Julia S van der Burgt1, Jaco J Geuchies1, Berend van der Meer1, Hans Vanrompay2, Daniele Zanaga2, Yang Zhang2, Wiebke Albrecht1, Andrei V Petukhov1,3, Laura Filion1, Sara Bals2, Ingmar Swart1, Daniël Vanmaekelbergh1. 1. Condensed Matter and Interfaces, Debye Institute for Nanomaterials Science, Soft Condensed Matter, Debye Institute for Nanomaterials Science, and Physical and Colloidal Chemistry, Debye Institute for Nanomaterials Science, Utrecht University, Utrecht 3508 TA, Netherlands. 2. Electron Microscopy for Materials Science, University of Antwerp, Antwerp 2000, Belgium. 3. Laboratory of Physical Chemistry, Department of Chemical Engineering and Chemistry, Eindhoven University of Technology, Eindhoven 5612 AZ, Netherlands.
Abstract
Colloidal CsPbBr3 nanocrystals (NCs) have emerged as promising candidates for various opto-electronic applications, such as light-emitting diodes, photodetectors, and solar cells. Here, we report on the self-assembly of cubic NCs from an organic suspension into ordered cuboidal supraparticles (SPs) and their structural and optical properties. Upon increasing the NC concentration or by addition of a nonsolvent, the formation of the SPs occurs homogeneously in the suspension, as monitored by in situ X-ray scattering measurements. The three-dimensional structure of the SPs was resolved through high-angle annular dark-field scanning transmission electron microscopy and electron tomography. The NCs are atomically aligned but not connected. We characterize NC vacancies on superlattice positions both in the bulk and on the surface of the SPs. The occurrence of localized atomic-type NC vacancies-instead of delocalized ones-indicates that NC-NC attractions are important in the assembly, as we verify with Monte Carlo simulations. Even when assembled in SPs, the NCs show bright emission, with a red shift of about 30 meV compared to NCs in suspension.
Colloidal CsPbBr3 nanocrystals (NCs) have emerged as promising candidates for various opto-electronic applications, such as light-emitting diodes, photodetectors, and solar cells. Here, we report on the self-assembly of cubic NCs from an organic suspension into ordered cuboidal supraparticles (SPs) and their structural and optical properties. Upon increasing the NC concentration or by addition of a nonsolvent, the formation of the SPs occurs homogeneously in the suspension, as monitored by in situ X-ray scattering measurements. The three-dimensional structure of the SPs was resolved through high-angle annular dark-field scanning transmission electron microscopy and electron tomography. The NCs are atomically aligned but not connected. We characterize NC vacancies on superlattice positions both in the bulk and on the surface of the SPs. The occurrence of localized atomic-type NC vacancies-instead of delocalized ones-indicates that NC-NC attractions are important in the assembly, as we verify with Monte Carlo simulations. Even when assembled in SPs, the NCs show bright emission, with a red shift of about 30 meV compared to NCs in suspension.
Over the
past decade, a lot
of research has been devoted to the properties and improvement of
hybrid organic–inorganic perovskite materials (e.g., CH3NH3PbX3, X = Cl, Br, or I) for solar
cell applications,[1] which has led to an
increase in efficiency from 3.8%[2] to over
20%.[3] This success has motivated researchers
all over the world to study other types of perovskite materials and
extend their optoelectronic applications to photodetectors,[4,5] light-emitting diodes,[6−8] and one-[9] and two-photon[10] pumped gain media for
lasers. Recently, cesium lead halide (CsPbX3) perovskite
nanocrystals (NCs) have been reported as a new and promising branch
of perovskites. These colloidal NCs can be synthesized with a facile
hot injection method and possess bright emission and readily tunable
opto-electronic properties.[11] In comparison
to the hybrid CH3NH3PbX3 compounds,
the all-inorganic CsPbX3 has a better temperature stability.[5] Moreover these NCs show narrow emission peaks
with a full width at half-maximum (fwhm) of 50–80 meV and exhibit
quantum yields of up to 90% without any extra passivation of the NC
surface.[12] The emission can be tuned over
the full visible spectrum by varying the composition of the X-anionic
sublattice, either directly or by postsynthetic anion-exchange reactions.[13−15]NCs with low polydispersity and high photoluminescence (PL)
quantum
yields are ideal building blocks for larger structures with predesigned
opto-electronic functionalities. In these superlattices, new properties
can arise from the electronic and/or magnetic coupling between constituent
NCs. For example, the new functionality that arises from the ordered
structures of NCs has been used already in magnetic devices[16] and for electronic applications.[17] Here, we report on the self-assembly of cubic-shaped
CsPbBr3 NCs into cuboidal supraparticles (SPs), each SP
consisting of several hundreds to thousands of NCs. Using in situ
X-ray scattering measurements, we show that the SPs nucleate and grow
in the solution upon increasing NC concentration or by addition of
a nonsolvent. The structure of the superlattices has been studied
with (scanning) transmission electron microscopy [(S)TEM], electron
diffraction (ED), and high-angle annular dark-field (HAADF) STEM tomography.
We show that the NCs form a simple cubic lattice and are atomically
aligned but not connected. We observed empty superlattice positions,
that is, vacancies similar to those in an atomic crystal, in the bulk
as well as on surface lattice positions. Vacancies positioned at the
lattice sites should be distinguished from delocalized vacancies predicted
for the hard-cube model.[18] Our results
thus indicate that attractive interactions between the constituent
NCs are important in the self-assembly process. We studied the PL
of individual SPs with confocal microspectroscopy. Despite the fact
that exciton energy transfer can occur, the SPs still show a bright
PL, red-shifted by 30 meV compared to a diluted NC solution.
Results
and Discussion
Formation of Cuboidal SPs
In general,
two methods can
be used to induce aggregation of NCs in solution: (1) solvent evaporation,
also shown by Kovalenko and Bodnarchuk,[19] which increases the NC concentration, and (2) addition of antisolvent,
increasing the potential energy per NC. We used methyl acetate as
an antisolvent because it is one of the few polar solvents that does
not damage the NCs.[20] The use of other
polar media, such as methanol, ethanol, acetone, and acetonitrile,
as an antisolvent leads to dissolution of the individual NCs. Both
methods give identical SPs, however larger SP sizes are obtained by
solvent evaporation (see the Supporting Information, Figure S4). The drawback of solvent evaporation is that the SP
formation takes rather long (>1 month at room temperature). When
we
consider self-assembly in a reasonable lab-time period of hours, care
has to be taken to operate the assembly under nearly reversible conditions,
as fast kinetics can lead to disordered, nonequilibrium structures.
Indeed, we observed large clusters with an irregular shape and disordered
structure on the NC length scale, when a relatively high amount of
antisolvent was added (see the Supporting Information, Figure S5) to speed up the SP formation. The structures that we
will discuss were obtained under nearly reversible conditions. A more
detailed description of the parameters, which were used during the
self-assembly process, can be found in the Supporting Information.Figure shows self-assembled NC SPs in a series of HAADF-STEM
images with increasing magnification. A low magnification image is
presented in Figure a. The SPs appear as large white cuboids on a dark background. At
higher magnifications, it becomes clear that each SP consists of several
hundreds of NCs (Figure b–d). Moreover, the columns of NCs are clearly distinguished
at the highest magnification image (Figure d). The degree of ordering inside the SP
is apparent through the observed contrast in the HAADF-STEM images,
hinting toward a simple cubic stacking of the NCs inside the cuboidal
SP. The NCs are clearly not atomically attached as has been observed
for superlattices of PbSe NCs,[21] but they
are still separated from each other most likely by their oleate and
oleylamine ligands. The alignment must hence be related to the dense
cubic stacking of the NC building blocks. The contrast in between
the particles seems to be blurred out slightly, which we tentatively
ascribe to some positional and rotational disorder of the NCs inside
the SP. More images of the formed SPs can be found in the Supporting Information, Figure S3. Although cubic
particles can theoretically achieve high packing fractions up to 100%,
they are not often encountered in nature. In the CsPbBr3 SPs discussed here, cubic symmetry emerges over three length scales:
on the scale of the atomic lattice, the NC building blocks in the
supra particle superlattice, and the cuboidal shape of the SP itself.
Figure 1
Structure
of CsPbBr3 NC SPs. (a) Low-magnification overview
image, showing a relatively low coverage of large, cuboid-shaped SPs.
(b) Zoom-in on the marked region in (a) showing a single SP with a
nearly cubic shape. (c) Zoom-in on the marked area in (b) showing
nearly aligned NCs in a dense cubic stacking. The individual NCs in
the SP are clearly visible. (d) Further zoom-in on the marked area
in (c). Clearly visible columns of NCs in the lattice of the SP, hinting
toward a simple-cubic packing of the constituent NCs inside the SP.
Structure
of CsPbBr3 NC SPs. (a) Low-magnification overview
image, showing a relatively low coverage of large, cuboid-shaped SPs.
(b) Zoom-in on the marked region in (a) showing a single SP with a
nearly cubic shape. (c) Zoom-in on the marked area in (b) showing
nearly aligned NCs in a dense cubic stacking. The individual NCs in
the SP are clearly visible. (d) Further zoom-in on the marked area
in (c). Clearly visible columns of NCs in the lattice of the SP, hinting
toward a simple-cubic packing of the constituent NCs inside the SP.To properly quantify the rotational
freedom of the CsPbBr3 NCs inside the SPs, we performed
ED experiments on the SP presented
in Figure . The selected
area ED pattern is shown in Figure a. The presence of well-defined diffraction spots,
instead of powder rings, already indicates that the particles are
(nearly) atomically aligned, with the three [001] type zone-axes directed
in the principal directions of the NC superlattice (i.e., one of the
NC {100} facets pointing upward). The diffraction pattern can be indexed
up to reflections from the {420} atomic planes. In Figure b, azimuthal traces at the
constant scattering vector q are presented for reflections
from the {100}, {110}, {200}, and {220} planes (depicted in red, blue,
green, and yellow, respectively). The 12 peaks are fitted with Gaussian
functions to obtain an average fwhm of 9.6° ± 1.4°,
which reflects an upper limit of their in-plane rotational freedom.
The high degree of atomic alignment of the NCs inside the SPs originates
from the densest possible stacking of NC building blocks with a uniform
cubic shape and size.
Figure 2
ED analysis of a single SP consisting of CsPbBr3 NCs.
(a) ED pattern acquired along the [001] NC zone axis of the SP shown
in Figure . The appearance
of well-defined diffraction spots indicate that the NCs are atomically
aligned. The diffraction spots can be indexed clearly up to reflections
from the {420} atomic planes. (b) Azimuthal traces at constant scattering
vectors q for different atomic reflections, as indicated
by the colored semicircles in (a). Fitting the 12 peaks with Gaussian
functions (solid lines) gives an average fwhm of 9.6° ±
1.4°, which provides an upper limit of the in-plane rotational
freedom of the NCs inside the SP. (c) Two-dimensional schemes of a
planar section perpendicular to the [001] superlattice axis, demonstrating
the effect of slight rotational disorder. The orange arrows represent
the [100] atomic axis of each of the nanocubes.
ED analysis of a single SP consisting of CsPbBr3 NCs.
(a) ED pattern acquired along the [001] NC zone axis of the SP shown
in Figure . The appearance
of well-defined diffraction spots indicate that the NCs are atomically
aligned. The diffraction spots can be indexed clearly up to reflections
from the {420} atomic planes. (b) Azimuthal traces at constant scattering
vectors q for different atomic reflections, as indicated
by the colored semicircles in (a). Fitting the 12 peaks with Gaussian
functions (solid lines) gives an average fwhm of 9.6° ±
1.4°, which provides an upper limit of the in-plane rotational
freedom of the NCs inside the SP. (c) Two-dimensional schemes of a
planar section perpendicular to the [001] superlattice axis, demonstrating
the effect of slight rotational disorder. The orange arrows represent
the [100] atomic axis of each of the nanocubes.It is also relevant to characterize the structure of the
NC assemblies
formed in the suspension before they are scooped and dried on a transmission
electron microscopy grid. Therefore, we performed in situ transmission
small/wide-angle X-ray scattering (SAXS, WAXS) measurements on pure
stable NC suspensions and on suspensions in which the NC assembly
was initiated by adding methyl acetate as an antisolvent. A schematic
of the experiment is shown in Figure a. A solution of NCs is added into a quartz capillary
with an outer diameter of 1.5 mm and placed inside a LINKAM stage,
positioned at a distance of 1 m from the SAXS detector. The WAXS detector,
collecting the atomic diffraction of the NCs, was positioned at the
inlet of the vacuum tube holding the SAXS detector. The SAXS and WAXS
patterns of a stable suspension are shown in Figure b. The SAXS region (light blue) only shows
form factor scattering from dispersed NCs. The WAXS signal (dark blue)
shows sharp diffraction peaks originating from the atomic perovskite
lattice of the NCs. A sample with a volume fraction of 0.2 methyl
acetate was measured after 3 days of incubation and is presented in Figure c. The SAXS pattern
shows clear Bragg peaks at positions of 0.55, 0.78, 1.1, and 1.23
nm–1. Their relative peak positions correspond to
1:√2:2:√5, which corresponds to scattering from the
{100}, {110}, {200}, and {210} lattice planes of a simple cubic lattice
of NCs. The measured NC–NC distance inside the SPs in suspension
is determined to be 11.4 ± 0.1 nm, corresponding to NCs still
separated by their oleate and oleylamine ligands. The fwhm of the
{100} reflection is 0.378 nm–1, which corresponds
to spatial extension of the periodic order of roughly 166.3 nm. The
SP formation is reversible, as the constituent NCs readily redisperse
upon diluting the solution of SPs (see the Supporting Information, Figure S6).
Figure 3
Transmission X-ray scattering of NC solutions
during the formation
of SPs. (a) Scheme of the experimental setup. A quartz capillary is
loaded with a solution of NCs and placed in a LINKAM stage, which
is located 1 m from the detector to collect the SAXS signal. The formation
of SPs can be initiated by addition of an antisolvent. (b) SAXS pattern
of the NC dispersion without addition of the antisolvent, showing
only form factor scattering of the individual NCs in solution. (c)
SAXS pattern of the diluted NC solution after 3 days of incubation
upon addition of 20% (v/v) of methyl acetate antisolvent; the Bragg
reflections indicate the formation of crystalline SPs in the solution.
The inset shows a zoom on the region with the Bragg peaks, which is
scaled by the form factor scattering from (b). The red lines indicate
the expected peak positions for a simple-cubic packing of the NCs
inside the SPs.
Transmission X-ray scattering of NC solutions
during the formation
of SPs. (a) Scheme of the experimental setup. A quartz capillary is
loaded with a solution of NCs and placed in a LINKAM stage, which
is located 1 m from the detector to collect the SAXS signal. The formation
of SPs can be initiated by addition of an antisolvent. (b) SAXS pattern
of the NC dispersion without addition of the antisolvent, showing
only form factor scattering of the individual NCs in solution. (c)
SAXS pattern of the diluted NC solution after 3 days of incubation
upon addition of 20% (v/v) of methyl acetate antisolvent; the Bragg
reflections indicate the formation of crystalline SPs in the solution.
The inset shows a zoom on the region with the Bragg peaks, which is
scaled by the form factor scattering from (b). The red lines indicate
the expected peak positions for a simple-cubic packing of the NCs
inside the SPs.Recent work by Bertolotti
et al. also showed that the CsPbBr3 NCs have the tendency
to stack in solution,[22] similar to concentrated
solutions of perovskite nanoplatelets.[23] They argue that the stacking direction occurs
most likely in the {100} NC direction, similar to the NCs in the SPs
in this work. Others have also found that these perovskite NCs readily
assemble into one-dimensional chains by addition of molecular clusters[24,25] or through ligand–solvent interactions.[26] Such stacks could hence be a precursor phase in the formation
of the SPs presented in this work. However, this remains to be verified
by performing in situ time-resolved SAXS and WAXS on the self-assembly
process, which is beyond the scope of this article.
NC Vacancies
inside a SP Studied with HAADF-STEM Tomography
and Monte Carlo Simulations
To study the geometric structure
of the NCs inside the SPs in more detail, a single SP was studied
using HAADF-STEM tomography. A series of projection images was acquired
with an angular range of −70° to 78° and a tilt increment
of 2°. Orthogonal to that, a second tilt series was acquired
from −76° to 76° to reduce the missing wedge of tilt
angles to a missing pyramid.[27] Using this
dual tilt series as an input for several mathematical reconstruction
algorithms, we found that the quality of the reconstruction was greatly
improved compared to the reconstruction that was obtained from just
using the one tilt series.A visualization
of the acquired tomogram is presented in Figure a. Unfortunately, we are not able to clearly
resolve all individual NC positions inside the SP; the missing pyramid
of tilt angles smears out the information in the z-direction of the cuboidal SP. Nonetheless, the orthoslices taken
through the center of the structure in Figure b–d show interesting features. In
all three orthogonal directions, we observe empty places on lattice
positions. They appear to be localized single point vacancies and
vacancy clusters. From the tomogram, we are able to estimate the volume
fraction of vacancies in the bulk of the SP of 3.3%, assuming cubic
symmetry of the NC lattice (see the Supporting Information for more details). Furthermore, we also performed
secondary electron-STEM to acquire information on the surface of the
SPs, which can be found in the Supporting Information (Figures S7 and S8). Also, on the surface of the SPs, we observe
localized NC vacancies and groups of missing NCs. We estimate that
the amount of missing surface NCs roughly equals 5.5 ± 1.5%.
Figure 4
Looking
inside SPs with HAADF-STEM tomography. (a) Visualization
of the three-dimensional reconstruction of a single SP, acquired by
rotating the sample over two orthogonal tilt series and reconstructed
using a SIRT algorithm. Orthoslices through the xy (b), xz (c), and yz (d) directions
of the tomogram reveal several localized NC vacancies and groups of
missing NCs in the bulk of the SP. All vacancies are clearly defined
on a lattice position.
Looking
inside SPs with HAADF-STEM tomography. (a) Visualization
of the three-dimensional reconstruction of a single SP, acquired by
rotating the sample over two orthogonal tilt series and reconstructed
using a SIRT algorithm. Orthoslices through the xy (b), xz (c), and yz (d) directions
of the tomogram reveal several localized NC vacancies and groups of
missing NCs in the bulk of the SP. All vacancies are clearly defined
on a lattice position.Interestingly, the localized nature of the NC vacancies is
in contrast
to previous Monte Carlo and molecular dynamics simulations for hard
cubes, that is, cubes that interact only through excluded volume interactions.
Smallenburg et al. showed that for hard cubes, vacancies manifest
themselves as a finite-length chain of particles along one of the
principal axes in the crystal.[18] The vacancy
is thus spread over many lattice positions. More specifically, in
this hard-cube system, the equilibrium concentration of vacancies
is very high (up to 6%) because of the fact that creating a vacancy
provides additional free volume for multiple nearby particles, which
increases the entropy of the crystal. Note that in a system of attractive
cubes, Rossi et al. observed localized NC vacancies, where colloidal
cubes of roughly 1 micron are self-assembled into simple cubic lattices
through addition of a depletant, which induces effective attractions
between the cube facets.[28] As such, we
propose that the localization and grouping of NC vacancies in our
experiments point toward attractive interactions between the NCs during
the self-assembly. Specifically, for sufficiently large NC–NC
attractions, the free energy will be minimized when the particles
in the vicinity of the vacancy are sitting on regular lattice positions,
as the cost of breaking cohesive bonds will dominate over the entropic
gain associated with vacancy delocalization.To understand the
observed localization of the vacancies, we performed Monte Carlo simulations
of cubes in the NVT ensemble, where N (the number of particles), V (the volume), and T (the temperature) are fixed, and the cube–cube
attraction can be changed from zero to several times the thermal energy.
As a minimal model that captures the repulsive core and attraction,
we model the interactions between the cubes as
a combination of a cubic hard-core repulsion and a spherically-symmetric
square well attraction originating from the center of the cube. The
range of the attractive square well was fixed to Δ = 1.14σ,
with σ is the edge length of the cubes, and we varied the well
depth βε, with β = 1/kbT. Thus, for βε = 0, our model reduces
to the hard-cube model with delocalized vacancies, and we can study
the effect of increasing the attraction strength βε on
the vacancy structure. While the nucleation of these SPs is described
in the Supporting Information (Figure S12),
we will now focus on the (de)localization of defects in the bulk of
the crystal. In these simulations, we start off from a perfect simple
cubic crystal at a fixed packing fraction of φ = 0.80 and initialized
a single vacancy by removing a single particle. In agreement with
ref (22) we observe
NC vacancies to be delocalized over many lattice positions for the
hard-cube system, βε = 0 (Figure a). Yet, upon introducing a slight attraction
between the cubes, βε = −0.5, the long delocalized
NC vacancies are not observed anymore (Figure b). Instead, the NC vacancies spread only
over a couple of lattice sites. Thus, already for a moderate attraction
strength between the particles, we observe the defect structure to
be dominated by the cohesive bonds between the particles. For slightly
larger attractions, βε = −1, we observe the particles
in the vicinity of the NC vacancies exclusively on lattice positions,
and no delocalized NC vacancies are observed (Figure c), similar to the tomography experiments
presented in Figure .
Figure 5
Monte Carlo simulations of (de)localized NC vacancies inside the
SPs as a function of the attraction strength. (a) ε = 0 kbT, the NC vacancy is delocalized
over many lattice positions. (b) ε = −0.5 kbT, long delocalized NC vacancies are
not observed anymore, only weak delocalization of the NC vacancies
over a maximum of a few lattice sites are observed. (c) ε =
−1 kbT, NC vacancies
are exclusively observed on lattice positions and do not delocalize.
The latter is also observed in the tomography experiments presented
in Figure .
Monte Carlo simulations of (de)localized NC vacancies inside the
SPs as a function of the attraction strength. (a) ε = 0 kbT, the NC vacancy is delocalized
over many lattice positions. (b) ε = −0.5 kbT, long delocalized NC vacancies are
not observed anymore, only weak delocalization of the NC vacancies
over a maximum of a few lattice sites are observed. (c) ε =
−1 kbT, NC vacancies
are exclusively observed on lattice positions and do not delocalize.
The latter is also observed in the tomography experiments presented
in Figure .
Optical Properties of SPs
Compared to NCs in Solution
Often, quantum dot (QD) solids
suffer from a reduced PL quantum yield,
as exciton energy transfer between the NCs in the solid enhances the
probability of nonradiative energy transfer. We found that CsPbBr3 SPs deposited on a glass slide remained strongly emissive,
with a PL quantum yield of 26%, as determined with an integrating
sphere. This is good news for several opto-electronic applications
requiring a QD solid as the optically active material. To study the
optical properties of individual SPs in more detail, we performed
confocal microspectroscopy. A dispersion containing the SPs was drop-cast
on a microscope slide, and the solvent of the dispersion was allowed
to evaporate. A droplet of immersion oil was placed on the sample,
and a second cover slide was placed on top of the sample. The results
are presented in Figure and in the Supporting Information, Figure
S13.
Figure 6
Confocal microspectroscopy on individual CsPbBr3 NC
SPs. (a) Confocal micrograph of a glass slide covered with SPs, which
appear as high intensity areas on a dark background. The circles indicate
the areas in which the PL signal was integrated. (b) PL spectra of
a large number of different regions in the confocal micrograph shown
in (a). The solid lines indicate raw data, whereas the dashed lines
are Gaussian fits to the data. (c) Comparing the peak intensity to
the PL energy, a red shift of roughly 30 meV is observed for the SPs
compared to the NC monolayer background.
Confocal microspectroscopy on individual CsPbBr3 NC
SPs. (a) Confocal micrograph of a glass slide covered with SPs, which
appear as high intensity areas on a dark background. The circles indicate
the areas in which the PL signal was integrated. (b) PL spectra of
a large number of different regions in the confocal micrograph shown
in (a). The solid lines indicate raw data, whereas the dashed lines
are Gaussian fits to the data. (c) Comparing the peak intensity to
the PL energy, a red shift of roughly 30 meV is observed for the SPs
compared to the NC monolayer background.Figure shows
the
PL spectra of many spots observed on the glass slide. Figure a shows the confocal micrograph
of a glass slide covered with SPs. The circled areas indicate the
regions where the PL signal was integrated. In Figure b, we present the peak energies of 30 spots
versus the intensity and observe a gradual red shift (shown in Figure c), which abruptly
becomes constant at a sufficiently high spot intensity, the overall
red shift being roughly 30 meV. A similar red shift is observed when
the PL spectra of dispersed NCs and dispersed SPs are compared (Supporting Information, Figures S14 and S15).
We hence conclude that the bright spots present regions of SPs and
the weak spots present individual NCs or smaller clusters.The
red shift observed in the emission of the SPs with respect
to individual NCs could have several causes: Kovalenko and Bodnarchuk
reported in their recent work[19] that the
relative increase in the dielectric constant of the NC environment
can cause a red shift if the electron and hole both become more delocalized.
Furthermore, the excitons in the NCs are in the weak confinement regime;
the electronic coupling between NCs inside a SP could thus also cause
a weak red shift of the PL spectrum. Third, exciton energy transfer
results in a spatial diffusion of the excitons over the superlattice,
always with a trend toward lower exciton energy. Di Stasio et al.
reported a similar shift of the PL in concentrated solutions of CsPbBr3 NCs and solid-state films.[29] This
effect is also shown in concentrated solutions of the cubic CsPbBr3 NCs and is reversible upon diluting the NC dispersion.[30] In our time-resolved PL experiments (Figure S14), we observe that the shift in the
PL peak position occurs faster than the time resolution of our setup
(∼250 ps), which is probably faster than the exciton energy
transfer. This indicates that energy transfer is not the cause for
the observed red shift. We must conclude here that although the red
shift occurring in CsPbBr3 NC SPs is well-established,
its main cause remains an outstanding question.In summary,
we have shown that CsPbBr3 nanocubes self-assemble
into SPs that have an overall cuboidal shape. In the superlattices,
the NCs are atomically aligned but separated by the capping ligands.
Vacancies form real point defects instead of being delocalized, indicating
that attractive interactions play a role in the self-assembly process.
The SPs show a bright PL, red-shifted by 30 meV with respect to that
of individual NCs. The fact that the obtained SPs are still highly
emissive makes them promising candidates for opto-electronic applications.
For example, they possibly can be used as microcavities to study the
confinement of the PL and lasing inside SPs.
Authors: Francesco Di Stasio; Muhammad Imran; Quinten A Akkerman; Mirko Prato; Liberato Manna; Roman Krahne Journal: J Phys Chem Lett Date: 2017-06-07 Impact factor: 6.475
Authors: Wiel H Evers; Bart Goris; Sara Bals; Marianna Casavola; Joost de Graaf; René van Roij; Marjolein Dijkstra; Daniël Vanmaekelbergh Journal: Nano Lett Date: 2012-11-07 Impact factor: 11.189
Authors: Ioannis Lignos; Stavros Stavrakis; Georgian Nedelcu; Loredana Protesescu; Andrew J deMello; Maksym V Kovalenko Journal: Nano Lett Date: 2016-02-04 Impact factor: 11.189
Authors: Biplab K Patra; Harshal Agrawal; Jian-Yao Zheng; Xun Zha; Alex Travesset; Erik C Garnett Journal: ACS Appl Mater Interfaces Date: 2020-06-30 Impact factor: 9.229
Authors: Muhammad Imran; Palvasha Ijaz; Dmitry Baranov; Luca Goldoni; Urko Petralanda; Quinten Akkerman; Ahmed L Abdelhady; Mirko Prato; Paolo Bianchini; Ivan Infante; Liberato Manna Journal: Nano Lett Date: 2018-11-12 Impact factor: 11.189
Authors: David Vila-Liarte; Maximilian W Feil; Aurora Manzi; Juan Luis Garcia-Pomar; He Huang; Markus Döblinger; Luis M Liz-Marzán; Jochen Feldmann; Lakshminarayana Polavarapu; Agustín Mihi Journal: Angew Chem Int Ed Engl Date: 2020-08-13 Impact factor: 16.823