The concentration and amounts of the colloidal solution the positively and negatively charged polystyrene spheres and
are also given in Table 1. the silica-coated gold shell silica core particles. Samples were
The amount of PVP was calculated to provide the colloids with prepared by dropping a dilute dispersion (c ≈ 0.05-0.005 g/L)
about 60 PVP molecules per nm2 surface, independent of the on a silicon substrate (used as received). Colloidal crystals of
molar mass of the PVP (Table 1). The PVP was dissolved in silica-coated gold-shell silica-core particles were obtained by
water by ultrasonication of the solution for 15 min. Subsequently, putting an ethanolic solution of these particles (c ) 1.43 g/L) on
the PVP and the colloidal solutions were mixed under stirring a silicon wafer and drying in air. The measurements were carried
(600 rpm in the case of the gold and silver colloids and 100 rpm out with a Philips XL30SFEG scanning electron microscope.
in the case of the boehmite rods and the gibbsite platelets.). To EDX. To determine the elemental composition of the particles
guarantee that adsorption was complete, the reaction mixture studied during electron microscopy measurements, an energy-
was stirred for 24 h at room temperature. dispersive X-ray (EDX) detector from EDAX was used together
The polystyrene particles were already in an ethanol solution. with the HRTEM or SEM. For the measurements in combination
These particles were sedimented, the supernatant was removed, with the HRTEM, the same samples were used as for the TEM
and a solution of PVP in ethanol was added. Because of the large imaging. For the measurements in combination with the SEM,
size of the PVP (PVP-360) used for these particles, the chosen diluted colloidal dispersions (c ≈ 5 mg/L) were dropped onto an
PVP concentrations were lower in relation to the total surface aluminum substrate (cleaned with acetone) to give well-separated
area. These solutions were directly used for the silica-coating particles on the surface.
step.
Transfer into Ethanol. To transfer the PVP-stabilized particles Results and Discussion
into ethanol, the solutions were centrifuged or sedimented (see
Table 1) and the supernatant was removed. Coating of Small Gold Colloids and Influence of
Synthesis of the Silica Shell. The sediments were redispersed the Polymer Length. A detailed discussion of our method
in a solution of ammonia in ethanol (4.2 vol % ammonia (29.3 wt is presented for the coating of small gold particles, because
% NH3 in water) in ethanol). In the case of the polystyrene spheres, under many conditions silica cannot be grown directly on
ammonia (29.3 wt % NH3 in water) was added to the PVP such particles.
polystyrene solutions. Immediately after this a TES solution (10 We began by trying a number of alternative methods
vol % in ethanol) was added under stirring (600 rpm in the case described in the literature. The growth of silica shells on
of the gold and silver colloids and 100 rpm in the case of the colloidal particles based on the method of Stöber et al.1 is
boehmite rods and the gibbsite platelets). In the case of the
boehmite rods the TES solution was added in four steps over 6
usually performed in an ethanol/ammonia mixture. If the
h (first 51.25 µL of TES was added, after 2 h again 51.25 µL, after total surface area of the seed particles per volume is large
4 h 102.5 µL, and after 6 h 205 µL). The total amount of TES enough compared to the added amount of TES, the
added was chosen depending on the desired thickness of the formation of new silica particles can be completely
silica shell. The reaction mixtures were then stirred for another suppressed and therefore the thickness of the silica shell
12 h. can be precisely controlled by the added amount of
Characterization. TEM. Samples for transmission electron TES.2,4,41,42 For gold particles made by the standard sodium
microscopy (TEM) were prepared by dipping copper 400-mesh citrate reduction method,33,34 we found it not possible to
carrier grids covered with carbon-coated Formvar films in the coat them directly in this way with silica. In this case
dispersions (c ≈ 0.5-5 g/L). The boehmite, gibbsite, and large aggregates of uncoated gold particles and silica are
polystyrene particles were imaged with a Philips Tecnai 20 FEG
high-resolution transmission electron microscope (HRTEM)
formed. However, recently a method was reported to grow
operated at 200 keV. All other colloids were imaged with a Philips a silica shell directly onto gold particles by addition of
Tecnai 12 transmission electron microscope operated at 120 keV. TES to a diluted 2-propanol solution.43 We found that this
procedure was not applicable to citrate-stabilized particles.
(38) Buining, P. A.; Pathmamanoharan, C.; Jansen, J. B. H.;
Lekkerkerker, H. N. W. J. Am. Ceram. Soc. 1991, 74, 1303. (41) Bogush, G. H.; Tracy, M. A.; Zukoski, C. F. J. Non-Cryst. Solids
(39) Goodwin, J. W.; Ottewil, R. H.; Pelton, R. Colloid Polym. Sci. 1988, 104, 95.
1979, 257, 61. (42) Chen, S.-H.; Dong, P.; Yang, G.-H.; Yang, J.-J. J. Colloid Interface
(40) Furusawa, K.; Norde, W.; Lyklema, J. Kolloid Z. Z. Polym. 1972, Sci. 1996, 180, 237.
250, 908. (43) Lu, Y.; Yin, Y. D.; Li, Z. Y.; Xia, Y. A. Nano Lett. 2002, 2, 785.
6696 Langmuir, Vol. 19, No. 17, 2003 Graf et al.

Figure 2. TEM pictures of gold particles (7 nm radius) coated with silica after functionalization with PVP-40 (a), PVP-10 (b), or
PVP-3.5 (c).
With the procedure of ref 43, only some of the gold colloids For all three chain lengths practically no secondary
and most of these particles were coated. Also, a lot of new nucleation of small silica colloids was observed, because
silica particles were formed. It is possible that on the the total particle surface per volume was sufficiently high
surface of the particles used in ref 43 a polymeric stabilizer (see ref 42). In the case of the particles stabilized with the
was present on the particles, but the supplier of the colloids longer polymer, mostly single gold colloids covered with
(Ted Pella, Inc.) did not provide this information. Liz- a silica shell were found (see Figure 2a for PVP-40 and
Marzan et al.21 developed a method to coat citrate- Figure 2b for PVP-10). However, the particles stabilized
stabilized gold particles with silica using the silane with PVP-3.5 form silica-covered aggregates of several
coupling agent (3-aminopropyl)trimethoxysilane as a gold particles (see Figure 2c). Apparently, the shorter
primer to reach a higher affinity of the gold surface to polymer cannot sufficiently stabilize the particles during
silica. After the formation of a thin silica layer in aqueous the growth of the silica shell, possibly because it does not
solution from sodium silicate, the particles were trans- sufficiently shield the large van der Waals forces between
ferred into ethanol for further growth using the Stöber these particles. The comparison of Figure 2a and Figure
method. In this way, homogeneously coated gold particles 2b shows a second important influence of the length of the
could be obtained. However, in this synthesis the gold polymer used: while the particles coated with PVP-10
particles easily aggregate during the transfer to ethanol have a relatively smooth and homogeneous coating
if the initial silica shell grown in water is not thick enough. thickness of the silica layer (see Figure 2b), the particles
To achieve both a stabilization of the gold colloids during coated with PVP-40 have a much more inhomogeneous
the shell growth and a higher affinity of the gold surface coating thickness and some particles with multiple gold
to silica, poly(vinylpyrrolidone) of different molar masses cores are observed (see Figure 2a). The sizes of the larger
(3.5, 10, and 40 kg/mol) was adsorbed onto the particles. two adsorbed polymers are comparable to the size of the
Poly(vinylpyrrolidone) (PVP) (see Figure 1) is an am-
gold colloids (r ) 7 nm) and the thickness of the silica
phiphilic polymer that is soluble in water and many
shell (about 10 nm); e.g., in water the hydrodynamic radius
nonaqueous solvents.44,45 This behavior arises from the
of PVP-10 is 4 nm and that of PVP 40 is 8 nm (from ref
presence of a highly polar amide group within the
29). The silica coating of the PVP-3.5 coated gold colloids
pyrrolidone ring and apolar methylene and methine groups
is also smooth and everywhere is of the same size. The
in the ring and along its backbone.29 Due to its amphiphilic
character it can be adsorbed onto many different surfaces. results obtained for the different PVP lengths show that
After 24 h of stirring in the presence of PVP, the particles the silica is growing directly onto the adsorbed polymer.
remained not visibly changed after sedimentation by If a polymer is used that is very large compared to the
centrifugation. This could be shown by electron microscopy colloid radius (such as PVP-40), sometimes particles
and UV-vis spectroscopy. To achieve complete dissolution appear to be only partially coated with silica (see Figure
of the PVP in the aqueous solution and a homogeneous 2c), so in this case the particles are not sufficiently
coating of the gold colloids, it is important to homogenize stabilized by an outer silica layer during the conditions
the polymer solution by ultrasonification before it is added of the Stöber growth and grow together with other
to the colloid solution. Otherwise the silica coatings were particles. This explains why some aggregation is observed
observed to be less homogeneous as shown by TEM despite the fact that a large polymer was used. The use
measurements. Redispersion of the gold colloids in a of PVP with an average molar mass of 10 kg/mol appears
mixture of ammonia and ethanol (gold particle concentra- to be an optimum for this size of particles, because it
tion 0.59 g/L) also gave no loss of particle stability, contrary provides sufficient steric stabilization of the colloids to
to the citrate-stabilized spheres. To these solutions a avoid aggregation during the silica shell growth, but it is
diluted TES solution was added, and in all three cases on the other hand small enough to form a relatively
after that the gold particles were covered with silica (see homogeneous layer onto colloids of this size to obtain a
Figure 2). smooth silica coating.
The PVP makes the affinity of the gold surface to silica It should be noted that the PVP functionalized colloids
sufficiently high, so that no coupling agent is necessary. should not be stored too long in water (less than 2 days)
before the silica shell is grown. If the particles are stored
(44) Molyneux, P. Water-soluble synthetic polymers: properties and for a longer time, the silica coatings become less complete
behaviour; CRC Press Inc.: Boca Raton, FL, 1983.
(45) Franks, F. Water: A Comprehensive Treatise; Plenum Press: and some particles remain uncoated. Three weeks after
New York, 1982; Vol. 3. the initial polymer coating it is not possible to grow any
Method for Coating Colloidal Particles with Silica Langmuir, Vol. 19, No. 17, 2003 6697

Figure 4. TEM pictures of silica particles (228 nm radius)

Figure 3. TEM pictures of gold particles coated with silica. (a) with a 38 nm thick gold shell (total polydispersity 4.4%), which
Gold particles (7 nm radius, polydispersity 15%) with an 18 nm have been coated with an outer silica shell of 21 ( 2 (a) or 60
shell (total polydispersity 9%) grown by two additions of TES. ( 6 nm (b) thickness. Due to the gold shell the silica core cannot
(b) Gold particles (20 nm radius, polydispersity 12%) with 12 be observed.
nm shell (total polydispersity 9%).

silica shell on the gold colloids. The reason for this is

probably a slow decomposition of PVP.
Thicker silica shells could be grown by repeated addition
of TES to the solution of the silica-covered gold colloids
(with PVP-10 adsorbed onto the gold colloids before the
shell growth) (see Figure 3a). With increasing silica shell
thickness the polydispersity decreases. (For a 9 nm thick
shell (Figure 2b) the polydispersity is 13% and for a 18 nm
thick shell (Figure 3a) the polydispersity is 9%. The
polydispersity of the gold cores is 15%.) This has been
reported before for the seeded growth of pure silica
particles.4 This effect is important if metal particles are
to be used as building blocks for photonic crystals.
Figure 3b shows particles with a 20 nm radius gold core
coated with a smooth silica shell using the same procedure.
Other Noble Metal Particles. To grow photonic
crystals of gold-shell silica-core particles with a high Figure 5. SEM picture of silica particles (228 nm radius) with
volume fraction, an outer silica shell is indispensable to a 38 nm gold shell and an outer silica shell of 10 nm thickness
reduce strong attractive interparticle forces.14 A method (total polydispersity 4.4%) dried from ethanolic solution on a
silicon wafer.
to coat such particles consisting of a water glass coating
step in water and a further silica shell growth in ethanol
similar to the method of Liz-Marzan et al.21 to coat small gold-shell silica-core particles (8.3 g/cm3 respectively 2
gold colloids was already described in ref 14. The gold g/cm3), these newly formed particles could be easily
shell of these particles was made not by the citrate method removed by repeated sedimentation and redispersion. The
but by reduction of HAuCl4 with hydroxylamine onto small initially relatively thin silica shell (about 10 nm) could be
gold clusters (made by reduction of HAuCl4 with tetra- grown further by repeated addition of TES (see Figure 4).
kishydroxymethylphosphonium chloride) attached onto These further growth steps were carried out at a higher
the silica spheres. Due to the different stabilization of particle and TES concentration to prevent further sec-
these particles, no initial coating with 3-aminopropyl- ondary nucleation. Aggregation of the particles during
trimethoxysilane (APS) is necessary for a water glass these further growth steps was not observed due to
coating. As mentioned above, the growth of sodium silicate shielding of the gold-gold van der Waals forces by the
is not well controlled. Despite the fact that these particles silica layer.
are also stable in ethanol, no shell growth was observed The silica layer stabilizes the particles. Because of their
by direct Stöber coating experiments with TES. low polydispersity (4.4%), the silica-coated gold-shell silica-
To coat these particles with silica, a longer PVP (PVP- core particles form large crystals in solution or when they
40 or PVP-360) was adsorbed onto the particles because are dried from ethanol onto a silicon wafer (see Figure 5).
of their much larger size (266 nm radius). This step was Uncoated gold shell particles form crystals in solution
carried out at a lower TES concentration (1.99 mM instead but aggregate during the drying process before they can
of 4.48 mM) to prevent aggregation of the particles during form an ordered structure.
the growth of the initial silica shell. At higher TES Because of their low bulk absorption, silver colloids are
concentrations some of the particles grew together during the most suitable metal particles for the creation for
the formation of the silica shells regardless of the length photonic crystals with a band gap in the visible.16 For the
of the polymer used. Because of the less favorable TES to synthesis of crystals with a high volume fraction of metal
surface area ratio as compared to the smaller gold particles, it is necessary to tune the particle surface
particles, secondary nucleation of small silica particles potential, which can easily be achieved by coating them
(less than 20 nm radius) was observed. However, due to with an outer silica layer. Under certain conditions silver
the much larger radius and much higher density of the particles can be directly coated with silica if they are
6698 Langmuir, Vol. 19, No. 17, 2003 Graf et al.

Figure 6. TEM pictures of large silver particles (320 nm radius, Figure 7. (a) HRTEM pictures of boehmite rods (thickness 9
polydispersity 15%) with 80 ( 12 nm silica shell (a; the inset nm, length 120 nm, length polydispersity 38%) coated with a
shows an enlarged part of one of the silica coated particles) and 10 nm thick silica layer. (b) Bright field image (top) and dark
of small silver particles (13 nm radius, polydispersity 23%) with field image (bottom) of the same rod.
a 47 nm silica shell (total polydispersity 5.9%) (b).

silica shell onto these particles (see Figure 6b). The

stabilized by organic surfactants.46 However, as for gold
combination of the polyol process with the silica-coating
particles this is not possible. Procedures for coating silver
method presented here opens a simple and quick way to
particles with silica based on the use of a silane coupling
obtain well-defined silica-coated metal colloids.
agent and a precoating with sodium silicate similar to
those mentioned above for small gold colloids have been Coating Metal Oxide Colloids with Silica. Here we
described.22 describe two examples of anisotropic metal oxide particle
Large silver colloids prepared by the reduction of silver systems that cannot be directly coated with silica. A
nitrate with ascorbic acid in the presence of the polymeric laborious method to coat boehmite rods with silica was
stabilizer gum arabic can be coated directly with silica in already developed by van Bruggen.25 This procedure
ethanol; however, the reaction is not very well controllable comprises one sodium silicate step in water, because
and most of the particles aggregate during shell forma- untreated boehmite rods are not stable in ethanol, and
tion.19 Furthermore, it turned out that if the particles then two further coating steps in ethanol, because also
were cleaned once by centrifugation or sedimentation in the precoated particles aggregate very easily.
water, more than 95% of the particles remained uncoated After stirring for 24 h in an aqueous PVP-10 solution,
during a Stöber growth process, very probably because boehmite rods (thickness 9 nm, length 120 nm, length
most of the gum arabic was removed from the colloidal polydispersity 38%) are already stable in ethanol. Cen-
surface. If PVP-40 was adsorbed onto these particles, all trifugation and the transfer into the ethanol/ammonia
of them were coated with a smooth silica layer (see Figure solution also did not influence the stability of the particles.
6a. In contrast to the unpurified particles with gum arabic To coat the rods with silica, a TES solution was added in
adsorbed onto their surface, nearly no aggregates were four portions to the particles under mild stirring. In this
formed during this reaction. Because the large silver way boehmite rods were directly coated with a smooth 10
colloids are formed through the aggregation of smaller ( 2 nm thick silica layer (see Figure 7). Practically no
(e20 nm) subunits, the surface roughness is on the order secondary nucleation was observed under the conditions
of about 10 nm. As shown in Figure 6a, the surface became chosen. To show more clearly the core shell structure of
smoother due to the outer silica layer. the coated rods, a single rod was imaged with HRTEM in
The polyol process47,48 is a widely used synthesis method bright field mode (Figure 7b, top) and in dark field mode
for various metal colloids49 in high concentration in organic (Figure 7b, bottom). In dark field mode a beam diffracted
solution. In a variant of this method PVP is already used by a correctly oriented crystal is used to form the image.
during the synthesis of metal colloids such as nickel,50 Therefore, the amorphous silica shell and all parts of the
copper,51 gold,52 silver,35 platinum, palladium, and ru- boehmite core, which do not have the correct crystal
thenium53 as a reductor and stabilizer. The PVP added in orientation, stay dark while the oriented crystalline core
this synthesis can be also used for coating with silica: For appears bright.
the coating of silver colloids of 13 nm radius prepared by A method to coat hexagonal gibbsite platelets with silica
such a polyol process,35 the PVP present has a suitable has not been published. Like boehmite rods gibbsite
molar mass (PVP-10) for the direct growth of a smooth platelets flocculate in ethanol, so a direct coating with
silica is not possible. With a PVP-10 coating the platelets
(46) Hardikar, V. V.; Matijevic, E. J. Colloid Interface Sci. 2000, 221, (10 nm thickness, 167 nm radius, radius polydispersity
133. 53%) could be redispersed in an ethanol/ammonia mixture.
(47) Figlarz, M.; Fievet, F.; Lagier, J. P. U. S. Patent No. 4,539,041, However, after the growth of a smooth silica shell of about
1985.
(48) Fievet, F.; Lagier, J. P.; Figlarz, M. MRS Bull. 1989, 14, 29. 8 nm thickness, some of the platelets were grown together.
(49) Kurihara, L. K.; Chow, G. M.; Schoen, P. E. Nanostruct. Mater. As was the case for the small gold particles (see above),
1995, 5, 607. it turned out that when a longer PVP was chosen (PVP-
(50) Hegde, M. S.; Larcher, D.; Dupont, L.; Beaudoin, B.; Tekaia-
Elhsissen, K.; Tarascon, J. M. Solid State Ionics 1996, 93, 33. 40) aggregation could be prevented under the same
(51) Saravanan, P.; Jose, T. A.; Thomas, P. J.; Kulkarni, G. U. Bull. reaction conditions, while a smooth and homogeneous
Mater. Sci. 2001, 24, 515. silica coating was still obtained (see Figure 8). Because
(52) Silvert, P. Y.; Tekaia-Elhsissen, K. Solid State Ionics 1995, 82, of the larger size of the gibbsite platelets (167 nm radius)
53.
(53) Bonet, F.; Tekaia-Elhsissen, K.; Sarathy, K. V. Bull. Mater. Sci. compared to the gold colloids (7 nm radius), a longer poly-
2000, 23, 165. (vinylpyrrolidone) (PVP-40 instead of PVP-10) was re-
Method for Coating Colloidal Particles with Silica Langmuir, Vol. 19, No. 17, 2003 6699

Figure 8. TEM picture of gibbsite platelets (167 nm radius,

polydispersity 53%) with 10 ( 2 nm thick silica shells (a) and
HRTEM-picture of a single silica coated gibbsite platelet (b).

Figure 10. TEM pictures of a cationic polystyrene sphere (186

nm radius) with 22 nm silica shell (a), an anionic polystyrene
Figure 9. EDX image (elementary composition as a function sphere (170 nm radius) with 10 nm silica shell (b), and cationic
of position) of cationic polystyrene spheres (386 nm radius, polystyrene spheres (386 nm radius) with 32 nm silica shell (c).
polydispersity 2.6%) with a 32 nm silica shell (total polydis- (d) SEM picture of particles with the same core size as in (c)
persity 2.4%). In (a) the carbon content (KR line) and in (b) the with a piece of a silica shell broken off a particle.
silicon content (KR line) per pixel is imaged. A lighter color
corresponds to a higher number of counts. the smaller positively charged polystyrene particles this
results, as already observed for the small gold colloids, in
quired in this case to obtain sufficient stability during the a relatively rough silica coating (see Figure 10a). In
shell growth process. contrast with this result, the negatively charged poly-
Coating of Positively and Negatively Charged styrene particles of nearly the same size have a smooth
Polystyrene Spheres with Silica. To test the generality silica coating (see Figure 10b). These findings are in good
of our method, we also coated a surface that is in part agreement with the results of Smith et al.29 for the
hydrophobic. For this we chose polystyrene particles. Such adsorption on negatively charged polystyrene in water.
particles can be prepared by various methods. Here we According to the photon correlation spectroscopy mea-
used two different emulsifier-free emulsion polymerization surements presented in ref 29, the adsorbed layer thick-
synthesis procedures resulting in colloidal particles with ness was only 1-3 nm for molar masses of PVP between
positively39 or negatively charged groups on the surface40 10 and 2500 kg/mol, indicating that the molecules were
resulting from the initiator. Here the PVP adsorption steps lying flat on the surface in the form of trains. For other
and the silica-coating steps were carried out in ethanol. types of colloids (gold colloids, positively charged poly-
The coverage of the particles with a silica layer could be styrene) it is more likely that the polymer is not adsorbed
proven for both systems by analyzing the elementary flat onto the gold colloids but is adsorbed more in the form
composition of the colloids with EDX. The results for large of coils, so that here the shorter the polymer the more
positively charged polystyrene spheres (386 nm radius) homogeneous is the obtained silica layer. For the larger
are shown in Figure 9. Similar results were obtained with positively charged polystyrene, the size of the polymer
small positively (186 nm radius) and negatively charged used is relatively small compared to the total particle size.
polystyrene spheres (170 nm radius). While the carbon Thus, here the silica coverage appears to be relatively
content is maximal in the center of the particles (see Figure smooth (see TEM image (Figure 10c) and SEM image of
9a), the silicon content is maximal at the edge (see Figure a part of a silica shell broken off during strong ultra-
9b), demonstrating the presence of the coating. sonification (Figure 10d)).
For three different systems of polystyrene (168 nm
radius and large 386 nm radius for positively charged Conclusions
polystyrene spheres and 170 nm radius for negatively We have developed a new and general method to coat
charged), a PVP polymer relatively large compared to the colloids with silica. By adsorbing poly(vinylpyrrolidone)
particle size (PVP-360, hydrodynamic diameter (in wa- (PVP) onto the colloidal surface, particles that are stable
ter29) ) 46 nm) was used to functionalize the surface. For in both water and ethanol were obtained. These particles
6700 Langmuir, Vol. 19, No. 17, 2003 Graf et al.

could be directly coated with a silica layer of variable our method is faster and requires neither the use of silane
thickness by addition of tetraethoxysilane (TES) to an coupling agents nor of a poorly reproducible precoating
ammonia-ethanol solution containing the particles in a step with sodium silicate.
seeded growth process. When longer PVP chains were
used, the particles were stabilized better. However, the
coatings obtained with longer polymers were also less Acknowledgment. The authors thank Marcel Gies-
smooth in most cases. An exception was the negatively berts (Utrecht University) for providing the gold colloids
charged polystyrene particles, probably because in this with 19 nm radius. Krassimir P. Velikov (UU) and Gabby
case even large molecular weights of PVP form only a thin E. Zeggers (UU) are thanked for the synthesis of the large
layer.29 silver colloids, and Judith E. G. J. Wijnhoven (UU) is
For colloidal particles between 7 and 320 nm radius thanked for the synthesis of the boehmite rods and the
optimal PVP lengths were found, which depended on the gibbsite platelets. Further we thank Carlos M. van Kats
size of the colloids. At the optimal PVP length stable (UU) for the SEM-measurements of the gold shell particles
colloidal particles coated with a smooth and homogeneous and Hans Meeldijk (UU) for the HRTEM measurements
silica layer were achieved. Due to the amphiphilic and and part of the EDAX analysis. Stephanie Eiden from the
nonionic character of the polymer used, this new process University of Konstanz (Germany) is thanked for the
can be applied to a broad range of colloids and was synthesis of the negatively charged polystyrene and
demonstrated for gold, silver, gibbsite, boehmite, and Patrick M. Johnson is thanked for a careful reading of the
positively and negatively charged polystyrene. This manuscript. This work was financially supported by the
method is especially useful for particles that cannot be Foundation for the Fundamental Research of Matter
coated directly by a Stöber-like1 growth process, for (FOM), which is part of The Netherlands Organization
instance when the particles are not stable in ethanol or for Scientific Research (NWO).
the affinity of the particle surface to silica is too low.
Compared to published methods for specific particles,21,22,25 LA0347859