From hot-injection synthesis to heating-up synthesis of cobalt nanoparticles: Observation of kinetically controllable nucleation

Article abstract of DOI:10.1002/anie.201005600

Almost equal: Particle nucleation in the widely studied hot-injection synthesis of magnetic cobalt nanoparticles (see TEM image) was found to take place well after the injection and to be temperature-controlled. This insight into particle nucleation led to a technologically relevant heating-up synthesis of nearly monodisperse cobalt nanoparticles on a multigram scale. The particle size could be tuned simply by adjusting the heating rate.

Full text of DOI:10.1002/anie.201005600

Communications
DOI: 10.1002/anie.201005600
Magnetic Nanoparticles
From Hot-Injection Synthesis to Heating-Up Synthesis of Cobalt
Nanoparticles: Observation of Kinetically Controllable Nucleation**
Jaakko V. I. Timonen,* Eira T. Seppꢀlꢀ, Olli Ikkala, and Robin H. A. Ras*
Monodisperse nanoparticles of well-defined size and shape
are required in several emerging applications, which take
advantage of their size-dependent properties such as the
superparamagnetic limit in the case of magnetic nanoparti-
cles.[1,2] Accurate tuning of the nanoparticle size and shape
requires understanding of the mechanisms involved in
particle nucleation and growth.[3–5] In spite of extensive
ongoing research, these mechanisms are still not fully under-
stood owing to their complexity and interplay. Moreover, the
current small-scale synthesis methods, such as the hot-
injection method, can be difficult to scale to industrially
relevant levels. Hence, more suitable methods are sought.[6–19]
Herein, we revisit a widely studied hot-injection synthesis
of monodisperse cobalt nanoparticles[20–26] and show that the
particle nucleation differs from what is expected for a hot-
injection synthesis. Evidence is given that the particles
nucleate several tens of seconds or a few minutes after the
injection, depending delicately on how the reaction temper-
ature is controlled after the sudden temperature drop caused
by the injection. The delayed nucleation is followed by a
period during which the cobalt precursor decomposes endo-
thermically, the temperature drops, carbon monoxide evolves,
and the nuclei rapidly grow into mature nanoparticles.
Particle growth after the endothermic period is negligible,
and we show that the final particle size is determined by the
rate of temperature increase after the injection-induced
temperature drop. A rapid increase results in a higher peak
temperature before the endothermic period and more nuclei,
hence smaller particles, in comparison to the case of a slower
rate of temperature increase. The contribution of the injection
to particle nucleation seems minor, and it is shown that
injection can be replaced entirely by an accurately controlled
heating up of the solution containing all reagents (including
the cobalt precursor) from room temperature to the nucle-
ation temperature. This synthetic method, which is often
termed either “non-injection synthesis”[15,16] or “heating-up
synthesis”,[3,11] results in nanoparticles that are nearly iden-
tical to those made by the hot-injection method.
We synthesized cobalt nanoparticles by injecting dicobalt
octacarbonyl, [Co2(CO)8], dissolved in a small amount of
ortho-dichlorobenzene (o-DCB, b.p. 1818C) into a solution of
oleic acid and trioctylphosphine oxide (TOPO) in o-DCB at
reflux.[20] The injection led to an immediate temperature drop
of several tens of degrees, which is characteristic of the hot-
injection method in general.[4] It has been shown that
[Co2(CO)8] undergoes partial decarbonylation during the
injection to form gaseous carbon monoxide and intermediate
cobalt carbonyl species (e.g. tetracobalt dodecacarbonyl,
[Co4(CO)12], and cobalt tetracarbonyl, [Co(CO)4]) in the
solution phase; these species then further decompose more
slowly to cobalt atoms.[25,27] It has been shown that both
maintaining the lower temperature and letting the temper-
ature recover to the reflux temperature after the injection can
lead to monodisperse nanoparticles.[20–23] In this study, we
concentrated on the latter approach and studied for the first
time in detail the kinetics of the temperature recovery to
reflux. The recovery rate can be conveniently controlled by
tuning the rate of heat transfer from the heat bath to the
reaction medium, for example, by using an oil bath at
different temperatures or an electric heating mantle with
different heating powers.
A typical development in the reaction temperature after
the injection is shown in Figure 1 for the hot-injection
synthesis HI1 (see the Experimental Section for a complete
list of syntheses with details). The temperature dropped from
180 to 1438C during the injection, after which it started to
recover, as heat was being transferred from the heat bath to
the reaction medium. In contrast to the expected continuous
increase until the reflux temperature was reached, one minute
after the injection we observed a characteristic endothermic
period during which the reaction temperature dropped
despite continuous heating. Interestingly, the peak temper-
ature (1748C) reached just before the endothermic period
was very close to the temperature prior to injection (1808C).
Vigorous evolution of carbon monoxide during the endother-
mic period indicated decomposition of the cobalt carbonyl
species and release of cobalt atoms.[27] Further evolution of
carbon monoxide after the endothermic period was negli-
gible, even when the reflux temperature was reached. This
observation indicated that nearly all cobalt carbonyl species
had decomposed during the endothermic period, which was
[*] J. V. I. Timonen, Prof. Dr. O. Ikkala, Dr. R. H. A. Ras
Department of Applied Physics, Aalto University
(formerly Helsinki University of Technology)
P.O. Box 15100, FI-02150 Espoo (Finland)
E-mail: jaakko.timonen@tkk.fi
robin.ras@tkk.fi
Homepage: http://tfy.tkk.fi/molmat/
Dr. E. T. Seppꢀlꢀ
Nokia Research Center
Itꢀmerenkatu 11–13, 00180 Helsinki (Finland)
[**] Funding from Nokia Research Center, the Finnish Funding Agency
for Technology and Innovation (Tekes), and the Academy of Finland
is acknowledged. Electron microscopy images were obtained at the
Nanomicroscopy Center at Aalto University. Dr. Christoffer Johans
(Aalto University), Dr. Markku Oksanen (Nokia Research Center),
and Dr. Andreas Walther (Aalto University) are acknowledged for
fruitful discussions.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201005600.
2080
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 2080 –2084

Figure 1. Typical double-drop of the reaction temperature in the hot-
injection synthesis HI1: the second temperature drop is an endother-
mic period between 50 and 65 s. Temperature readings are marked as
filled circles and particle diameters as observed by TEM as open
circles (see the Supporting Information for TEM images and particle-
size distributions).
verified also by Fourier transform infrared spectroscopy
(FTIR). Small aliquots were extracted during the synthesis
and cast immediately on carbon-coated transmission electron
microscopy (TEM) grids. TEM analysis indicated that very
few nanoparticles existed in the reaction medium before the
peak temperature, whereas after that point the medium was
concentrated with nanoparticles. Determination of the aver-
age particle diameter as a function of reaction time and
temperature indicated that the main growth took place during
the endothermic period, and that growth thereafter was
negligible (Figure 1). This result is in agreement with the
observed decomposition of the precursor during the endo-
thermic period. After that period, the particles were consid-
ered to be full-grown; further annealing at 1808C or storage at
room temperature resulted in negligible increase in particle
size (see the Supporting Information).
Figure 2. Effect of the temperature-recovery rate on the size distribu-
tion of the nanoparticles. a) Temperature histories. HI2: rapid temper-
ature recovery, HI3: medium-rate recovery, HI4: slow recovery. b) Cor-
responding TEM images of full-grown particles (scale bars are 20 nm).
c) Corresponding particle size distributions with Gaussian fits.
from HI2–HI4 was also dependent on the recovery rate. The
trend of an increasing particle size with a decreasing temper-
ature-recovery rate is clear from the TEM images of the full-
grown nanoparticles (Figure 2b) and the corresponding
particle size distributions (Figure 2c).
The effect of the temperature-recovery rate on the peak
temperature and on the resulting nanoparticles was inves-
tigated in syntheses HI2–HI4. These syntheses were similar to
HI1, except that the temperature-recovery rate after the
injection was varied by tuning heat transfer to the reaction
medium (HI2: rapid temperature recovery owing to a high
temperature of 2158C of the heating oil bath, HI3: medium-
rate recovery owing to a lower temperature of 1958C of the
heating oil bath, HI4: slow recovery owing to the very low
heating power of an electric heating mantle). The temper-
ature histories of these syntheses (Figure 2a) featured
qualitatively similar double-drop behavior as observed in
HI1. However, a decrease in the heating rate led to a decrease
in the peak temperature reached before the endothermic
period and also to a decrease in the magnitude of the
temperature drop. Furthermore, the endothermic period
shifted toward later times as the recovery rate decreased.
Despite these differences, the growth pattern of the nano-
particles was qualitatively the same as in HI1. That is, the
particles appeared very close to the peak temperature, and
the main growth took place during the endothermic period.
Importantly, the average diameter of the full-grown particles
It has been suggested that nucleation in the studied
synthesis takes place nearly instantaneously during the
injection.[21] Our results from syntheses HI1–HI4 suggest
another mechanism. The nearly complete absence of particles
before the peak temperature and the rapid appearance of
particles after the peak temperature already suggest that the
nucleation takes place near the peak temperature instead of
the injection. Unfortunately, although the direct measure-
ment of nuclei concentration and size distribution is well-
established for optically active quantum-dot nanoparticles, it
is more difficult for particles that are not optically active, such
as cobalt nanoparticles.[11,28–31] Therefore, we used an indirect
method to resolve the final concentration of the full-grown
nanoparticles. The concentration can be calculated from the
amount of carbonyl present initially and the average size of
the nanoparticles at the end of the synthesis if the conversion
of cobalt into nanoparticles (crystallization yield) is known.
We determined the crystallization yield to be approximately
90% in these syntheses (see the Supporting Information);
hence, the concentrations of the full-grown nanoparticles and
also the nuclei are approximately 24.9 mm (HI2), 5.7 mm (HI3),
Angew. Chem. Int. Ed. 2011, 50, 2080 –2084
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2081

Communications
and 3.6 mm (HI4). This variation in the concentration cannot
high-resolution TEM, electron diffraction, and magnetometry
(see the Supporting Information), which indicated that
the particles were superparamagnetic and single-crystalline
e-cobalt, like those formed by the hot-injection synthesis.
To demonstrate that nucleation in the heating-up syn-
thesis can be controlled by temperature kinetics as in the hot-
injection method, we performed two heating-up syntheses
with different heating rates. In HU2, the flask containing the
precursor reagents was heated rapidly by immersing it in an
oil bath at 2158C, whereas in HU3, the heating rate was
slower as a result of a lower temperature of 1958C of the oil
bath (Figure 4a). Both syntheses exhibited a temperature
be explained by random variation in the number of nuclei
formed during the injection since special care was taken to
make the injections identical: the injection was always started
at 180.0 Æ 0.58C, its duration was intentionally stretched to
5 s, and the lowest temperature reached was always observed
to be 143.0 Æ 1.08C. No difference in the final particles was
observed if the injection was carried out more rapidly (in less
than 1 s). Moreover, the temperature behavior during the first
15 s after the injection was almost the same in all syntheses as
a result of dominant heat transfer from the reaction-vessel
walls rather than from the heat bath (Figure 2a). Hence, the
difference in the final particle size stems from reactions taking
place after the first 15 s and before the end of the endothermic
period (after which growth is terminated).
In some hot-injection syntheses of other nanoparticles, the
number of nuclei formed during the injection is initially high,
but afterwards some nuclei dissolve back to monomers. Thus,
the nuclei concentration is decreased, and the remaining
nuclei can grow into larger particles.[28–30] However, the partial
dissolution of nuclei is not the reason for the formation of
differently sized nanoparticles at different temperature-
recovery rates in the synthesis under investigation: the hot-
injection synthesis HI2 was successfully converted into a
heating-up synthesis HU1, in which the injection of the
carbonyl compound was replaced by mixing of the carbonyl
compound directly with all the other reagents at room
temperature and heating of the solution to reflux. The heating
rate was adjusted to be very close to that of HI2 near the peak
temperature, and a very similar temperature drop resulted
(Figure 3a). Furthermore, the particles formed in the heating-
Figure 4. Temperature control of nanoparticle size in the heating-up
synthesis. a) Temperature histories of HU2 (rapid heating) and HU3
(slow heating). b) Corresponding particle size distributions with Gaus-
sian fits. Corresponding TEM images of full-grown particles are shown
on the right. Scale bars are 20 nm.
peak near 1708C, which was followed by an endothermic
period. The diameters of the full-grown nanoparticles showed
a standard deviation of only 7% in both cases. However, the
particles produced during more rapid heating were consid-
erably smaller than those produced during slower heating
(Figure 4b). Because of the similar crystallization yield in
both cases, more nuclei had to form under the conditions of
more rapid heating.
The experimental observations can be discussed in terms
of nucleation theory to shed more light on the formation of
the nanoparticles, even though the exact chemical composi-
tion of the particle-forming monomer often cannot be
resolved.[5,11] Nucleation becomes thermodynamically
increasingly favorable as the concentration of the monomers
increases beyond the saturation limit.[3,4] In an ideal hot-
injection synthesis, the monomer concentration greatly
exceeds the saturation limit during the injection, which results
in very rapid burst nucleation. Our observations suggest that
during injection in the studied synthesis, the saturation limit is
not exceeded enough for burst nucleation to happen. Instead,
the experimental results can be understood if the injection
Figure 3. Comparison of the hot-injection and heating-up methods.
a) Temperature histories of the HI2 and HU1 syntheses under identical
conditions (same reagents and heating rate). b) TEM image of full-
grown particles from HU1 (scale bar is 100 nm; inset shows diffracto-
gram of the image).
up synthesis appeared similarly in the vicinity of the peak
temperature and had approximately the same size of 5 nm as
in the hot-injection synthesis (Figure 3b). This similarity
between the hot-injection and heating-up methods indicates
that the injection has only a minor effect on nucleation, since
it can be omitted without a notable change in the nano-
particles. If there was nucleation during the injection, practi-
cally all nuclei would redissolve after the injection and reform
after a delay just before the endothermic period. The particles
synthesized by the heating-up method were characterized by
2082 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 2080 –2084

only causes a preliminary increase in the monomer concen-
tration, which continues to increase after the injection as the
carbonyl species decompose (which is observed as slow
evolution of carbon monoxide before the peak temperature
and which has also been shown by in situ Fourier transform
infrared spectroscopy (FTIR)[25]). When the monomer con-
centration reaches the burst-nucleation limit, the particles
nucleate and grow rapidly (as demonstrated in Figure 1)
along with more intensive evolution of carbon monoxide (as
also observed by FTIR[25]). Similar delayed nucleation after
precursor injection has been observed in the synthesis of iron
oxide nanodisks,[32] and a similar gradual increase in monomer
concentration during the heating up of iron oleate has been
shown to lead to the nucleation of monodisperse iron oxide
nanoparticles.[11]
Both the peak temperature around which nucleation takes
place and the number of nuclei formed increase with the
heating rate. This correlation may be understood by consid-
ering the rate of monomer formation. Since the decomposi-
tion of cobalt carbonyl complexes is more rapid at higher
temperatures,[27] the burst-nucleation limit is reached in less
time under conditions of rapid heating (HI2 and HU2), and
hence the temperature at the nucleation point is higher than
under conditions of slower heating (HI3,4 and HU3). A
higher temperature during the nucleation results in more
rapid monomer formation, which partially compensates for
heating-up synthesis to produce nanoparticles (see the
Supporting Information). Finally, even though the injection
was shown to be less important than previously suggested, at
least in the case where the temperature is raised back to reflux
after the injection, the injection may still have significance
since precursor decomposition and complexation with sur-
factants may proceed differently in the heating-up synthesis
compared to the hot-injection synthesis. Such a difference
may explain why the heating-up method results in somewhat
broader polydispersity than the hot-injection method.
In conclusion, it has been shown that the number of nuclei
formed in the hot-injection synthesis of cobalt nanoparticles
depends more on temperature kinetics after the injection than
on the injection itself. We suggest that the injection leads to
supersaturation that is not high enough to cause burst
nucleation, and hence the nucleation is delayed until
enough monomers are created from the decomposing pre-
cursor. The number of nuclei formed in the delayed nucle-
ation can be controlled by kinetic tuning of the temperature at
which the nucleation takes place. This insight led to a
technologically relevant heating-up synthesis of nearly mono-
disperse cobalt nanoparticles that is readily scalable to a
multigram level and in which the particle size can be
controlled simply by the heating rate.
the monomers consumed in the nucleation and thus enables Experimental Section
Hot-injection syntheses HI1–HI4: [Co2(CO)8] (1080 mg) dissolved in
more nuclei to be formed before growth takes over. Another
contribution to the increased number of nuclei comes from
the nucleation process itself, which is more rapid at higher
temperatures.[5] The temperature drop due to the endother-
micity of the decomposition of cobalt carbonyl complexes can
also play a role in particle formation, since a decrease in
temperature can help quench the nucleation.
Significant effort has been made recently in the develop-
ment of techniques for the large-scale synthesis of mono-
disperse nanoparticles, such as different heating-up[6,9–19] and
pressure-drop methods.[7,8] One advantage of the presented
heating-up synthesis is that it can be readily scaled to
multigram quantities (see the Supporting Information for a
one-pot heating-up synthesis with a yield of about 2 g of
particles). Furthermore, the heating-up synthesis may enable
the development of an industrially relevant continuous-flow
process in which the temperature of the precursor solution is
increased in a controlled way to the nucleation temperature
while the evolving carbon monoxide is removed from the
system.
o-DCB (6 mL) was injected into a solution of TOPO (200 mg) and
oleic acid (360 mg, 0.4 mL) in o-DCB (24 mL) at 180 Æ 0.58C under
N2. Postinjection temperature recovery was tuned by adjusting heat
transfer from the heat bath to the reaction medium to either rapid
(HI1 and HI2: an oil bath with efficient heat transfer and a high set
temperature of 2158C), medium (HI3: as HI1 and HI2, but with a
lower set temperature of 1958C), or slow (HI4: electric heating
mantle with reduced heat transfer).
Heating-up synthesis HU1: As for HI1 and HI2, except that all
reagents were mixed at room temperature, followed by heating to
reflux by immersion of the flask in the oil bath. Heating-up syntheses
HU2 and HU3: A solution of [Co2(CO)8] (1080 mg), TOPO (300 mg),
and oleic acid (270 mg, 0.3 mL) in o-DCB (30 mL) was heated under
N2 to reflux either rapidly (HU2: by immersion of the flask in an oil
bath at 2158C) or slowly (HU3: by immersion of the flask in an oil
bath at 1958C).
See the Supporting Information for synthesis details.
Received: September 7, 2010
Revised: December 7, 2010
Published online: January 26, 2011
Even though the hot-injection synthesis of cobalt nano-
particles has been investigated intensively, no detailed
information on the temperature development and its effect
on the nanoparticles have been provided previously. Kinetic
control of the nucleation may explain why differently sized
particles have been obtained under otherwise identical
conditions (same injection temperature and reagents).[20–23,33]
The kinetics of carbon monoxide formation and removal are
also relevant, but have been studied less.[7,8] They may play
some role, as carbon monoxide controls metal-nanostructure
growth.[34] We also demonstrated that other types of cobalt
carbonyl complexes, such as [Co4(CO)12], can be used in the
Keywords: cobalt · hot-injection method · nanoparticles ·
nucleation · synthetic methods
.
[1] Z. Nie, A. Petukhova, E. Kumacheva, Nature Nanotech. 2010, 5,
15.
[2] V. Skumryev, S. Stoyanov, Y. Zhang, G. Hadjipanayis, D. Givord,
J. Nogues, Nature 2003, 423, 850.
[3] J. Park, J. Joo, S. G. Kwon, Y. Jang, T. Hyeon, Angew. Chem.
2007, 119, 4714; Angew. Chem. Int. Ed. 2007, 46, 4630.
[4] C. de Mello Donegꢀ, P. Liljeroth, D. Vanmaekelbergh, Small
2005, 1, 1152.
Angew. Chem. Int. Ed. 2011, 50, 2080 –2084
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 2083

Communications
[5] E. V. Shevchenko, D. V. Talapin, H. Schnablegger, A. Kornow-
[19] Y. Chen, E. Johnson, X. Peng, J. Am. Chem. Soc. 2007, 129,
10937.
[20] V. F. Puntes, K. M. Krishnan, A. P. Alivisatos, Science 2001, 291,
2115.
[21] V. F. Puntes, D. Zanchet, C. K. Erdonmez, A. P. Alivisatos, J.
Am. Chem. Soc. 2002, 124, 12874.
[22] A. C. S. Samia, K. Hyzer, J. A. Schlueter, C.-J. Qin, J. S. Jiang,
S. D. Bader, X.-M. Lin, J. Am. Chem. Soc. 2005, 127, 4126.
[23] Y. Bao, W. An, C. H. Turner, K. M. Krishnan, Langmuir 2010,
26, 478.
[24] G. Cheng, C. L. Dennis, R. D. Shull, A. R. H. Walker, Langmuir
2007, 23, 11740.
[25] A. Lagunas, C. Jimeno, D. Font, L. Solꢄ, M. A. Pericꢄs,
Langmuir 2006, 22, 3823.
[26] G. Cheng, J. D. Carter, T. Guo, Chem. Phys. Lett. 2004, 400, 122.
[27] R. Tannenbaum, G. Bor, J. Organomet. Chem. 1999, 586, 18.
[28] C. R. Bullen, P. Mulvaney, Nano Lett. 2004, 4, 2303.
[29] J. van Embden, P. Mulvaney, Langmuir 2005, 21, 10226.
[30] M. Tiemann, ꢁ. Weiß, J. Hartikainen, F. Marlow, M. Lindꢅn,
ChemPhysChem 2005, 6, 2113.
ski, ꢁ. Festin, P. Svedlindh, M. Haase, H. Weller, J. Am. Chem.
Soc. 2003, 125, 9090.
[6] N. A. D. Burke, H. D. H. Stꢂver, F. P. Dawson, Chem. Mater.
2002, 14, 4752.
[7] C. Johans, M. Pohjakallio, M. Ijꢃs, Y. Ge, K. Kontturi, Colloids
Surf. A 2008, 330, 14.
[8] M. Huuppola, N. Doan, K. Kontturi, C. Johans, J. Colloid
Interface Sci. 2010, 344, 292.
[9] M. M. Bull, W. J. Chung, S. R. Anderson, S.-j. Kim, I.-B. Shim,
H.-j. Paik, J. Pyun, J. Mater. Chem. 2010, 20, 6023.
[10] J. Park, K. An, Y. Hwang, J.-G. Park, H.-J. Noh, J.-Y. Kim, J.-H.
Park, N.-M. Hwang, T. Hyeon, Nature Mater. 2004, 3, 891.
[11] S. G. Kwon, Y. Piao, J. Park, S. Angappane, Y. Jo, N.-M. Hwang,
J.-G. Park, T. Hyeon, J. Am. Chem. Soc. 2007, 129, 12571.
[12] Y. C. Cao, J. Wang, J. Am. Chem. Soc. 2004, 126, 14336.
[13] Y. A. Yang, H. Wu, K. R. Williams, Y. C. Cao, Angew. Chem.
2005, 117, 6870; Angew. Chem. Int. Ed. 2005, 44, 6712.
[14] L. Li, P. Reiss, J. Am. Chem. Soc. 2008, 130, 11588.
[15] T.-Y. Liu, M. Li, J. Ouyang, M. B. Zaman, R. Wang, X. Wu, C.-S.
Yeh, Q. Lin, B. Yang, K. Yu, J. Phys. Chem. C 2009, 113, 2301.
[16] J. Ouyang, J. Kuijper, S. Brot, D. Kingston, X. Wu, D. M. Leek,
M. Z. Hu, J. A. Ripmeester, K. Yu, J. Phys. Chem. C 2009, 113,
7579.
[31] B. Abꢅcassis, F. Testard, O. Spalla, P. Barboux, Nano Lett. 2007,
7, 1723.
[32] M. F. Casula, Y.-w. Jun, D. J. Zaziski, E. M. Chan, A. Corrias,
A. P. Alivisatos, J. Am. Chem. Soc. 2006, 128, 1675.
[33] G. Cheng, D. Romero, G. T. Fraser, A. R. Hight Walker,
Langmuir 2005, 21, 12055.
[17] L. M. Bronstein, X. Huang, J. Retrum, A. Schmucker, M. Pink,
B. D. Stein, B. Dragnea, Chem. Mater. 2007, 19, 3624.
[18] A. Puglisi, S. Mondini, S. Cenedese, A. M. Ferretti, N. Santo, A.
Ponti, Chem. Mater. 2010, 22, 2804.
[34] Y. Kang, X. Ye, C. Murray, Angew. Chem. 2010, 122, 6292;
Angew. Chem. Int. Ed. 2010, 49, 6156.
2084 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 2080 –2084