COM

MUNIC

ATIO

N

DOI: 10.1002/adem.201000108

Synthesis of Monodisperse Silica Nanoparticles Dispersablein Non-Polar Solvents**

By Eoin Murray, Philip Born, Anika Weber and Tobias Kraus*

Three synthetic routes to hydrophobic silica nanoparticles are compared in this paper. First, theestablished synthetic method based on the Stober process was examined. Monodisperse colloidal silicaparticles with diameters of 15–25 nm were prepared via the hydrolysis of tetraethyl orthosilicate(TEOS) by aqueous ammonia in ethanol. The surfaces of these particles were rendered hydrophobicwith octadecyltrimethoxysilane (ODTMS) after the reaction or, more conveniently, during the growthphase. Secondly, silica particles with diameters of 15–50 nm were prepared using a one-pot synthesis inwhich TEOS was hydrolyzed by an amino acid and the resulting particles were coated with ODTMS.Lastly a novel, direct approach to the synthesis of hydrophobic organosilica nanoparticles wasdeveloped using ODTMS as the single silica source. Hydrolysis of the ODTMS by aqueous ammoniain ethanol yielded monodisperse colloidal organosilica particles with diameters of 15–30 nm.

[*] Dr. T. Kraus, Dr. E. Murray, P. Born, A. WeberINM – Leibniz Institute for New Materials, StructureFormation Group Campus D2 2, 66123 Saarbruecken,GermanyE-mail: tobias.kraus@inm-gmbh.de

[**] We thank the Physical Analysis Group at INM for technicalassistance and Christian Cavelius for scientific discussions. Theauthors thank Eduard Arzt for his continuing support.

374 � 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ADVANCED ENGINEERING MATERIALS 2010, 12, No. 5

Silica nanoparticles are common nanoscale components.

Polydispersed nanoscale silica is used as a transparent filler in

polymers, modifies the flow properties of emulsions and

improves the properties of construction materials. With a

narrow size distribution, silica nanoparticles can introduce

nanoscale regularity on surfaces or in bulk materials either as

component of the final material or as an etch mask in further

processing steps. In all cases, particle-matrix compatibility

and composite behavior depend on the surface chemistry of

the particles.

Here we address organically functionalized particles,

which have found application in many areas including paints

and coatings,[1,2] sensors,[3] catalysis,[4,5] and drug delivery

methodologies.[6,7] Hydrophobic silica nanoparticles interact

strongly with the constituent molecules of the polymer matrix

and have been shown to enhance physical and mechanical

properties of nanocomposites.[8,9] Assembly of monodis-

persed silica nanoparticles into regular superstructures is a

promising route to materials with rationally designed

microstructures. Stable suspensions of hydrophobic silica

nanoparticles in apolar solvents are also good model systems

for studying the equilibrium and transport properties of

colloidal dispersions; their refractive index matches that of

some apolar solvents, minimizing multiple scattering in light

scattering experiments.[10–12]

Three solution-based routes to hydrophobic silica nano-

particles are compared in this paper. First, the established

synthetic method based on the Stober process was examined.

Monodispersed colloidal particles with diameters of 15–25 nm

were prepared via the hydrolysis and condensation of

tetraethyl orthosilicate (TEOS) by aqueous ammonia in

ethanol. The surfaces of these particles were rendered

hydrophobic with octadecyltrimethoxysilane (ODTMS) after

the reaction or, more conveniently, during the growth phase.

Secondly, silica particles with diameters of 15–50 nm were

prepared using a one-pot synthesis in which TEOS was

hydrolyzed by an amino acid and the resulting particles were

coated with ODTMS. Lastly a novel, direct approach to the

synthesis of hydrophobic organosilica nanoparticles was

developed using ODTMS as the single silica source. Hydro-

lysis of the ODTMS by aqueous ammonia in ethanol yielded

monodisperse colloidal organosilica particles with diameters

of 15–30 nm.

The first approach used here is based on the commonly

applied hydrolysis and condensation of tetraalkyl orthosili-

cates by a strong base as originally developed by Stober.[13]

The resulting hydrophilic particles can then be modified by

the grafting of organosilicate compounds onto their surfaces

or by the co-hydrolysis of a tetraorthosilicate with the required

functional organosilane. Using these techniques, silica nano-

COM

MUNIC

ATIO

N

E. Murray et al./Synthesis of Monodisperse Silica Nanoparticles Dispersable in . . .

particles were coated with a number of functional groups

including vinyl, carboxylate, thiol, mercaptopropyl,

amine,[14–17] and a variety of biomolecules,[3,18] among many

others. The surface of particles from the Stober synthesis can

also be made hydrophobic by the esterification of the surface

hydroxyl groups using octadecanol.[10,11,14] The resulting

so-called octadecylsilica is a popular model system for

hard-sphere colloids. Badley et al.[15] introduced a more

convenient route to a similar octadecyl-coated silica system by

the reaction of ODTMS with TEOS using a modified Stober

route. He showed that these particles were thermally more

stable than those produced via esterification as the esterified

groups leave the surface at 240 8C while those attached by a

siloxane linkage were stable to over 350 8C.There are limitations to the silica nanoparticles attainable

using variations of the Stober process. While narrow size

distributions can be achieved for larger particles (>200 nm),

smaller particles generally exhibit more irregular shapes and

sizes. Indeed the smallest usable particles that have been

produced using the Stober process are approximately

15–20 nm and have a polydispersity of over 20%.[19,20] To

overcome this limitation, a new route to silica nanoparticles

was recently developed, replacing ammonia as the base

catalyst with basic amino acids such as L-lysine, L-arginine,

and L-histidine.[21–23] Combined with the heterogeneous

delivery of TEOS using a water/alkane biphasic system, it

resulted in improved control of the hydrolysis rate and the

growth of monodisperse silica nanoparticles in the range of

10–20 nm. We have attempted to modify their surfaces and

thus obtained hydrophobic, monodispersed, small silica

particles. A markedly different approach to functionalized

silica nanoparticles was reported recently.[24] Instead of using

two different silanes, only one of which carries the desired

functionality, synthesis is done starting solely from the

functionalized silane. The resulting organosilica particles

carry organic moieties not only on their surface, but also

throughout their core, which can thus host suitable dyes or

other apolar agents.[25]

In the present paper, we compare all three approaches to

hydrophobic silica (Fig. 1). Silica particles were prepared via

hydrolysis of TEOS with ammonia or via hydrolysis of TEOS

Fig. 1. Schematic of synthetic routes featured in this paper and the resulting silicananoparticles. Reactions that are crossed out did not lead to the desired product. ‘‘Polar’’denotes particles that are dispersible in polar solvents while those marked with ‘‘apolar’’are dispersible in apolar solvents. EtOH denotes ethanol, MeOH methanol, and NH4OHaqueous ammonia.

ADVANCED ENGINEERING MATERIALS 2010, 12, No. 5 � 2010 WILEY-VCH Verl

with lysine. The particles were rendered hydrophobic by

addition of ODTMS either during growth or in a second step.

Finally, pure ODTMS was hydrolyzed with ammonia to

directly yield hydrophobic organosilicate particles. The

various particles were analyzed using electron microscopy

and light scattering methods and the different routes were

compared in terms of synthetic yield (the fraction of silane that

is converted into particles) and particle quality.

Experimental

Materials

All chemicals were used as purchased without further

purification. Heptane, toluene, methanol, ammonium hydro-

xide (25wt%), ethanol, and TEOS (tetraorthosilicate), all

puriss p.a. grade, were purchased from Fluka. ODTMS (tech.

grade) and L-lysine (purum) were obtained from Aldrich.

Deionized millipore water (>18mV) was used in all

preparations.

Preparation

Silica nanoparticles were synthesized using a variation of

the standard synthetic route developed by Stober [13] andwere

modified using Badley’s method.[15] Aqueous ammonia

(25wt%, 0.8mL), was mixed in 25mL of ethanol and placed

in an ultrasonic bath. Upon sonication, 0.8mL of TEOS was

added and after 5 h ODTMS (0.3mL) was added and

sonication was continued for an additional 5 h. The sample

was removed from the ultrasonic bath and left to further react

by stirring overnight, by which time the octadecylsilica had

precipitated. The sample was washed twice by centrifugation

and resuspension to remove any unreacted silane and finally

resuspended in a suitable non-polar solvent (usually heptane)

using ultrasonication. Similar results could be obtained by

adding the silane following the washing steps rather than

during the reaction.

To produce similar particles without the use of strong

bases, the route developed by Yokoi and Hartlen [21–23] was

adapted to yield organically coated silica nanoparticles. In a

typical synthesis, 60mg of L-lysine was dissolved in 20mL of

methanol and stirred for 15min. Heptane (20mL) was then

carefully added, and the stirring was adjusted to adequately

mix the lower part of the liquid volume without strongly

perturbing the interface. After 5 h, 0.1mL of ODTMS was

carefully added to the top layer and left to react at room

temperature for 30 h. After this time, the ocytadecyl-coated

silica precipitated and the lower layer could be separated and

centrifuged. The residue was washed twice to remove any

unreacted silane and finally resuspended in a suitable

non-polar solvent (usually heptane) using ultrasonication.

The same protocol without the silane addition was also used

to produce uncoated nanoparticles.

Organosilica nanoparticles were synthesized by replacing

the TEOS in the standard synthetic route developed by Stober

with ODTMS as the single silica source. In a typical synthesis,

aqueous ammonia (25wt%, 0.8mL) was mixed with 25mL of

ag GmbH & Co. KGaA, Weinheim http://www.aem-journal.com 375

COM

MUNIC

ATIO

N

E. Murray et al./Synthesis of Monodisperse Silica Nanoparticles Dispersable in . . .

Fig. 3. Particle size distribution from dynamic light scattering (DLS) of silane modifiedhydrophobic silica in heptane, produced using the Stober synthesis with addition ofoctadecylsilane after the reaction (light) compared to the size distribution of the uncoatedparticles in ethanol from a similar Stober synthesis (dark). The modified particles exhibita slight increase in radius over the modified particles.

ethanol and stirred vigorously at ambient temperature. After

20min, 0.1mL of ODTMS was added and stirring was

continued for 3 h, by which time the coated silica was

precipitating. The sample was washed twice by centrifugation

and resuspension to remove any unreacted silane and finally

resuspended in a suitable non-polar solvent using ultra-

sonication. Similar results could be obtained by using

ultrasonication during the reaction instead of stirring.

Analytical Methods

Transmission electron microscopy (TEM) was performed

using a Philips CM 200 TEM operating at 200 keV with a

point-to-point resolution of 0.24 nm and a lattice resolution of

0.14 nm. For particle characterization, copper TEM grids with

carbon coating were dipped into the suspension and were

dried in an oven at 105 8C. The silica nanoparticles were

characterized with respect to particle size, shape, and degree

of aggregation.

Dynamic light scattering (DLS) analysis was performed

using a Wyatt Technology DynaPro Titan operating at a

wavelength of 831.2 nm to measure the hydrodynamic radius

and standard deviation in the dispersity of the particles in

suspension in an appropriate solvent. Samples for DLS were

prepared by suspending of a small concentration of particles

in a good solvent under sonication. The DynaLS algorithm of

the Alango Company, based on an inverse Laplace transfor-

mation for polydisperse systems which derives an optimal

regularization parameter based on the noise of the data, was

used to fit all autocorrelation data. The DLS distributions are

given here in percent mass.

Results and Discussion

The particles obtained by the modified ambient-

temperature Stober synthetic route were found, by both

TEM and DLS, to have diameters in the range of 15–25 nm

with a polydispersity (by DLS) of approximately 10–15%

(Fig. 2). These particles could subsequently be modified and

made hydrophobic by the addition of ODTMS without

Fig. 2. Transmission electron microscopy images of octadecyl-coated silica nanoparticles symodified Stober route at ambient temperature, (a) with silane addition after the reactionreaction. Scale bars are 100 nm.

376 http://www.aem-journal.com � 2010 WILEY-VCH Verlag GmbH & C

significant increase in either the size of the particles or their

polydispersity. Similarly modified particles could be pro-

duced by in situ coating with ODTMS added during the

hydrolysis reaction. These particles showed a slight increase

in particle radius over bare silica in ethanol (Fig. 3).

We attempted to produce high-quality hydrophobic

nanoparticles using the amino-acid-based silica particle

synthesis. In water, TEOS was hydrolyzed in the presence

of lysine to form silica particles with amean diameter of 20 nm

and 10% polydipersity. However, we were unable to modify

the particles by esterification using the octadecanol route of

Van Helden,[8,10] probably because residual amino acids on

the particle surface inhibited the esterification reaction that is

required for a stable surface modification.

In a second set of experiments, we co-hydrolyzed TEOS

and ODTMS to achieve concurrent synthesis and coating of

the formed nanoparticles. The long alkyl chain renders

ODTMS highly hydrophobic, so that the reaction could not

nthesized using theand (b) during the

o. KGaA, Weinheim

take place in an aqueous system; however,

the catalytic amino acids are only sparingly

soluble in other suitable solvents. We chose

lysine, as it had previously been used as a

base catalyst for silica nanoparticle synthesis

and is soluble in methanol. In methanol,

TEOS was successfully hydrolyzed in the

presence of lysine and the particle surface

modified during growth. The hydrolysis

rate of TEOS in methanol differs from that

in water, partially due to the different

solubility of the catalytic amino acids. At

elevated temperatures (60 8C), hydrolysis is

fast and the resulting particles are large and

polydisperse. At lower temperatures (20 8C),hydrolysis is slower, resulting in slower,

more controllable growth of the nanoparti-

ADVANCED ENGINEERING MATERIALS 2010, 12, No. 5

COM

MUNIC

ATIO

N

E. Murray et al./Synthesis of Monodisperse Silica Nanoparticles Dispersable in . . .

Fig. 4. Particle size distribution from DLS of unmodified silica synthesized using theamino acid route at 37 8C in methanol (dark) and following silane modification andresuspension in heptane (light).The non-silanized particles are largely agglomerated dueto the limited solubility of the stabilizing amino acid residues in methanol .Thehydrophobic particles show good dispersity in heptane with a minimum of largeagglomerates.

Fig. 5. Transmission electron microscopy images of octadecyl modified silica nanoparticles synthesized using theamino acid route at ambient temperature, with silane addition after (a) 5 h and (b) after 12 h. Scale bars are 50 nm.

Fig. 6. On the left is the particle size distribution from DLS of organosilica in heptaneafter 3 h (dark) shows a very narrow particle size distribution and a broader distributionafter 24 h (light). On the right are representative transmission electron microscopyimages of organosilica nanoparticles synthesized using octadecyltrimethoxysilane as thesingle silica source. The scale bar is 10 nm for the very small particles and 100 nm for thelarger.

cles. In the absence of ODTMS, uncoated particles form and

rapidly agglomerate owing to the limited solubility of the

stabilizing amino acid residues in methanol (Fig. 4). Addition

of ODTMS appears to inhibit further growth of the

nanoparticles and as such, the timing of the addition of the

octadecylsilane is very important for the resultant hydro-

phobic particle size and morphology. Addition of the

octadecyl silane after 5 h to a reaction at room temperature

results in very small (10–20 nm) particles with poorly defined

morphology, whereas addition of the silane after 12 h results

in more monodisperse but larger particles (Fig. 4). Similarly,

the increased reaction rate at an intermediate temperature

(37 8C) and silane addition after 5 h result in particles

comparable with those produced with the later addition at

room temperature. All of the resulting particles were found to

be suspendable in a number of apolar solvents such as

heptane, toluene, and chloroform (Fig. 5).

While the synthetic routes detailed above yield hydro-

phobic silica nanoparticles with acceptable quality, they are

ADVANCED ENGINEERING MATERIALS 2010, 12, No. 5 � 2010 WILEY-VCH Verl

not useful in the preparation of very small (<15 nm), spherical

particles. To obtain such particles, we resorted to a route that

was again similar to the Stober synthesis, but used ODTMS as

the single silica source in an ethanol/ammonia solution. In

this synthesis, the size of the organosilica nanoparticles could

be controlled by varying the reaction time. After 3 h with

typical reactant concentrations at ambient temperatures, the

resulting particles were smaller than 10nm and exhibited

below 10% polydispersity. The particles continued to grow

and after 72 h were 20–40 nm in diameter, with polydisper-

sities above 25% (Fig. 6). This reaction is therefore a quick and

facile route to hydrophobic silica nanoparticles with size and

shape dispersity comparable to or better than those produced

by both the amino acid and Stober routes. If the organic groups

inside the particle core are acceptable (or even beneficial, as in

the addition of hydrophobic agents such as dyes), this is the

preferred route to monodispersed hydrophobic nanoparticles

below 20 nm in diameter.

ag GmbH & Co. KGaA

Conclusion

In this paper two novel, one-pot synthetic

routes to hydrophobic silica nanoparticles

have been developed and compared to

proven routes based on the Stober synthesis.

In the standard synthetic route, nanoparticles

in the diameter range of 15–25nm with

polydispersity below 15% were prepared

via hydrolysis of TEOS by ammonia under

sonication and then modified with ODTMS

to yield hydrophobic surfaces. This protocol

was modified by introducing lysine as the

catalyst for the hydrolysis of TEOS in a

biphasic solution. The further addition

of ODTMS to the reaction mix yielded

, Weinheim http://www.aem-journal.com 377

COM

MUNIC

ATIO

N

E. Murray et al./Synthesis of Monodisperse Silica Nanoparticles Dispersable in . . .

hydrophobic nanoparticles in the range of 10–50 nm with 20%

polydispersity.

Neither of the two approaches yielded small (<10 nm),

monodisperse spherical nanoparticles. We thus developed a

new synthetic route, parting with the dual precursor concept.

The synthesis of hydrophobic organosilica particles from

ODTMS as the single silica source in a mixture of ethanol and

ammonia yielded very small monodisperse nanoparticles

from 8 to 50 nm with polydispersities between 10 and 25%.

The organosilica core can possibly also be used as a host for

hydrophobic agents.

Received: December 11, 2009

Final Version: March 3, 2010

Published online: May 6, 2010

[1] M. M. Jalili, S. Moradian, H. Dastmalchian, A. Karbasi,

Prog. Org. Coat. 2007, 59, 81.

[2] J. Douce, J. P. Boilo, J. Biteau, L. Scodellaro, A. Jimenez,

Thin Solid Films 2004, 466, 114.

[3] M. Qhobosheane, M. Santra, P. Zhang, Analyst 2001, 126,

1274.

[4] M. T. Reetz, A. Zonta, J. Simpelkamp, Angew. Chem. Int.

Ed. 1995, 34, 301.

[5] M. T. Reetz, A. Zonta, J. Simpelkamp, W. Konen, Chem.

Commun. 1996, 11, 1397.

[6] S. Brandriss, G. Borchardt, J. Kreuter, S. Margel, React.

Polym. 1995, 25, 111.

[7] M. M. van Schooneveld, E. Vucic, R. Koole, Nano Lett.

2008, 8, 2517.

378 http://www.aem-journal.com � 2010 WILEY-VCH Verlag GmbH & C

[8] H. Zou, S. S. Wu, J. Shen, Chem. Rev. 2008, 108, 3893.

[9] E. Bourgeat-Lami, J. Lang, J. Colloid Interface Sci. 1998,

197. 293.

[10] A. K. Van Helden, J. Colloid Interface Sci. 1981, 81, 354.

[11] A. K. Van Helden, J. Colloid Interface Sci. 1980, 76,

418.

[12] A. Van Blaaderen, A. Vrij, J. Colloid Interface Sci. 1993,

156, 1.

[13] W. Stober, J. Colloid Interface Sci. 1968, 26, 62.

[14] A. P. Philipse, A. Vrij, J. Colloid Interface Sci. 1989, 128,

121.

[15] R. D. Badley, W. T. Ford, F. J. McEnroe, R. A. Assink,

Langmuir 1990, 6, 792.

[16] R. P. Bagwe, L. R. Hilliard, W. Tan, Langmuir 2006, 22,

4357.

[17] C. Beck,W.Hartl, R. Hempelmann,Angew. Chem. Int. Ed.

1999, 38, 1297.

[18] X. J. Zhao, R. Tapec-Dytioco,W.H. Tan, J. Am. Chem. Soc.

2003, 125, 11474.

[19] H. Giesche, J. Eur. Ceram. Soc. 1994, 14, 205.

[20] S. Wong, V. Kitaev, G. A. Ozin, J. Am. Chem. Soc. 2003,

125, 15589.

[21] K. D. Hartlen, A. P. T. Athanasopoulos, V. Kitaev,

Langmuir 2008, 24, 1714.

[22] T. Yokoi, Y. Sakomoto, O. Terasaki, Y. Kubota, T. Okubo,

T. Tatsumi, J. Am. Chem. Soc. 2006, 128, 13664.

[23] T. M. Davis, M. A. Snyder, J. E. Krohn, M. Tsapatsis,

Chem. Mater. 2006, 18, 5814.

[24] M. Nakamura, K. Ishimura, J. Phys. Chem. C 2007, 111,

18892.

[25] M. Stjerndahl, P. Jarvoll, M. Andersson, R. Kohout, R. S.

Duran, J. Phys. Chem. C 2008, 112, 17063.

o. KGaA, Weinheim ADVANCED ENGINEERING MATERIALS 2010, 12, No. 5