Journal of Colloid and Interface Science 346 (2010) 317–323

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

www.elsevier.com/locate/jcis

The synthesis of zinc oxide nanoparticles from zinc acetylacetonate hydrate

and 1-butanol or isobutanol
Gabriela Ambrožič a, Srečo D. Škapin b, Majda Žigon a, Zorica Crnjak Orel a,*
a
National Institute of Chemistry, Hajdrihova 19, SI-1000 Ljubljana, Slovenia
b
Jožef Stefan Institute, Jamova cesta 39, SI-1000 Ljubljana, Slovenia

a r t i c l e i n f o a b s t r a c t

Article history: ZnO nanoparticles of different sizes, from 20 to 200 nm in length, and morphologies, nanorods and coral-
Received 4 December 2009 like structures, were synthesized via a simple one-pot synthesis by refluxing an oversaturated solution of
Accepted 1 March 2010 zinc acetylacetonate hydrate in 1-butanol and isobutanol.
Available online 6 March 2010
On the basis of 1H and 13C NMR experiments, the reactions in both alcohols were found to proceed via
the alcoholytic C–C cleavage of the acetylacetonate ligand, followed by the hydrolytic formation of the
Keywords: reactive Zn–OH intermediate from the water molecules present in the precursor hydrate species and/
ZnO
or those released during the condensation cycle. The zinc acetylacetonate conversion into ZnO in isobu-
Nanostructures
Sol–gel synthesis
tanol is significantly slower than in the case when 1-butanol was used as both the medium and the
Crystal growth reagent.
FE-SEM studies showed that in 1-butanol the growth of the rod-shaped particles occurs via the agglom-
eration of ZnO primary particles that are less than 10 nm in size. The morphology of the particles formed
in the isobutanol is time dependent, with the final coral-like structures developing from initially formed
bundle-like structures.
Ó 2010 Elsevier Inc. All rights reserved.

1. Introduction performing the experiments in a series of alcohols containing dif-

Zinc oxide (ZnO) is one of the most extensively studied materi-
als because of its outstanding optoelectronic properties, with po-
tential applications in many different fields of technology [1–4].
Over the past few years, enormous efforts have been devoted to
exploring synthetic ways of preparing nano- to micrometer-size
ZnO with a well-defined particle shape and a uniform particle size.
Hydrothermal sol–gel chemistry has been one of the most
extensively used in the production of bulk metal oxides [5–7].
The high reactivity of the metal oxide precursors in the water leads
to a rapid growth rate of the ZnO, which makes it difficult to con-
trol the synthesis of nanometer-scale particles in these systems.
The uncontrolled growth of ZnO nanoparticles can be prevented
by performing the conversion of zinc precursors in organic sol-
vents, which act both as the reagents and the media [8–10].
Non-aqueous systems are particularly suitable for studies of the
reaction mechanism since the organic by-products that are formed
in the reaction can be determined using conventional analytical
techniques.
The majority of studies on the influence of the structure of the
alcohol on the morphology and size of the ZnO particles in cases
where alcoholytic cleavages of the precursor are involved rely on
* Corresponding author. Fax: +386 1 4760 300.
E-mail address: zorica.crnjak.orel@ki.si (Z.C. Orel).
ferent alkyl chain lengths [11,12]. In these systems the as-synthe-
sized ZnO particle morphologies change from roughly spherical to
rod-like and their average length increases from 184 to 506 nm
when 1-butanol and 1-decanol are used as the solvents. However,
it should be mentioned that under the described reaction condi-
tions (in an autoclave at 250 °C) the particles obtained in all of
the solvents seemed to be non-uniform and had a wide size
distribution.
Although zinc acetylacetonate as a bidendate chelate complex
provides for more controlled reaction rates in the synthesis of zinc
oxide than the precursors with monodendate ligands, i.e., metal alk-
oxides or metal halides [13,14], only a little information is available
in the current literature. Some of the methods for the preparation of
ZnO from zinc acetylacetonate include CO2 laser decomposition
[15], chemical spray deposition [16] or solution chemistry methods
in alcoholic solutions with the presence of a surfactant [17] or an
inorganic base [18].
Here we present a surfactant- and catalyst-free synthesis of ZnO
particles from an oversaturated solution of zinc acetylacetonate
hydrate (Zn(acac)2xH2O, 1), in the alcohols 1-butanol and isobuta-
nol under reflux conditions.
The various sizes and shapes (rod-like particles and complex
coral-like architectures) of the obtained nano- and submicron-
sized ZnO particles depend on the type of alcohol used, as well
as on the amount of precursor and/or the reaction time. In

0021-9797/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcis.2010.03.001
318 G. Ambrožič et al. / Journal of Colloid and Interface Science 346 (2010) 317–323
particular, in the 1-butanol/1 system, a time- and initial-precursor-
quantity-independent particle shape can be obtained. The chosen
reaction system, 1-butanol/1, enabled us to perform an FE-SEM
in situ monitoring of the formation and growth of the nanoparti-
cles from the zinc precursor, which is of particular importance in
gaining an insight into the dynamics of the formation of inorganic
nanoparticles under defined experimental conditions. The influ-
ence of the non-dissolved zinc precursor on the rate of particle for-
mation was investigated, which, to the best of our knowledge, has
not yet been reported.
Here, we also report on the reproducibility when transferring
the ZnO production from small to larger laboratory quantities, thus
acquiring valuable data for scaling-up the synthesis to pilot ZnO
production.
2. Experimental
2.1. Materials and methods
Zinc acetylacetonate hydrate (powder, 1), anhydrous 1-butanol
(99.8%) and anhydrous isobutanol (P99.0%) were purchased from
Sigma–Aldrich and used in the as-received condition.
2.1.1. Nanoparticle synthesis
In a typical synthesis, 4.6 mmol of 1 was dispersed in 17 ml of
preheated alcohol (the temperature of the oil bath was kept at
130 °C and 120 °C) depending on the alcohol used (1-butanol and
isobutanol, correspondingly). Aliquots of 1.5 ml were taken at de-
fined reaction times (10, 20, 30, 40, 50, 60, 120, 180, 240, 300
and 360 min) and then quenched by cooling to room temperature.
For the FE-SEM, XRD and FTIR studies of the ZnO particle formation
and growth, the milky suspension was filtered (MILIPORE GTTP fil-
ters) and/or centrifuged (13,000 rpm, 20 min) and washed several
times with the alcohol used. The obtained white precipitates were
dried overnight under vacuum at 40 °C. For the mechanistic study,
the supernatant solution was separated from the precipitated ZnO
nanoparticles and analyzed using NMR spectroscopy. The particle
sizes and size distributions were determined from FE-SEM images.
At least 100 crystals were measured for each sample. To avoid er-
rors due to projections of the three-dimensional crystal shape onto
a two-dimensional image, only crystals that were clearly visible
and lying flat on the substrate were considered during the
measurements.
2.2. Characterization
Fourier-transform infrared spectroscopy (FTIR) spectra were re-
corded on a Spectrum One spectrometer (Perkin Elmer Instru-
ments). The samples were prepared in the form of KBr platelets.
The morphology and particle size were investigated by using a
FE-SEM SUPRA 35VP (Carl Zeiss) equipped with an energy-disper-
sive spectrometer (EDXS, Inca 400, Oxford Instruments). X-ray dif-
fraction (XRD) was performed on a Bruker AXS D4 Endeavor
diffractometer with Cu Ka radiation (k = 0.154 nm) and a Sol-X en-
ergy-dispersive detector within the angular range 2H from 20° to
70° with a step size of 0.02° and a collection time of 3 s. Nuclear
magnetic resonance (NMR) analyses were performed on a Varian
VXR 300 NMR spectrometer at 300 MHz for 1H and 75 MHz for
13
C measurements. DMSO-d6 was used as the solvent. Transmis-
sion electron microscopy was employed using a JEOL-2100 micro-
scope operating at 200 kV.
3. Results and discussion
Fig. 1 shows FE-SEM images of as-prepared ZnO nanoparticles
synthesized after different reaction times from oversaturated solu-
tions of 2.3 mmol of 1 in 17 ml of 1-butanol (Fig. 1a and b) and iso-
butanol (Fig. 1c and d).
The particles prepared in 1-butanol after a reaction time of 1 h
are well separated and rod shaped with an average size of approx-
imately 50 nm (Fig. 1a). The magnified part of Fig. 1a shows that
they are aggregates of small elementary particles, with an esti-
mated size of about 5 nm. As the reaction time increases, these
aggregates transform into smooth nanorods with an average
length of approximately 90 nm (Fig. 1b).
The rod-shaped particles were formed from all of the investi-
gated oversaturated 1-butanol solutions containing different initial
amounts of precursor 1 (0.8, 1.5, 2.3, 4.6 and 6.9 mmol in 17 ml of
1-butanol). Further analyses of the FE-SEM data show that as the
mean particle length increases, the size distribution also expands
a great deal. As an example, Fig. 2 shows the length distribution
of the ZnO particles obtained from the oversaturated solution of
0.8 mmol of the precursor 1 in 17 ml of 1-butanol after different
reaction times. The broadening of the particle size distribution
can be attributed to Ostwald ripening, which involves the dissolu-
tion of the fine particles and the growth of the larger particles.
Besides the reaction time, the ZnO particle lengths and the dis-
tributions of the investigated system also depend on the amount of
initial precursor. By analyzing the FE-SEM data of the particles ob-
tained after 1 h of reaction for different initial quantities of 1 (0.8,
1.5, 2.3, 4.6 and 6.9 mmol in 17 ml of 1-butanol), the mean particle
size increased with an increasing amount of precursor from
approximately 20–70 nm.
Unlike the uniform nanoparticle growth in 1-butanol, the mor-
phology of the complex architectures is time dependent in isobuta-
nol systems (Fig. 1c and d). The final coral-like particles in Fig. 1d
develop from the initially formed bundle-like structures (Fig. 1c).
They are polydispersed in size and, although they appear to be sep-
arate, some of them are partially fused together. A magnified part
of the micrograph (Fig. 1d) shows that they consist of a few-nano-
meters-large ZnO crystallites, which form approximately 100-nm-
long and 10-nm-thick nanowires growing on one side of the
agglomerated, densely packed, quasi-spherical ZnO particles.
Fig. 3 shows the XRD patterns of the purified products obtained
from 2.3 mmol of 1 in 17 ml of 1-butanol after 1 h (Fig. 3, pattern a)
and isobutanol after 1 h (Fig. 3, pattern b). All of the diffraction
peaks of the nanoparticles correspond to a wurtzite hexagonal
structure and there are no additional reflections representing 1
or any intermediate structures.
To obtain evidence of the reaction mechanisms of ZnO forma-
tion, we performed 1H and 13C NMR measurements on the final
reaction solutions after the precipitation of the ZnO. Fig. 4 shows
the 13C NMR spectra of the solutions obtained after 5 h of refluxing
2.3 mmol of 1 in 17 ml of 1-butanol (Fig. 4a) and isobutanol
(Fig. 4b). In addition to the alcohol signals (S) and the small inten-
sity signals of unreacted 1, the spectra consist of the peaks of buty-
lacetate (2a, Fig. 4a), isobutylacetate (2b, Fig. 4b) and acetone (3).
For the reactions in both 1-butanol and isobutanol we propose a
single-pathway reaction mechanism, which is based on the cleav-
age of the C–C bond of the acetylacetonate ligand in 1 (Scheme 1).
The most abundant alcohol molecules attack the electron-deficient
carbonyl C-atom of the acetylacetonate ligand in 1 activated by the
electropositive zinc precursor center. The cleavage of the C–C bond
leads to butyl acetate (2a) and isobutyl acetate (2b), correspond-
ingly. We assume that the next hydrolysis step proceeds via the
nucleophilic attack of the water molecule present in the precursor
hydrate species on the zinc center, the proton-transfer to the eno-
late, and the subsequent cleavage of the acetone (3). This would re-
sult in the formation of Zn–OH species, which represent the
monomeric units for the nucleation of ZnO nanocrystals, which
proceeds through the formation of Zn–O–Zn bonds via the conden-
sation reaction [10].
 G. Ambrožič et al. / Journal of Colloid and Interface Science 346 (2010) 317–323 319

Fig. 1. FE-SEM images of ZnO nanoparticles synthesized from the oversaturated solution of 2.3 mmol of the precursor 1 in 17 ml of 1-butanol after 1 h (a) and 5 h (b),
isobutanol after 1 h (c) and 5 h (d). The insets show the magnified parts of a, b, c and d.

Fig. 2. Length distribution of ZnO particles obtained from the oversaturated solution of 0.8 mmol of the precursor 1 in 17 ml of 1-butanol after different reaction times.

This mechanism can be rationalized by considering the eventual
competitive reaction pathway. If, supposing, the reaction were to
proceed through the ligand-exchange mechanism, there would
be a formation of zinc alkoxide and/or zinc hydroxide intermedi-
ates and acetylacetone. Since we were not able to detect acetylace-
tone in the supernatant solution, we believe that the initial
dissociation of the Zn–O bond in 1 by alcohol or water molecules
does not occur. This can be interpreted as a consequence of the
strong coordination of the bidentate acetylacetonate ligand to the
metal center [19]. As a reference, we performed simple experi-
ments of refluxing acetylacetone in 1-butanol and isobutanol,
which showed that acetylacetone is inactive toward the alcoholy-
sis, even after 24 h of reaction. This additionally confirms that
the alcoholysis reaction of 1 does not proceed through the
320 G. Ambrožič et al. / Journal of Colloid and Interface Science 346 (2010) 317–323

metal chelate, in which the acetylacetonate ligand is activated to-

(101) ward the attack of the protic solvent.
(100) In accordance with the proposed mechanism, Garnweitner [20]
(002)
and Pinna et al. [21] reported on the C–C bond cleavage in the syn-
(110) (103)
(102) (112) thesis of metal oxides from the corresponding metal acetylaceto-
Intensity (a.u.)

nates and protic solvents under autoclave reaction conditions.

b However, in the mentioned systems the decompositions of the ace-
tylacetonate ligand occur simultaneously through several reaction
pathways, all strictly involving solvent molecules as nucleophiles.
We should point out that the reaction rate of ZnO formation in
isobutanol is slower: after 5 h of reflux, the quantity of unreacted
zinc precursor 1 is significantly higher than that in 1-butanol for
a the same amounts of 1 (0.8, 2.3 and 6.9 mmol in 17 ml of isobuta-
nol). Besides the relatively small difference in the boiling points
20 30 40 50 60 70 (DT = 9 °C) and dielectric constants (lower solubility of the precur-
sor 1), we also ascribe this phenomenon to the lower reactivity of
2 Θ (o)
the isobutanol, whose hydroxyl group is sterically hindered from
Fig. 3. X-ray diffractograms of ZnO, by refluxing the oversaturated solution of carrying out any nucleophilic attack on the precursor 1 carbonyl
2.3 mmol of the precursor 1 in 17 ml of 1-butanol after 1 h (a) and isobutanol after group.
1 h (b). To gain a better understanding of the evolution mechanisms of
the ZnO particles from an oversaturated solution of zinc precursor
formation of acetylacetone and its subsequent decomposition into 1 in 1-butanol, the growing process was monitored by FTIR and FE-
alkylacetate and acetone, but rather that the reaction species is the SEM for different reaction times.

13
Fig. 4. C NMR spectra of the reaction solutions after ZnO precipitation in 1-butanol (a) and isobutanol (b).
 G. Ambrožič et al. / Journal of Colloid and Interface Science 346 (2010) 317–323
Scheme 1. Proposed mechanism of Zn–O–Zn bond formation from the precursor 1
and 1-butanol or isobutanol.
The FTIR spectroscopy (Fig. 5) of the precipitates obtained from
4.6 mmol of precursor 1 in 17 ml of 1-butanol at different reaction
times showed that after 10 min of reflux the sample mainly con-
sists of the precursor 1 (Fig. 5, spectrum b), where the initial for-
mation of ZnO is represented by an additional band centered at
470 cm1 (tZn–O stretching vibration). After 1 h the precursor 1
had completely disappeared (Fig. 5, spectrum c). The as-synthe-
sized ZnO nanoparticles obtained after 3 h of reaction are repre-
sented in spectrum d (Fig. 5, spectrum d).
In accordance with the FTIR spectroscopy, FE-SEM observations
(Fig. 6) of the precipitate obtained after 10 min of reaction showed
the initial formation of small, quasi-spherically shaped ZnO parti-
cles close to the surface of the solid (undissolved) precursor 1
(Fig. 6a). The particle size estimated from the FESEM micrographs
is approximately 10 nm. During the early stages of the reaction,
somewhat larger particles are formed, while the rest of the solid
precursor 1 is continuously consumed to form new ZnO nanopar-
ticles (Fig. 6b). After 1 h, well-separated, rod-shaped ZnO aggre-
gates with a mean length of 50 nm are formed (Fig. 6c).
In order to simulate the particle formation in a homogeneous
solution under the same reflux conditions, we performed the
experiment by rapidly removing the non-dissolved precursor 1
immediately after adding it to boiling 1-butanol.
After 6 h of reaction and several washings, we isolated particles
up to 20 nm in size (Fig. 7a). The HR-TEM images (Fig. 7b) of the re-
dispersed particles showed that the sample consists of crystallized
Fig. 5. FTIR spectra of the precursor 1 (reference spectrum a), and the precipitates
obtained from the oversaturated solution of 4.6 mmol of the precursor 1 in 17 ml of
1-butanol after 10 min (b), 1 h (c) and 3 h of reaction (d).
321
cores in the amorphous matrix. The presence of the ZnO phase was
confirmed by the corresponding SAED pattern (inset in Fig. 7b).
On the basis of the performed experiments, we suggest that the
formation of the ZnO nanocrystals in the oversaturated system is
due to the limited solubility of the precursor 1. This provides for
a constant supply of precursor in the solution, and the formation
of the reactive species (i.e., monomers) can proceed continuously
under steady-state conditions. Since the homogeneous solution
does not have a constant supply of ‘‘new” precursor 1 molecules,
the reaction turned out to be very slow. Even after six reaction
hours, well-crystallized ZnO particles did not form.
During very early stages of the reaction in the oversaturated
system, steady-state conditions are not likely to be reached, so
the nucleation process in the solution is difficult to begin. How-
ever, the local increase in the concentration of ZnO monomers
around the solid precursor (local supersaturation) could lead to
an enhanced rate of formation of ZnO primary particles near the
precursor 1 particle surface (Fig. 6a). These particles grow until
reaching a critical size, when they diffuse from the surface of the
solid precursor 1 (Fig. 6b). The obtained small nanocrystals form
large aggregates with a defined rod-like morphology (Fig. 6c). With
time the ‘‘bounded” nanocrystals gradually coalesce within the
aggregate into smooth nanorods [22], as shown in Fig. 1b.
When comparing the results obtained in both 1-butanol and
isobutanol, it is evident that the chemical structure and physical
properties of the alcohol determine the morphology and size of
ZnO particles as well as the rate of their formation. With the regard
to the particle nucleation and growth, the nature of the alcohol
should also be of key importance. However, the comparison of
these processes in different media is quite challenging and it can-
not be simplified since it is necessary to take into account not only
the difference in the structure of the alcohol used but also various
other parameters (density and dielectric constants at boiling tem-
peratures as well as the reaction temperature itself).
There is, however, still an underlying question of how to scale-up
the production of well-separated nanoparticles with a controllable
size and shape. We performed the synthesis of ZnO from the precur-
sor 1 in 1-butanol under the same experimental conditions, but using
96 mmol of the precursor 1 in 340 ml of 1-butanol. The obtained ZnO
nanoparticles have the same morphology (rod-like) and a size com-
parable to the ones prepared in the smaller batch at the same reaction
time, meaning that our approach is reproducible and suitable for
medium-to-large scale ZnO nanoparticle production.
4. Conclusion
In summary, we report here on a synthesis and the underlying
reaction mechanism of uniform ZnO nanoparticle formation ob-
tained by refluxing an oversaturated solution of Zn(acac)2xH2O
(1) in 1-butanol and isobutanol. The single-pathway reaction is
suggested to proceed via the nucleophilic attack of the alcohol on
the carbonyl acetylacetonate group of the precursor 1, followed
by acetone release, which is promoted by water molecules present
in the precursor and/or formed during the polycondensation of the
Zn–OH species.
The formation and growth of the ZnO rod-like particles in the
oversaturated 1-butanol solutions were studied in detail. In con-
trast to the homogeneous solution, the enhanced decomposition
rate of the dissolved precursor in the oversaturated systems is sug-
gested to be due to steady-state conditions providing a constant
supply of precursor in the solution and, consequently, the continu-
ous formation of reactive species (i.e., monomers). The enhanced
rate of formation of the ZnO elementary particles near the precur-
sor surface might be due to the local supersaturation of the
reactive monomer species around the solid precursor. Once
322 G. Ambrožič et al. / Journal of Colloid and Interface Science 346 (2010) 317–323
Fig. 6. FE-SEM micrographs of the particles suspended in 1-butanol at different reaction times (initial amount of the precursor 1 4.6 mmol in 17 ml of 1-butanol): 10 min (a),
20 min (b) and 60 min (c). The insets show the magnified parts of a, b and c.
Fig. 7. FE-SEM (a) and HR-TEM (b) images of clusters formed from the homogeneous solution of the precursor 1 in 1-butanol.
formed, the ZnO nanocrystals self-assemble into well-separated,
pearl-chain aggregates with a mean length of 20–70 nm. The ZnO
morphology is supposedly tailored by the interface interactions
of the ZnO with 1-butanol and the organic species formed during
the reaction, while the size and distribution of the ZnO nanorods
are governed by the initial amount of the precursor 1 in the 1-buta-
nol oversaturated solution. The analogous single-path reaction
mechanism was also confirmed for an isobutanol-mediated reac-
tion. However, the reaction rate for isobutanol is lower and the ob-
tained coral-like particles are polydispersed and their morphology
is time dependent.
By comparing 1-butanol with isobutanol as a medium for the
formation of ZnO nanoparticles from the precursor 1, it can be con-
cluded that 1-butanol is the most convenient solvent as it enables
time-independent product-shape control, control over the particle
size and dispersity, and the preparation of a large quantity ZnO
nanoparticles, which are the crucial prerequisites for the reproduc-
ible preparation of metal oxides in organic solvents.
Acknowledgments
This work was funded by the Ministry of Higher Education, Sci-
ence and Technology (Project Novapol of MNT-ERA.NET) and the
Slovenian Research Agency (research programs P2-0145 and P2-
030). We appreciate the help of Ms. Ines Bracko in performing
the TEM analysis.
References
[1] Y. Nakayama, P.J. Pauzauskie, A. Radenovic, R.M. Onorato, R.J. Saykally, J.
Liphardt, P.D. Yang, Nature 447 (2007) 1098.
[2] C. Klingshirn, ChemPhysChem. 8 (2007) 782.
[3] Z.L. Wang, Adv. Mater. 19 (2007) 889.
[4] M. Bitenc, P. Podbršček, Z. Crnjak Orel, M.A. Cleveland, J.A. Paramo, R.M. Peters,
Y.M. Strzhemechny, Cryst. Growth Des. 9 (2009) 997.
[5] K.S. Krishna, U. Mansoori, N.R. Selvi, M. Eswaramoothy, Angew. Chem., Int. Ed.
46 (2007) 5962.
[6] M. Bitenc, M. Marinšek, Z. Crnjak Orel, J. Eur. Ceram. Soc. 28 (2008) 2915.
[7] M. Li, H. Bala, X. Lv, X. Ma, F. Sun, L. Tang, Z. Wang, Mater. Lett. 61 (2007) 690.
[8] M. Niederberger, Acc. Chem. Res. 40 (2007) 793.
[9] M.M. Demir, R. Muñoz-Espí, I. Lieberwirth, G. Wegner, J. Mater. Chem. 16
(2006) 2940.
[10] X. Zhong, Y. Feng, Y. Zhang, I. Lieberwirth, W. Knoll, Small 3 (2007) 1194.
[11] P. Praserthdam, P.L. Silveston, O. Mekasuwandumrong, V. Pavarajarn, J.
Phungphadung, A. Kanyanucharat, Cryst. Growth Des. 3 (2003) 215.
[12] P. Tonto, O. Mekasuwandumrong, S. Phantanasri, V. Pavarajarn, P.
Praserthdam, Ceram. Int. 3 (2008) 57.
[13] T. Tsuchiya, A. Watanabe, Y. Imai, Jpn. J. Appl. Phys. 2 (38) (1999) L1112.
[14] A.L. Willis, Z. Chen, J. He, Y. Zhu, N.J. Turro, S.J. Ó0 Brien, Nanomater. 12 (2007)
14858.
[15] C. Fauteux, R. Longtin, J. Pegna, D. Therriault, Inorg. Chem. 46 (2007) 11036.
 G. Ambrožič et al. / Journal of Colloid and Interface Science 346 (2010) 317–323 323

[16] A. Maldonado, M. de la Luz Olvera, S. Tirado Guerra, R. Asomoza, Sol. Energy
Mater. Sol. Cells 82 (2004) 75.
[17] J.F. Liu, Y.Y. Bei, H.P. Wu, D. Shen, J.Z. Gong, X.G. Li, Y.W. Wang, N.P. Jiang, J.Z.
Jiang, Mater. Lett. 61 (2007) 2837.
[18] M. Iwasaki, Y. Inubushi, S. Ito, J. Mater. Sci. Lett. 16 (1997) 1503.
[19] S. Doeuff, M. Henry, C. Sanchez, J. Livage, J. Non-Cryst. Solids 89 (1987) 206.
[20] G. Garnweitner, Nonaqueous Synthesis of Transition Metal-oxide
Nanoparticles and their Formation Mechanism, Max-Planck-Institute für
Kolloid- und Grenzflächenforschung, Postdam, 2005.
[21] N. Pinna, G. Garnweitner, M. Antonietti, M. Niederberger, J. Am. Chem. Soc. 127
(2005) 5608.
[22] M. Ethayaraja, R. Bandyopadhyaya, Langmuir 23 (2007) 6418.