Crystal structure control of CdSe nanocrystals in growth and nucleation: Dominating effects of surface versus interior structure

Article abstract of DOI:10.1021/ja5020025

For the most studied nanocrystal system in the literature, experimental results in this paper revealed that formation of either zinc blende or wurtzite CdSe nanocrystals was dominated by the ligand-surface interaction, instead of the interior structure difference. This conclusion was considered to be reasonable, given the very small energy difference between wurtzite and zinc blende CdSe (only 1.4 meV per CdSe unit and about 1000 times smaller than the energy of a single cadmium-ligand bond). Cadmium carboxylate ligands in the form of cadmium fatty acid salts promoted formation of the zinc blende structure. Conversely, cadmium phosphonate ligands with a long hydrocarbon chain favored the formation of the wurtzite structure. The effects of either cadmium carboxylate or cadmium phosphonate ligands were found to play a determining role during both nucleation and growth. Different from our expectation, fatty amine was found to be only a secondary factor for crystal structure determination. With an appropriate choice of capping ligands, it was possible to achieve precise control of the crystal structure of the CdSe nanocrystals in both nucleation and growth for either the zinc blende or wurtzite structure.

Full text of DOI:10.1021/ja5020025

Article
pubs.acs.org/JACS
Crystal Structure Control of CdSe Nanocrystals in Growth and
Nucleation: Dominating Effects of Surface versus Interior Structure
Yuan Gao and Xiaogang Peng*
Center for Chemistry of Novel & High-Performance Materials and Department of Chemistry, Zhejiang University, Hangzhou,
310027, People’s Republic of China
*
S Supporting Information
ABSTRACT: For the most studied nanocrystal system in the literature, experimental
results in this paper revealed that formation of either zinc blende or wurtzite CdSe
nanocrystals was dominated by the ligand−surface interaction, instead of the interior
structure difference. This conclusion was considered to be reasonable, given the very
small energy difference between wurtzite and zinc blende CdSe (only 1.4 meV per
CdSe unit and about 1000 times smaller than the energy of a single cadmium−ligand
bond). Cadmium carboxylate ligands in the form of cadmium fatty acid salts
promoted formation of the zinc blende structure. Conversely, cadmium phosphonate ligands with a long hydrocarbon chain
favored the formation of the wurtzite structure. The effects of either cadmium carboxylate or cadmium phosphonate ligands were
found to play a determining role during both nucleation and growth. Different from our expectation, fatty amine was found to be
only a secondary factor for crystal structure determination. With an appropriate choice of capping ligands, it was possible to
achieve precise control of the crystal structure of the CdSe nanocrystals in both nucleation and growth for either the zinc blende
or wurtzite structure.
INTRODUCTION
respect to each other. Although this structural difference seems
to be subtle, it breaks the inversion symmetry in the wurtzite
structure and thus makes the optical properties of the two types
of nanocrystals differ from each other, which was suggested by
■
The synthesis of size- and shape-controlled colloidal semi-
conductor nanocrystals in their quantum confinement size
regime has achieved great success in the past couple decades.
Extensive studies revealed that, to fully realize their size-
dependent optical properties, semiconductor nanocrystals
should be not only monodisperse in size but also single-
1
,2
3
−5,16
both theoretical and experimental studies.
Therefore, if
the nanocrystals were synthesized with an imperfect structure, a
new type of unnecessary inhomogeneity of their optical
properties would result.
3
−5
crystalline with identical crystal structures. This has inspired
great interest in identification of parameters that control the
crystal structure of semiconductor nanocrystals in both
The difficulty in controlling the structural perfection of either
zinc blende or wurtzite is their quite small structural difference
discussed in the above paragraph, which in turn results in a very
small energy difference. For CdSe, theoretical calculations
revealed that the energy difference between its zinc blende and
6
,7
8−13
nucleation
and growth.
Unfortunately, even for the
most studied CdSe nanocrystal system, perfection of the crystal
structure of the nanocrystals has not reached a satisfactory level
yet, especially in the wurtzite structure.
described below suggest that, instead of considering the interior
structure difference, it is necessary to take surface effects into
account. In the specific case of the CdSe nanocrystal system, in
this paper we will reveal that the surface effects dominated the
structural difference during both nucleation and growth of
CdSe nanocrystals for either the wurtzite (hexagonal) or zinc
blende (face-centered cubic) structure.
17
1
4,15
wurtzite structures is only 1.4 meV per CdSe unit. Because
the phase transition temperature from zinc blende (room
temperature stable) to wurtzite (high temperature stable) is
The results
18
about 100 °C for the bulk, the entropy difference between the
two phases should also be quite small. It is interesting to notice
that synthetic methods for good quality zinc blende as well as
wurtzite CdSe nanocrystals were all established at a temper-
1
,2
ature significantly above 100 °C, and the structural perfection
of zinc blende nanocrystals (typically formed between 200 and
II−VI and III−V binary systems, at present, are the main
focus of both fundamental study and technical applications of
semiconductor nanocrystals. These binary compounds can all
appear in either the zinc blende or wurtzite crystal structure. In
both types of crystal structures, each cation (anion) is
coordinated with four counterions in a tetrahedral config-
uration. While the local coordination and bond length in both
crystal structures are identical, the hexagonal planesalong the
3
00 °C) seemed to be at least as good as that of their wurtzite
counterparts.
There is a substantial difference between a bulk crystal and a
nanocrystal, which is the role of surface coordination. To
achieve satisfactory optical properties and colloidal stability,
good quality semiconductor nanocrystals available today were
(
111) direction for the zinc blende structure and the (001)
Received: February 26, 2014
Published: April 8, 2014
direction for the wurtzite structureare rotated 60° with
©
2014 American Chemical Society
6724
dx.doi.org/10.1021/ja5020025 | J. Am. Chem. Soc. 2014, 136, 6724−6732

Journal of the American Chemical Society
Article
all synthesized with a monolayer of organic ligands, such as
ligands varies from case to case. As a rough reference, a typical
metal−ligand coordination bond should be between ∼50 to
150 kJ/mol, which is extremely large in comparison with the
energy difference between the zinc blende and wurtzite crystal
structures.
For the case of nucleation, the size of the particles should be
2 nm (or smaller). Taking a 2 nm CdSe particle coated with
carboxylate ligandsusually from fatty acidsas an example,
the total energy difference between the zinc blende and
wurtzite interior structures for this particle is only ∼0.11 eV.
This value is even smaller than the energy of a single hydrogen
bond, which is roughly 20 kJ/mol (or 0.21 eV/bond). As the
total number of carboxylate−Cd bonds for a 2 nm CdSe
particle should be around 40−50, the energy difference
between two crystal structures would thus unlikely play a
significant role in the nucleation stage.
For the growth stage after nucleation, the energy difference
contributed by the interior structures was further reduced in
comparison to that of the surface ion−ligand coordination. Let
us consider a layer-by-layer growth model for conservative
reasons. The number of bonds formed in the middle of a layer
is proportional to its area, and the number of coordinations on
the surface should be proportional to the number of dangling
bonds at the edge of the layer. Therefore, with this conservative
model, one could safely conclude that the role of surface
coordination in the growth stage should be at least as important
as in the case of nucleation. For an aggressive model, one could
argue that every step of growth involves surface ion−ligand
bonds.
19
carboxylate, amine, phosphonate, and phosphine ligands. The
bonding energy between the surface cadmium ions and surface
ligands must be substantially larger than the energy difference
between zinc blende and wurtzite crystal structures (for details
see the Results and Discussion). This suggested to us that
control of the crystal structure of the CdSe nanocrystal system
should be dominated by the ligand bonds with the surface ions
of the nanocrystals.
A systematic study was carried out to test the hypothesis
outlined in the above paragraph, i.e., formation of a specific
crystal structure being determined by capping ligands. By
judiciously designing the reaction system, it was possible to
separate the initial formation (or nucleation) stage and the
following growth stage, which demonstrated somewhat differ-
ent trends of ligand effects. While the main hypothesis was
generally verified, this work also revealed some other
interesting results. Unlike the studies described in the
2
0,21
literature,
this study showed that synthesis of stable
wurtzite CdSe nanocrystals in the commonly known nucleation
size regime, i.e., ∼2 nm, was possible with proper choices of the
ligands. Using these newly discovered wurtzite CdSe small
particles as the seeds, it will be demonstrated below that high
crystal purity wurtzite CdSe nanocrystals could be synthesized
at relatively low temperature.
RESULTS AND DISCUSSION
■
Choice of Systems for Study. The CdSe nanocrystal
system, including both the wurtzite and zinc blende structures,
has been the most studied system up to now. In this study, as
the noncoordinating solvent approach in various forms gave the
The entropy difference of the interior structures would also
unlikely play a decisive role in control of the crystal structure
for the CdSe nanocrystal system. An early work reported that
the phase transition temperature from zinc blende (room
temperature stable) to wurtzite (high temperature stable) for
2
2,23
best performance,
it was applied. Octadecene (ODE) was
the noncoordinating solvent. Organic ligands studied by this
work mainly included the three most common ones, namely,
cadmium carboxylates, amines, and cadmium phosphonates
with a long hydrocarbon chain. The cadmium precursor was
either CdO reacting with fatty acids in situ or isolated cadmium
fatty acid (phosphonic acid) salts. The Se precursor was mostly
18
bulk CdSe is about 100 °C. From this and the enthalpy of the
phase transition, one could roughly estimate the entropy
difference between two phases should be ∼0.38 J/(K mol),
which is very small. A computational work supported this
estimation, which reported the entropy difference between zinc
2
4
a suspension of Se in ODE (Se-SUS), a Se−ODE
7
,23,25
27
solution,
or a Se−amine solution. The commonly used
blende and wurtzite bulk CdSe is negligible.
14
28
Se−phosphine precursors were also tested as well. The
reactions were typically carried out at temperatures between
The Rosenthal group reported that sub ∼2 nm CdSe
nanocrystals could be amorphous even at room temperature.
They suggested that, as the energy difference between the zinc
blende and wurtzite structures is so small, nanoparticles with
such small sizes have a somewhat difficult time choosing a
structure. Because the reaction temperature in a typical
synthesis for good quality CdSe nanocrystals is always
significantly higher than room temperatures, it is thus not
possible for the CdSe nanocrystals to make a choice based on
the energy difference between the two structures. As a result,
the final structure of the nuclei and the growth of additional
materials might be solely determined by the surface ligands.
Effects of Fatty Amine on the Growth of Wurtzite
CdSe Nanocrystals. Fatty amines have been widely used in
the synthesis of wurtzite CdSe and other types of II−VI (or
III−V) semiconductor nanocrystals. However, in terms of
control of the crystal structure of a given type of nanocrystal,
2
00 and 300 °C under an Ar flow. The details are provided in
the Experimental Section.
For simplicity, we employ an expanded abbreviation system
26
suggested previously for fatty amine, fatty acid, and fatty acid
salts. Fatty amines are abbreviated as NH Xy, where Xy is the
2
first two letters of the common name of the corresponding
amine. In this abbreviation system, the first letter of the
common name is capital and the second is lower case. For
example, oleylamine would be abbreviated as NH Ol. For fatty
2
acid and cadmium fatty acid salts, the abbreviations are quite
similar to those of fatty amines, but the “NH ” in the
2
abbreviation is replaced by either “H” or “Cd”, respectively. For
instance, stearic acid and cadmium stearate would be written as
HSt and Cd(St) , respectively.
2
Estimation of the Transition Enthalpy and Entropy
from Zinc Blende to Wurtzite CdSe Nanocrystals. For
CdSe, theoretical calculations revealed that the energy
difference between its zinc blende and wurtzite structures is
9
,29
different studies usually led to different conclusions.
To
clarify this issue, a systematic study was carried out with a fixed
solution composition other than the concentration of the
amine.
Figure 1 illustrates a series of powder X-ray diffraction
(XRD) patterns of CdSe nanocrystals (∼5 nm in size for all
17
only 1.4 meV per CdSe unit, which corresponds to about 0.14
kJ/mol of CdSe units. For the CdSe nanocrystal system, the
bonding energy between the surface cadmium ions and surface
6
725
dx.doi.org/10.1021/ja5020025 | J. Am. Chem. Soc. 2014, 136, 6724−6732

Journal of the American Chemical Society
Article
Figure 1. XRD patterns of CdSe nanocrystals (∼5 nm in size)
synthesized with different oleylamine (NH Ol) concentrations varied
2
Figure 2. Representative TEM images, high-resolution TEM images,
and SAED patterns of CdSe nanocrystals synthesized with either 0.60
from 0 to 0.6 mM. Three unique diffraction peaks of the wurtzite
structure are indicated by dashed arrows (WZ(102), WZ (103), and
WZ(203)) in the plot. The short bars on the top (bottom) of the plot
represent the diffraction pattern of the bulk wurtzite (zinc blende)
CdSe structure.
mM (left panel) or 0 mM (right panel) NH Ol.
2
to be wurtzite nanocrystals with a significant number of
stacking faults. Interestingly, the stacking faults were not
randomly distributed. The high-resolution TEM picture of a
cases) formed with different NH Ol concentrations. In all
2
reactions, the cadmium precursor (Cd(St) ) was formed by
reaction of 0.1 mmol of CdO and 0.4 mmol of HSt in 4 mL of
ODE, and the Se precursor was a Se solution made by
typical CdSe nanocrystal synthesized with a high NH Ol
2
2
concentration is shown in Figure 2 (left panel) and reveals that
the particle actually has a section of zinc blende at one end.
Also consistent with the XRD analysis, the zinc blende
nanocrystals were found to be of good quality, with one
stacking fault on average along the (111) direction per dot (see
Figure 2, right panel for one example with a single stacking
fault). With such atomic level information from the high-
resolution TEM measurements, one could simulate the XRD
pattern (simulation details to be provided below) with good
confidence (see Figure S1, Supporting Information).
dissolving Se in ODE and NH Ol. Introduction of amine from
2
the Se precursor was found to be beneficial in terms of
controlling the stability of the cadmium carboxylates in the
reaction system. To make all reactions comparable, the reaction
temperature was controlled between 280 and 300 °C until the
size of the nanocrystals reached ∼5 nm (see the transmission
electron microscopy (TEM) images in Figure 2 as examples).
The XRD patterns in Figure 1 reveal a gradual change of the
CdSe nanocrystals from zinc blende to wurtzite structure with
The above observation is consistent with the suggestion that
nucleation of zinc blende CdSe nanocrystals is relatively
an increase of the NH Ol concentration. All diffraction peaks in
2
2
0,21,30
Figure 1 support this trend. As the NH Ol concentration
easy.
Thus, a plausible interpretation of the results
2
gradually increased in Figure 1, the (111) peak of the zinc
blende structure at about 26° evolved into a broad envelop due
to the overlap of the (100), (002), and (101) Bragg peaks of
the wurtzite structure. Unlike the completely indistinguishable
diffraction peaks between the zinc blende and wurtzite
structures, the unique Bragg peaks of the wurtzite structure,
namely, the (102), (103), and (203) peaks, are all identifiable
by the patterns associated with reactions with a reasonably high
amine concentration.
illustrated above is that the initial seedsor nucleiof CdSe
nanocrystals were zinc blende seeds. When no amine was
present in the reaction solution, the nanocrystals were grown in
zinc blende structure with only fatty acid salts (fatty acids
excluded by late experiments; see the next subsection) as the
ligands. When amine was added as an additional ligand, the
system was promoted to the growth of the wurtzite structure.
Because the reaction temperature was significantly below the
phase transition temperature needed for the nanocrystals in this
11
When no NH Ol was present in the reaction solution, the
size range, the zinc blende structure remained as a section
along with the wurtzite section in the final nanocrystals.
Growth of CdSe Nanocrystals on Existing Zinc Blende
Seeds. Though the addition of amine leads to the growth of
wurtzite CdSe, one could argue that amine may activate the Cd
precursor, leading to a faster growth. According to the
2
XRD pattern (the lowest pattern in Figure 1) matched quite
well with the standard zinc blende diffraction pattern. It should
be noticed that even the XRD pattern with the highest NH Ol
2
concentration in Figure 1 reveals significant imperfection in
comparison to the standard wurtzite diffraction pattern. A
11
further increase of the NH Ol concentration apparently did not
literature, the growth rate might dictate the crystal structure
of the final nanocrystals. Thus, the results in Figures 1 and 2
cannot distinguish whether the structure is the effect of either
the ligands or the reaction rate.
2
improve the quality of the samples as judged by XRD
diffraction. This imperfection is evidently similar to that of
the common XRD patterns of wurtzite CdSe nanocrystals
1
4,15
reported in the literature.
Furthermore, selective area
A set of comparative experiments were designed to clarify the
issue discussed in the above paragraph. In this set of reactions,
zinc blende CdSe nanocrystals with only 2.5 nm size were
synthesized without addition of amine. The as-synthesized
electron diffraction (SAED) patterns shown in Figure 2 also
confirm the perfection of the zinc blende nanocrystals and
imperfection of the wurtzite nanocrystals.
High-resolution TEM was applied to further examine the
structure of the nanocrystals. Consistent with the hypothesis
illustrated in the above paragraph, the nanocrystals were found
5
seeds were purified according to the literature to remove
unreacted precursors. As shown in Figure S2 (Supporting
Information), no shift of the first peak of absorption was
6
726
dx.doi.org/10.1021/ja5020025 | J. Am. Chem. Soc. 2014, 136, 6724−6732

Journal of the American Chemical Society
Article
observed after the purification. The purified seeds were mixed
with 2 mL of NH Ol and 4 mL of ODE and heated to 220 °C
2
under an Ar flow. A slight blue shift (2 nm) of the first
absorption peak was observed, indicating the partial displace-
31
ment of metal carboxylate complexes by amine. Other than
that, the system was found to be stable upon prolonged heating.
At this temperature, 0.6 mL of 0.1 M Se-SUS was injected into
24
the reaction flask. According to a recent study, Se powder in
the suspension would immediately yield very reactive H Se, and
2
the reactive H Se would attack the Cd−carboxylate bonds on
2
the surface of the nanocrystals. In the current system, this
rapidly initiated the growth of the nanocrystals through the
Ostwald ripening process. In this process, CdSe monomers
were dissolved from small dots and grown on the bigger ones,
which resulted in the red shift of the first peak of UV−vis
absorption (Figure S2).
The comparative experiment was performed by simply
replacing NH Ol with oleic acid (HOl). Again, the nanocrystals
2
remained stable by heating the mixture of the 2.5 nm zinc
blende seeds, HOl, and ODE to 220 °C. However, the amount
of Se-SUS with the same volume fraction of Se was reduced to
0.2 mL. Injection of the same amount of Se-SUS used in the
first reaction described in the above paragraph caused rapid
aggregation of CdSe nanocrystals. Infrared (IR) measurements
suggested that the nanocrystal precipitate was nearly naked
(
Figure S3, Supporting Information). Because there was a large
amount of HOl in the solution, the precipitation of the
nanocrystals implied that HOl could not be a good ligand for
the nanocrystals. This seems to be reasonable if one notices
that free fatty acids cannot coordinate well with either the
surface cations or surface anions.
The XRD patterns of the final products (∼5 nm in size) of
two comparative reactions are plotted in Figure 3a. The pattern
for the reaction with NH Ol resembled the pattern with a high
2
concentration of amine in Figure 1, namely, a pattern with all
wurtzite features and a certain degree of structural imperfection.
The corresponding high-resolution TEM picture (see Figure 3b
as one example) also showed a combination of a zinc blende
section and a wurtzite section. In comparison, the XRD pattern
and high-resolution TEM picture for the reaction with HOl
Figure 3. (a) XRD patterns of the 2.5 nm zinc blende (ZB) core and 6
nm final nanocrystals formed with different ligands. (b, c) HR-TEM
images of the final samples. (d, e) Time evolution curves of the
absorption peak of the nanocrystals for the growth reactions with
different ligands.
(
instead of NH Ol) revealed the final CdSe nanocrystals had
acids did not. This fact inspired an extensive amount of work to
achieve structure-controlled nucleation by fatty amine ligands.
Several types of cadmium precursors were tested, including
2
the zinc blende structure.
Moreover, parts d and e of Figure 3 suggest that two
comparative reactions, though they resulted in growth of either
the wurtzite structure (with NH Ol) or zinc blende structure
(
means that the growth rate was not the determining parameter
for growth in either the wurtzite or zinc blende structure.
Instead, it would be safe to conclude that amine ligands
induced the growth of the wurtzite structure on these zinc
blende seeds through ligand−surface interaction.
It is worth noting that the results above contradict a recent
proposal that Ostwald’s rule of stages dictates the isolated phase
of the final nanocrystals without a structure-directing ligand.
There could be two reasons for this inconsistency. First, their
system possessed a relatively complex ligand composition, i.e., a
mixture of amine, cadmium carboxylate, and tributylphosphine.
Second, the variation of the Cd:Se ratio and Cd(St)2
concentrations in their system could also lead to an
uncontrolled surface.
Attempts To Control Nucleation by Amine Ligands.
The experimental results shown in Figures 1−3 reveal that fatty
amine promoted growth of the wurtzite structure while fatty
cadmium nitrate, cadmium acetylacetonate, Cd(Ac) , and
2
2
Cd(Fa) . Unfortunately, none of the reactions with these
2
with HOl), occurred at about the same growth rate. This
nonligand Cd precursors gave us positive results on nucleation
control. Usually, the reaction was often too fast to be
controlled, which reveals that fatty amines are much weaker
ligands in comparison to cadmium fatty acid salts.
Reactions attempted to consume the carboxylate groups
from long-chain fatty acids during formation of CdSe
nanocrystals were also unsuccessful. Although formation of
amides by the reaction of cadmium fatty acid salts with amine
was observed, the XRD patterns of the resulting CdSe
nanocrystals were still not much better than those with a
high amine concentration in Figure 1. Furthermore, employ-
ment of other noncharge ligands, such as trioctylphosphine
oxide (TOPO) and tributylphosphine (TBP), did not improve
the situation, too.
11
Nucleation in Wurtzite Structure by Cadmium
Phosphonate Ligands. The results discussed above suggest
that, with cadmium carboxylates and/or fatty amines, it would
not be possible to obtain wurtzite CdSe nanocrystals with a
6
727
dx.doi.org/10.1021/ja5020025 | J. Am. Chem. Soc. 2014, 136, 6724−6732

Journal of the American Chemical Society
Article
high structural purity. Likely, the problem is a lack of ligand-
dominated structure control for wurtzite CdSe during the
nucleation stage. By going through the literature, it is found
that cadmium phosphonate often fostered growth of wurtzite
rods and tetrapods out of a zinc blende core with good crystal
was typical, i.e., 250−280 °C, about 100 °C lower than the
temperature (∼370 °C) used for formation of wurtzite CdSe
36
nanocrystals reported in literature.
Interestingly, when fatty amine was introduced into this new
system, zinc blende CdSe nanocrystals with at least two
stacking faults were obtained (Figure 4, bottom). Considering
there were only seven atom layers in such a small nanocrystal,
“polycrystalline” may be a more appropriate description of the
dots synthesized with amine added to this new system.
3
2,33
quality.
More importantly, the side facets of the wurtzite
32,33
rods and tetrapods were typically nonpolar ones.
This
implies that, though cadmium phosphonate ligands are also
negatively charged, they might be able to stabilize all facets of
the wurtzite structure of CdSe nanocrystals. The difference
between carboxylate and phosphonate ligands might be a result
of their distinguishable coordinating habits. Usually, a
phosphonate group can coordinate with three or more cation
High-resolution TEM could directly offer structural
information on nanocrystals at the atomic level, especially
37
with the help of exit wave reconstruction. However, a particle
with a specific orientation on the substrate would be required
to give a definite answer. For CdSe nanocrystals with wurtzite
and zinc blende structures, such a specific orientation does not
occur frequently on TEM substrates. Thus, a more conclusive
method is to combine high-resolution TEM experiments with
XRD simulation, which could give a global picture of the entire
sample. The XRD patterns in this work were all simulated
3
4,35
centers, instead of one (or two) for the carboxylate group.
Fortunately, employment of cadmium phosphonate as the
Cd precursor (Se-SUS as the Se precursor) indeed yielded
small-sized wurtzite nanocrystals with very high structure
purity. The XRD pattern matched well with a pure wurtzite
stacking sequence (Figure 4, top). The reaction temperature
14
following a literature method. The nanocrystals were built by
layering either the (111) plane of zinc blende or the (001)
plane of wurtzite one by one. The bulk bond length (2.63 Å for
CdSe) was used. The sequence used for zinc blende was
ABCABC..., while ABABAB... was used for the wurtzite
structure. Atoms falling within the defined sphere of a given
size were selected and calculated using a discrete form of the
38
Debye equation:
f2 (S)
πS
p(rk)
I(S) = I0 2
∑
sin(2πr S)
k
rk
(1)
k
where I(S) is the diffracted intensity, I is the incident intensity,
0
f(S) is the scattering factor, S is the scattering parameter (S = 2
sin(q)/λ for X-rays of wavelength λ diffracted through angle q),
r is the interatomic distance, and p(r ) is the number of times a
k
k
Figure 4. XRD patterns of 2.5 nm CdSe nanocrystals synthesized with
given interatomic distance r occurred.
k
(
bottom) and without (top) NH Ol using cadmium dodecylphosph-
2
Simulation of the XRD patterns (Figure 4) confirmed that
cadmium phosphonate ligands indeed behaved differently from
cadmium carboxylate ligands. Surprisingly, the effect of amine
revealed in Figure 4 was actually to promote formation of the
zinc blende structure, instead of the wurtzite structure
onate (Cd(DDPA) ) as the Cd precursor. Red lines are the computer
2
simulation results. Stacking sequences (shown to the right of the plot)
applied to the simulations are ABABABA (ABC*BAC*A) for particles
synthesized without (with) NH Ol.
2
9
suggested in the literature. Moreover, the results in Figures
1
−3 definitely reveal amine-promoted growth of the wurtzite
Figure 5. (a) XRD patterns of CdSe nanocrystals (5 nm in size) grown with wurtzite seeds using a cadmium carboxylate precursor with (or without)
amine. Red lines are the computer simulation results. (b) Schematic diagram of the CdSe nanocrystals used for XRD simulations and the
corresponding typical HR-TEM images.
6
728
dx.doi.org/10.1021/ja5020025 | J. Am. Chem. Soc. 2014, 136, 6724−6732

Journal of the American Chemical Society
Article
Figure 6. XRD pattern and a representative high-resolution TEM image of 5 nm CdSe nanocrystals formed by seeded growth with cadmium
dodecylphosphonate (Cd(DDPA) ) as the Cd precursor and NH Ol added. Red lines are the computer XRD simulation results using the sequence
2
2
and shape in the corresponding schematic diagram. Left: 2.5 nm wurtzite seeds. Right: 2.5 nm zinc blende seeds.
structure in the growth stage. Evidently, the main difference
between the reactions associated with Figure 4 and those
shown in Figures 1−3 was the cadmium precursor usedalso
potentially the major ligands. Presumably, the existence of
amine in either the cadmium phosphonate or cadmium
carboxylate system would change the composition of the
ligands, which in turn would weaken the structure choice of the
nuclei for both cases. The experiments below further confirm
this suggestion.
had a zinc blende section on one edge of the particle. For the
nanocrystals grown without amine, high-resolution TEM
showed CdSe nanocrystals with a small section of wurtzite
structure sandwiched in the middle of two zinc blende regions
(Figure 5b). More quantitatively, the maximum size of the
wurtzite section in the sandwich structure was found to be ∼7
atomic layers according to the simulation results. This was the
size of the seeds used in the growth (see Figure 4). It should be
mentioned that, as amine could also be a ligand in the system,
the final Cd:Se ratio in such reactions should not exceed 1:1.
Otherwise, the cadmium carboxylate ligands would be
converted into fatty acid. Similar experimental results were
also obtained if the Se precursor was the traditional selenium
tributylphosphine (Figure S4, Supporting Information) instead
of Se-SUS.
Growth onto Wurtzite Seeds with Cadmium Carbox-
ylate. After the wurtzite CdSe nanocrystals were obtained in
the nucleation stage, it became possible to complete the study
on ligand-induced structure control in the growth stage with
cadmium carboxylate. Such studies using zinc blende seeds in
2
9
the literature as well as this study (see Figures 1−3 as
examples) are already in place.
The results in this subsection combined with the results
about growth on zinc blende seeds (see the text related to
Figures 1−3) reveal that the ligand effects at the growth stage
could be separated from that in the nucleation stage.
Independent of the crystal structure of the seeds, cadmium
carboxylate ligands promoted the growth of zinc blende CdSe
nanocrystals. The existence of amine along with cadmium
carboxylate ligands would alter the growth habit from zinc
blende to wurtzite nanocrystals. Again, similar to what was
proposed for amine in altering the nucleation habit of cadmium
phosphonates, amine in the current case could also affect the
ligand environment dictated by the cadmium carboxylate
ligands in the solution. This in turn would change the growth
habit of the nanocrystals. The results above further imply that,
if a high concentration of amine and an excess of Se precursor
are in place, the CdSe nanocrystals obtained will be in a
structure close to wurtzite with zinc blende stacking faults. This
was actually the condition mostly used for synthesis of such a
The purified wurtzite CdSe nanocrystal seeds (∼2.5 nm in
size) synthesized using cadmium phosphonate (Figure 4, no
39
amine) with known concentration were mixed with Cd(La)2
and NH Ol in ODE. The mixture was heated to 220 °C under
2
Ar and was found to be stable. The growth was initiated by
injection of Se-SUS into the flask (see the Experimental Section
for details). For comparison, a reaction without any amine was
performed in parallel with all other conditions the same. The
XRD patterns show a dramatic difference between the
nanocrystals obtained in this set of comparative reactions
(
Figure 5a). The XRD pattern of the final CdSe nanocrystals
grown with the existence of NH Ol was likely among the best
2
quality diffraction patterns reported up to now for wurtzite
10,15
CdSe nanocrystals.
Conversely, the XRD pattern for the
reaction without amine was quite similar to that of zinc blende
with some stacking faults. The analytical technique with XRD
simulation (Figure 5a) coupled with high-resolution TEM
15
(
Figure 5b) was applied to yield more quantitative structural
type of CdSe nanocrystals in the literature.
information for both cases. The simulation results indicated
that, though the quality of the XRD pattern of the nanocrystals
grown with amine in place (top curve in Figure 5a) was among
the best in the literature for wurtzite CdSe nanocrystals, it still
Synthesis of Wurtzite CdSe Nanocrystals with Differ-
ent Types of Seeds. Initially, we suspected that cadmium
phosphonate (or cadmium phosphonate with some free
phosphonic acid) would yield high-purity wurtzite CdSe
6
729
dx.doi.org/10.1021/ja5020025 | J. Am. Chem. Soc. 2014, 136, 6724−6732

Journal of the American Chemical Society
Article
nanocrystals with different sizes under certain conditions.
However, this was not successful. The system was often found
to be too reactive when excess phosphonic acid was added (or
produced by the reaction of the Se precursor with excess
cadmium phosphonate), which would thus induce Ostwald
ripening rapidly. Because phosphonic acid is a pretty strong
acid, we thought it should be possible to neutralize it by fatty
amine. The results in Figure 6 (right) reveal that, by addition of
conditions at least when the size of the nanocrystals was 2.5 nm
or larger. By controlling both nucleation and growth ligand
systems, it was possible to obtain high-purity CdSe nanocrystals
for both the wurtzite and zinc blende structures. Selection of
one crystal structure over the other was found to be
independent of the reaction rates, and in fact, formation of
high-purity wurtzite CdSe nanocrystals with a decent size (∼5
nm) could only be achieved with a very low growth rate in this
work. Although the exact ligand−surface coordination is
unknown due to limited access to the surface structure of
nanocrystals at present, the results shown in this work shed new
light on the precise control of the crystal structure of the
nanocrystals. This is not only important for synthesis of
nanocrystals with desired properties, but also of interest for
understanding crystallization in general. In principle, any
crystallization event starts from crystals in the nanometer size
regime.
NH Ol to the reaction system with wurtzite seeds formed by
2
the approach shown in Figure 4, large-sized CdSe nanocrystals
with a high-purity wurtzite crystal structure were indeed
obtained. The XRD pattern with very sharp (102) and (103)
peaks in Figure 6 (left) was substantially better in comparison
to that shown in Figure 5 (top curve). High-resolution TEM
and computer simulation of the XRD pattern gave us wurtzite
nanocrystals with at most one stacking fault per dot (the high-
resolution TEM image in Figure 6 (left) shows a wurtzite
nanocrystal without any stacking fault).
It should be noted that this approach is not optimized yet for
production (see the Experimental Section for details), though it
does yield high-purity wurtzite CdSe nanocrystals. At present,
the Se precursor is selenium tributylphosphine as the greener
Se-SUS would result in rapid nucleation of small-sized CdSe
nanocrystals. This is very likely because cadmium phosphonates
are bound onto the surface of nanocrystals quite tightly. As a
result, when active Se precursors are added to the system, they
react with cadmium phosphonates in solution.
Experimental results further revealed that the high-purity
wurtzite CdSe nanocrystals in large sizes also depended on the
structure of the seeds. As shown in Figure 6 (right), the
structural purity of the final nanocrystals grown on the zinc
blende seeds using cadmium phosphonate and amine was quite
poor. Quantitatively, the final nanocrystals grown with zinc
blende seeds were identified as having one section of zinc
blende and one section of wurtzite. More importantly, the size
of the zinc blende region agreed with the size of the seeds used
EXPERIMENTAL SECTION
■
Chemicals. Stearic acid (HSt; 90+%), lauric acid (HLa; 99.5%),
tetramethylammonium hydroxide (98%), octylamine (98%), cadmium
oxide (CdO; 99.998%), selenium powder (200 mesh, 99.999%), 1-
dodecylphosphonic acid (DDPA; 95%), 1-octadecene (ODE; 90%),
and oleic acid (HOl; 90%) were purchased from Alfa-Aesar.
Oleylamine (NH Ol; C content 70%) was purchased from Aldrich.
2
18
Tributylphosphine (TBP) was purchased from Shanghai Titanchem.
Cadmium acetate dihydrate (Cd(Ac) ·2H O; 98.5%) was purchased
from Shanghai Tingxin Reagents. All organic solvents were purchased
from Sinopharm Reagents. All chemicals were used directly without
any further purification.
2
2
Synthesis of Cadmium Stearate (Cd(St) ). The synthetic
2
procedure was similar to the room temperature synthesis of Mn(St)
2
4
3
reported in the literature. Typically, HSt (20 mmol) and 20 mmol of
tetramethylammonium hydroxide were dissolved in 200 mL of
methanol. To this solution was added dropwise under vigorous
stirring Cd(Ac) ·2H O (10 mmol) dissolved in 50 mL of methanol.
2
2
White Cd(St) immediately precipitated, and the mixture was stirred
2
for another 20 min to ensure the reaction was completed. The
precipitate was washed three times with methanol and dried under
vacuum overnight.
(
Figure 6, right).
The combined results in Figure 6 (left and right) reveal that
cadmium phosphonate promotes growth of wurtzite CdSe no
matter what the crystal structure of the seeds is. This is
consistent with our hypothesis that the crystal structures of
CdSe nanocrystals in both nucleation and growth should be
mostly determined by the ligands. However, this result
contradicts the suggestions in the literature. Cadmium
phosphonate was usually thought to “arrest” the structure of
the nuclei due to the strong bonding feature and thus should
Synthesis of Cd(La) . The procedure was similar to the synthesis
2
of Cd(St) described above, except the substitution of HSt by HLa.
2
Preparation of the Se Precursor. Se solutions in ODE and in
NH Ol (0.1 mM for both cases) were prepared by mixing 1 mmol of
2
Se powder with 10 mL of ODE and NH Ol, respectively. The mixtures
2
were degassed for 10 min and heated to 180 °C, and the reactions
were stopped after the full dissolution of Se. A 0.1 mM Se suspension
was prepared by dispersing 0.5 mmol of Se powder in 5 mL of ODE
by sonication for 5 min. A 0.5 mM TBP−Se solution was made by
dissolving 5 mmol of Se powder in 1.18 g of TBP in a glovebox and
further dilution with 7.5 g of ODE.
1
2,40
yield final nanocrystals with the same structure as the seeds.
CONCLUSION
Synthesis of 5 nm CdSe Nanocrystals under Different NH Ol
2
■
Concentrations. CdO (0.1 mmol), 0.4 mmol of HSt, and 4 mL of
ODE were mixed in a 25 mL flask. After the mixture was stirred and
argon was bubbled for 10 min, the contents of the flask were heated to
In summary, instead of the interior crystal structure, surface
ion−ligand coordination was found to be the key factor in
determining the crystal structure of CdSe nanocrystals, i.e.,
wurtzite versus zinc blende. Overall, the results above showed
that cadmium phosphonate ligands (cadmium carboxylate
ligands) promoted formation of the wurtzite (zinc blende)
CdSe structure in both nucleation and growth. Conversely,
fatty amines, fatty acids, phosphonic acids, and phosphines did
not show decisive effects. After extensive and systematic studies,
it was found that fatty amines usually act as a secondary factor
3
00 °C to get a clear solution. A Se solution was made by mixing the
desired volumes of 0.1 M Se−ODE and 0.1 M Se−NH Ol. A 1 mL
2
sample of the Se solution was injected into the reaction flask at 300
°C, and the temperature was dropped from 300 to 280 °C. The
temperature remained the same until the nanocrystals grew to 5 nm.
Synthesis of 2.5 nm Zinc Blende CdSe Seeds. Zinc blende
CdSe nanocrystals were synthesized using a modified procedure
24
reported in the literature. Cd(St) (0.1 mmol) was loaded into a 25
2
mL flask with 3 mL of ODE. After the mixture was stirred and argon
was bubbled for 10 min, the contents of the flask were heated to 260
C. A 1 mL sample of 0.1 M Se-SUS was swiftly injected into the flask
at 260 °C, and the reaction temperature was set to 240 °C for further
growth of the nanocrystals. After 5 min of growth, the particle reached
41
in structure determination likely due to their weak bonding
42
and highly dynamic nature. Experimental results further
revealed that structure conversion from wurtzite (zinc blende)
to zinc blende (wurtzite) would not occur in typical reaction
°
6
730
dx.doi.org/10.1021/ja5020025 | J. Am. Chem. Soc. 2014, 136, 6724−6732

Journal of the American Chemical Society
Article
the desired size (with the first excitonic absorption peak at 500 ± 5
nm), and further annealing at 240 °C for 5 min was carried out to
obtain a stable surface. Needle-tip aliquots were taken and dissolved in
toluene for monitoring the reaction using UV−vis.
transmission electron microscope at 100 kV using copper grids coated
with a pure carbon support film. HR-TEM images were taken on a
JEM 2100F transmission electron microscope at 300 kV using a
copper grid coated with an ultrathin carbon film as the substrate.
Synthesis of 2.5 nm Wurtzite CdSe Seeds. In a typical
synthesis, CdO (0.1 mmol) and 0.2 mmol of DDPA were loaded into
a 25 mL flask with 2 mL of ODE. After the mixture was stirred and
argon was bubbled for 10 min, the contents of the flask were heated to
ASSOCIATED CONTENT
■
*
S
Supporting Information
280 °C to get a clear solution. A 1 mL sample of 0.1 M Se-SUS was
Figures showing the XRD patterns of 5 nm CdSe nanocrystals
synthesized with or without amine, evolution of UV−vis spectra
of 2.5 nm zinc blende CdSe nanocrystals ripened in oleylamine
and oleic acid, FT-IR spectra of the precipitate yield with excess
Se suspension with oleic acid and the corresponding washed
precipitate, and isomaterial seeded growth on wurtzite seeds
swiftly injected into the flask at 280 °C, and the reaction temperature
was set to 250 °C for further growth of the nanocrystals. The reaction
was stopped at 30 s to get 2.5 nm nanocrystals.
In Situ Purification of Seed Nanocrystals. For zinc blende
CdSe seeds, octylamine (0.2 mL), hexane (4 mL), and methanol (8
mL) were added to the reaction flask at 50 °C and stirred for 1 min.
After stirring, the colorless methanol layer was separated from the top
ODE/hexane layer by a syringe. This extraction procedure was
repeated three times. The hexane left in the ODE layer was removed
by argon bubbling at about 70 °C. For wurtzite CdSe seeds, TBP (0.2
mL), octylamine (0.2 mL), hexane (3 mL), and methanol (6 mL) were
added to the reaction flask at 50 °C and stirred for 5 min. TBP was
used to get rid of the unreacted elemental Se, considering the very
short reaction time. After stirring, the colorless methanol layer was
separated from the top ODE/hexane layer by syringe. This extraction
procedure was repeated three times, but neither TBP nor octylamine
wes added for the last purification. The hexane left in the ODE layer
was removed by argon bubbling at about 70 °C.
with Cd(La) and Se-TBP. This material is available free of
2
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
xpeng@zju.edu.cn
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the National Natural Science
Foundation of China (Grants 21233005 and J1210042) and
Fundamental Research Fund for the Central Universities
(Grant 2014FZA3006).
Synthesis of CdSe Nanocrystals on Wurtzite Seeds with
Carboxylate. Purified 2.5 nm wurtzite seeds (3.0 × 10− mol) were
mixed with 0.2 mmol of Cd(La) and 1 mL of NH Ol, and the mixture
7
2
2
was further diluted by ODE to 4 mL of total volume. The mixture was
degassed for 10 min and heated to 220 °C, and 0.3 mL of 0.1 M Se-
SUS was swiftly injected into the flask. More Se-SUS was added
dropwise to get 5 nm quantum dots at 240 °C. Control experiments
REFERENCES
■
(
1) Murray, C. B.; Kagan, C. R.; Bawendi, M. G. Annu. Rev. Mater.
Sci. 2000, 30, 545.
2) Peng, X. G. Nano Res. 2009, 2, 425.
(3) Ninomiya, S.; Adachi, S. J. Appl. Phys. 1995, 78, 4681.
(4) Efros, A. L.; Rosen, M.; Kuno, M.; Nirmal, M.; Norris, D. J.;
Bawendi, M. Phys. Rev. B 1996, 54, 4843.
were carried out by simply exchanging NH Ol with the same volume
2
of ODE.
(
Growth of CdSe Nanocrystals on Different Seeds with
Cadmium Phosphonate. CdO (0.2 mmol) and 0.4 mmol of DDPA
(
stoichiometric ratio) were loaded into a 25 mL flask with 2 mL of
ODE. After 10 min of degassing, the mixture was heated to 280 °C to
get a clear solution. The Cd(DDPA) precursor was cooled to room
(5) Nan, W. N.; Niu, Y. A.; Qin, H. Y.; Cui, F.; Yang, Y.; Lai, R. C.;
Lin, W. Z.; Peng, X. G. J. Am. Chem. Soc. 2012, 134, 19685.
(6) Alivisatos, A. P. Ber. Bunsen-Ges. Phys. Chem. 1997, 101, 1573.
(7) Jasieniak, J.; Bullen, C.; van Embden, J.; Mulvaney, P. J. Phys.
Chem. B 2005, 109, 20665.
(8) Yang, Y. A.; Wu, H. M.; Williams, K. R.; Cao, Y. C. Angew. Chem.,
Int. Ed. 2005, 44, 6712.
(9) Mahler, B.; Lequeux, N.; Dubertret, B. J. Am. Chem. Soc. 2010,
132, 953.
(10) Huang, J.; Kovalenko, M. V.; Talapin, D. V. J. Am. Chem. Soc.
2010, 132, 15866.
(11) Washington, A. L.; Foley, M. E.; Cheong, S.; Quffa, L.; Breshike,
C. J.; Watt, J.; Tilley, R. D.; Strouse, G. F. J. Am. Chem. Soc. 2012, 134,
17046.
(12) Yu, W. W.; Wang, Y. A.; Peng, X. G. Chem. Mater. 2003, 15,
4300.
(13) Manna, L.; Milliron, D. J.; Meisel, A.; Scher, E. C.; Alivisatos, A.
P. Nat. Mater. 2003, 2, 382.
(14) Murray, C. B.; Norris, D. J.; Bawendi, M. G. J. Am. Chem. Soc.
2
−7
temperature and mixed with 1 mL of NH Ol and 3.0 × 10 mol of 2.5
2
nm CdSe seeds with either the zinc blende or wurtzite crystal
structure. The mixture was degassed for 10 min and heated to 220 °C,
and then 0.4 mL of 0.5 mM Se−TBP was injected into the flask. After
injection, the temperature was reset to 240 °C for the growth of CdSe
nanocrystals. When the first peak of absorption did not shift any
further, 0.2 mL of 0.5 mM Se−TBP was added dropwise to the
solution to reach the desired size.
X-ray Powder Diffraction (XRD). XRD patterns were acquired
using a Rigaku Ultimate-IV X-ray diffractometer operating at 40 kV/30
mA using the Cu Kα line (λ = 1.5418 Å). To get high-quality XRD
patterns, purification of the nanocrystal products was carried out
before the measurements. Nanocrystals synthesized with cadmium
carboxylate were purified using the same procedure for zinc blende
seeds described above. Nanocrystals synthesized with cadmium
phosphonate were precipitated by ethanol. Precipitates collected by
centrifugation and decantation were redissolved in toluene, and
unreacted phosphonate was separated by centrifugation at 8000 rpm
for 5 min. Powder samples for XRD measurements were prepared by
precipitation of the nanocrystals from the solution using ethanol as the
precipitant, and the precipitate was separated by decantation of the
solution waste. Collected precipitates were redissolved in hexane and
precipitated by ethanol again. The final precipitates were transferred
onto a glass slide for XRD measurements.
1993, 115, 8706.
(15) Qu, L. H.; Peng, X. G. J. Am. Chem. Soc. 2002, 124, 2049.
(16) Qin, H. Y.; Niu, Y.; Meng, R. Y.; Lin, X.; Lai, R. C.; Fang, W.;
Peng, X. G. J. Am. Chem. Soc. 2014, 136, 179.
(17) Yeh, C.-Y.; Lu, Z.; Froyen, S.; Zunger, A. Phys. Rev. B 1992, 46,
10086.
(18) Fedorov, V. A.; Ganshin, V. A.; Korkishko, Y. N. Phys. Status
Solidi A 1991, 126, K5.
(19) Qu, L. H.; Peng, Z. A.; Peng, X. G. Nano Lett. 2001, 1, 333.
(20) Soloviev, V. N.; Eichhofer, A.; Fenske, D.; Banin, U. J. Am.
Chem. Soc. 2000, 122, 2673.
Optical Measurements. UV−vis spectra were recorded on an
Analytik Jena S600 UV−vis spectrophotometer. Before measurement,
the nanocrystal sample was diluted to an optical density below 0.2.
Transmission Electron Microscopy (TEM) and High-Reso-
lution TEM (HR-TEM). TEM images were taken on a Hitachi 7700
6
731
dx.doi.org/10.1021/ja5020025 | J. Am. Chem. Soc. 2014, 136, 6724−6732

Journal of the American Chemical Society
Article
(
21) Dance, I. G.; Choy, A.; Scudder, M. L. J. Am. Chem. Soc. 1984,
106, 6285.
(
(
22) Yu, W. W.; Peng, X. G. Angew. Chem., Int. Ed. 2002, 41, 2368.
23) Deng, Z. T.; Cao, L.; Tang, F. Q.; Zou, B. S. J. Phys. Chem. B
2
(
005, 109, 16671.
24) Pu, C. D.; Zhou, J. H.; Lai, R. C.; Niu, Y.; Nan, W. N.; Peng, X.
G. Nano Res. 2013, 6, 652.
25) Bullen, C.; van Embden, J.; Jasieniak, J.; Cosgriff, J. E.; Mulder,
R. J.; Rizzardo, E.; Gu, M.; Raston, C. L. Chem. Mater. 2010, 22, 4135.
26) Xie, R. G.; Rutherford, M.; Peng, X. G. J. Am. Chem. Soc. 2009,
31, 5691.
(
(
1
(
27) Li, S.; Yang, G. W. J. Phys. Chem. C 2010, 114, 15054.
(
28) McBride, J. R.; Pennycook, T. J.; Pennycook, S. J.; Rosenthal, S.
J. ACS Nano 2013, 7, 8358.
29) Chen, O.; Yang, Y. A.; Wang, T.; Wu, H. M.; Niu, C. G.; Yang, J.
H.; Cao, Y. C. J. Am. Chem. Soc. 2011, 133, 17504.
30) Erk, C.; Rohner, C.; Schlecht, S. Z. Anorg. Allg. Chem. 2008, 634,
001.
31) Anderson, N. C.; Hendricks, M. P.; Choi, J. J.; Owen, J. S. J. Am.
Chem. Soc. 2013, 135, 18536.
32) Peng, X. G.; Manna, L.; Yang, W. D.; Wickham, J.; Scher, E.;
Kadavanich, A.; Alivisatos, A. P. Nature 2000, 404, 59.
(
(
3
(
(
(
(
33) Peng, Z. A.; Peng, X. G. J. Am. Chem. Soc. 2002, 124, 3343.
34) Fredoueil, F.; Evain, M.; Massiot, D.; Bujoli-Doeuff, M.; Janvier,
P.; Clearfield, A.; Bujoli, B. J. Chem. Soc., Dalton Trans. 2002, 1508.
35) Cao, G.; Lynch, V. M.; Yacullo, L. N. Chem. Mater. 1993, 5,
000.
36) Carbone, L.; Nobile, C.; De Giorgi, M.; Sala, F. D.; Morello, G.;
(
1
(
Pompa, P.; Hytch, M.; Snoeck, E.; Fiore, A.; Franchini, I. R.; Nadasan,
M.; Silvestre, A. F.; Chiodo, L.; Kudera, S.; Cingolani, R.; Krahne, R.;
Manna, L. Nano Lett. 2007, 7, 2942.
(
37) Bertoni, G.; Grillo, V.; Brescia, R.; Ke, X. X.; Bals, S.; Catellani,
A.; Li, H. B.; Manna, L. ACS Nano 2012, 6, 6453.
(
(
38) Hall, B.; Monot, R. Comput. Phys. 1991, 5, 414.
39) Yu, W. W.; Qu, L. H.; Guo, W. Z.; Peng, X. G. Chem. Mater.
2
(
003, 15, 2854.
40) Mohamed, M. B.; Tonti, D.; Al-Salman, A.; Chemseddine, A.;
Chergui, M. J. Phys. Chem. B 2005, 109, 10533.
41) Ji, X. H.; Copenhaver, D.; Sichmeller, C.; Peng, X. G. J. Am.
Chem. Soc. 2008, 130, 5726.
42) Pradhan, N.; Reifsnyder, D.; Xie, R. G.; Aldana, J.; Peng, X. G. J.
Am. Chem. Soc. 2007, 129, 9500.
43) Chen, Y. F.; Johnson, E.; Peng, X. G. J. Am. Chem. Soc. 2007,
29, 10937.
(
(
(
1
6
732
dx.doi.org/10.1021/ja5020025 | J. Am. Chem. Soc. 2014, 136, 6724−6732