Atom‐Resolved Investigation on Dynamic Nucleation and Growth of Platinum Nanocrystals (Small Methods 6/2022)

Abstract

Inside Front Cover In article number 2200171, Sun, Wang, and coworkers focused on the nucleation and growth mechanism of nanocrystals. The dynamic transformation of a solid‐phase model platinum precursor was investigated at atomic resolution. Those findings shall provide a deeper understanding of the solid‐phase nucleation mechanism. The in‐situ method can be applied to study the nucleation process of other systems.

Full text

2200171 (1 of 8) © 2022 Wiley-VCH GmbH
www.small-methods.com
ReseaRch aRticle
Atom-Resolved Investigation on Dynamic Nucleation
and Growth of Platinum Nanocrystals
Huanyu Ye, Feng Yang, Yinghui Sun,* and Rongming Wang*
H. Ye, Y. Sun, R. Wang
Beijing Advanced Innovation Center for Materials Genome Engineering
Beijing Key Laboratory for Magneto-Photoelectrical Composite
and Interface Science
School of Mathematics and Physics
University of Science and Technology Beijing
Beijing 100083, China
E-mail: yhsun@ustb.edu.cn; rmwang@ustb.edu.cn
F. Yang
Department of Chemistry
Southern University of Science and Technology
Shenzhen, Guangdong 518055, China
DOI: 10.1002/smtd.202200171
at the atomic scale, especially in an in
situ growth environment still remains
a challenge, because of the restric-
tions of spatial and temporal resolution
of in situ characterization techniques.
The nucleation process of crystals was
roughly described by the classical nuclea-
tion theory, revealing a one-step nuclea-
tion path where monomers overcome
a single free-energy barrier to form a
stacking nucleus.[3] This theory ascribed
the driving force for the crystal nucleation
and growth to the variation of Gibbs free
energy in an ideal thermodynamic system.
However, the mechanism of nucleation
and growth derived from the statistical
analysis and ex situ observation showed
a discrepancy with the theoretical predic-
tion.[4] Each nanocrystal and its nuclea-
tion process in dierent systems require
advanced characterization techniques to
study further. With the development of
in situ characterization technology, some
attractive mechanisms of nucleation and growth of nanocrys-
tals have been discovered,[5] such as the reversible disorder-to-
order transition of metastable gold (Au) clusters,[3b] two-step
heterogeneous nucleation of iron, Au, and rhenium,[6] and the
hybrid process of nucleation on twin boundaries and surface
diusion of atoms in the growth of Au shell on a platinum (Pt)
icosahedron.[7] In these studies, the change of free energy was
used to explain the growth mechanism, however, the dier-
ence that existed in the experimental results between dierent
systems made it confusing.
In recent years, advanced characterization techniques have
been rapidly developed, such as spherical aberration correc-
tion environment transmission electron microscopy (ETEM),[8]
atomic force microscopy,[9] X-ray diraction,[10] and synchrotron
X-ray technique,[11] which can reveal the details of crystal nuclea-
tion and growth in dierent materials. Among these tech-
niques, the sophisticated in situ transmission electron micros-
copy (TEM) can achieve high spatial and temporal resolutions
simultaneously. The chamber in an in situ TEM can act as a
mini reaction system, which enables the study of the dynamic
nucleation and growth process of nanocrystals at the atomic
scale.[1d,12] For example, the in situ liquid-phase TEM experiment
combined with quantitative analysis and simulation confirmed
a multi-step pathway for the nucleation and growth of Au and
silver (Ag) nanocrystals in solution.[13] The atomic concentration
and eective planar diusion coecients of Au nanoparticles
in solution were calculated based on the quantitative scanning
Understanding the mechanism of nucleation and growth of nanocrystals
is crucial for designing and regulating the structure and properties of
nanocrystals. However, the process from molecules to nanocrystals remains
unclear because of the rapid and complicated dynamics of evolution under
reaction conditions. Here, the complete evolution process of solid-phase
chloroplatinic acid during the electron beam irradiation triggered reduction
and nucleation of platinum nanocrystals is recorded. Aberration-corrected
environmental transmission electron microscopy is used for direct visualiza-
tion of the dynamic evolution from H2PtCl6 to Pt nanocrystals at the atomic
scale, including the formation and growth of amorphous clusters, crystalliza-
tion, and growth of clusters, and the ripening of Pt nanocrystals. At the first
two stages, there exists a critical size of ≈2.0nm, which represents the start
of crystallization. Crystallization from the center and density fluctuation are
observed in the second stage of the crystallization of a few clusters with a
size obviously larger than the critical size. The work provides valuable infor-
mation to understand the kinetics of the early stage of nanocrystal nucleation
and crystallization at atomic scale.
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/10.1002/smtd.202200171.
1. Introduction
Nucleation is a crucial issue in the crystallization process and
has significant influences on the morphology and the crystal
structure of nanocrystals, including defects, atomic steps,
active sites, etc. Therefore, the nucleation further modulates
the properties of functional nanomaterials, from catalysts to
semiconductors.[1] In the past few decades, most of the pro-
posed nucleation kinetics were investigated based on the
speculation from macro-experiments, theoretical calcula-
tions, and model predictions.[2] The dynamic characterization
Small Methods 2022, 6, 2200171

© 2022 Wiley-VCH GmbH
2200171 (2 of 8)
www.advancedsciencenews.com www.small-methods.com
transmission electron microscopy (STEM) high angle annular
dark field (HAADF) image analysis.[13] Non-classical nucleation
processes aected by structure defects[1c,14] and atomic steps[7]
have also been revealed by in situ TEM. The two-step nucleation
mechanism identifying the existence of amorphous intermedi-
ates was discussed under solid[15] and liquid phases.[10b]
In this work, we in situ investigated the evolution process
of the decomposition of solid-phase Pt precursor compound,
the nucleation and growth into Pt nanocrystals at atomic
resolution by ETEM. We chose a beam-sensitive solid-phase
chloroplatinic acid (H2PtCl6) as the precursor and performed
the e-beam irradiation to induce decomposition and reduction
of H2PtCl6. We captured and identified the early stage of the
structure collapse of precursor, the formation of intermediate
amorphous clusters, and the nucleation and crystallization
of Pt nanocrystals at the atomic scale. The mechanism of
the evolution of amorphous clusters and the formation of Pt
nanocrystals were investigated by in situ aberration-corrected
TEM and STEM-HAADF observation coupled with semi-
quantitative analysis. Our results confer a deep understanding
of the kinetics of nanostructure transformation and provide
a promising way for in situ observation of the solid-phase
decomposition, nucleation, and growth process with high spa-
tial and temporal resolutions.
2. Results and Discussion
2.1. In Situ Observation of the Nucleation and Crystallization
Process
The H2PtCl6 was used as the precursor and dispersed the ethanol
dispersion onto a TEM silicon nitride (SiNx) film. The TEM elec-
tron beam was adopted to decompose and reduce the H2PtCl6
crystals to form metallic Pt nanocrystals.[14] To regulate the decom-
position rate of H2PtCl6, the dose rate was precisely controlled
from 14 to 118 e Å–2s–1 to ensure the atomic spatial resolution and
millisecond temporal resolution. When the dose rate was lower
than 14 e Å–2s–1, no significant change occurred in the H2PtCl6
crystal after 10-min irradiation. While for the dose rate higher
than 118 e Å–2s
–1, the precursor decomposed into Pt nanocrys-
tals too rapid to observe the initial stage of nucleation. Figure S1
(Supporting Information) schematically illustrates the nuclea-
tion dynamics transition from H2PtCl6 precursor to metallic Pt
nanocrystal.
Figure1a shows a typical TEM image of the H2PtCl6 crystal
distributed on SiNx film at room temperature. It is found that
the large nanoparticles distributed around the holes of SiNx
film with a size of 100–300nm provide a good sight for in situ
observation. The diraction contrast in the bright-filed TEM
Figure 1. Structure characterization and element distribution of the H2PtCl6 crystal. a) Bright-field TEM image of H2PtCl6 crystallites at room tempera-
ture in a high vacuum. b) A zoom-in TEM image of a square-like plate of precursor from a region marked in (a). c) Diraction pattern of the H2PtCl6
crystallite in (b) along [001] zone axis. d,e) Elemental mappings of the H2PtCl6 crystal. f) Schematic atomic model of H2PtCl6 with space group of
3Fm m
.
Small Methods 2022, 6, 2200171

© 2022 Wiley-VCH GmbH
2200171 (3 of 8)
www.advancedsciencenews.com www.small-methods.com
image (Figure1b) indicates the crystal nature of the nanopar-
ticle with a size of 370nm. The nonuniform diraction contrast
in Figure1b reveals the existence of the local crystal defects in
the crystallite. The corresponding selected area electron dirac-
tion (SEAD) pattern (Figure1c) confirmed the cubic structure of
H2PtCl6 (space group:
3Fm m
[16]) along the [001] direction with
exposed {100} facets as shown in Figure 1b. Elemental map-
ping of Figure1b obtained using energy-dispersive X-ray spec-
troscopy (EDS) in STEM mode exhibit the uniform distribution
of Pt and chlorine (Cl) throughout the crystallite (Figure1d,e).
Figure1f shows the atomic model of H2PtCl6. Previous work
has shown that the weak ionic bond between Pt4+ and Cl− can
be broken to form Pt monomer under e-beam irradiation.[14]
Thus Pt nanocrystals can form under the preset dose irradia-
tion of a high-energy electron beam and the reaction rate can be
accurately controlled by tuning the beam dose rate.
In the in situ experiment, we chose a typical dose rate of
118 e Å–2s
–1 to reduce the H2PtCl6. Based on our observa-
tion in Figure2a–d and Movie S1 (Supporting Information),
the evolution from H2PtCl6 to Pt nanocrystals showed a dis-
tinct feature of the multi-step process. At the initial stage, the
H2PtCl6 decomposed under the e-beam irradiation and formed
a great amount of tiny clusters. The clusters with sizes of <1nm
were observed in Figure 2a, which are partially marked with
yellow shadows for better illustration. As shown in Figure S2
(Supporting Information), the electron energy loss spectroscopy
(EELS) intensity of Cl L2,3-edge gradually decreased with reac-
tion proceeding, implying the main structure of the amorphous
clusters was composed of PtClx (x<6, decreased with reaction
proceeding), and the dose broke the ionic bond between Pt4+
and Cl−. As shown in Figure 2b, some tiny clusters (<1 nm)
grew to larger ones (1–2nm) by coalescence and aggregation.
When the reaction proceeded, in situ observation revealed the
amorphous clusters gradually show clear lattice fringes, thus we
considered the reaction entered the second stage. The typical
feature of this stage is the crystallization of metastable clus-
ters. We measured the sizes of clusters and found sizes of the
clusters are mostly in the range of 1.8–2.2nm. The statistical
analysis of the cluster size will be discussed later. As shown in
Figure2c, most clusters became stable crystallites and few clus-
ters remained amorphous. At this stage, we captured various
crystallization behaviors and they cannot be fully described by
the classical crystal nucleation theory.[17] After the crystallization
and growth process, the ripening process of Pt crystallites was
observed, similar to Ostwald ripening and orientation attach-
ment as reported.[18] This process can be considered the final
stage of the reaction. Figure2d shows the polycrystalline nano-
particles with clear lattice fringes of ≈0.23nm attributed to the
Pt (111) plane, which indicates that the nanocrystals are metallic
Pt without the formation of any other phase (e.g., alloy or metal
salt). These results revealed the formation of amorphous clus-
ters was a crucial intermediate state during the nanocrystal
nucleation and crystallization process.
Advanced spatial and temporal observation and quantitative
analysis were further performed to reveal the growth mecha-
nism. The structural evolution of the region marked with a
dashed blue square in Figure 2a was recorded to reveal the
details of the early stage of nucleation and crystallization pro-
cess. The high resolution TEM (HRTEM) images and cor-
responding fast-fourier transform (FFT) patterns at dierent
times are shown in Figure2e–m. As shown in Figure2e–h, the
HRTEM images revealed the precursor decomposed to form tiny
Figure 2. In situ observation of the nucleation and growth process of Pt nanocrystals under e-beam irradiation. a–d) Sequential TEM images of typical
stages for the precursor evolution. Some typical clusters are marked with yellow shadows: i) 0–125 s, the precursor collapsed to form sub-nano clus-
ters and grew gradually; ii) 125–1090 s, the clusters entered a metastable state and unstable nucleation appeared. e–m) Sequential HRTEM images
and corresponding FFT patterns of the area marked with the dashed blue frame in (a). The dierent color of the curve shows the clusters evolution
process at the atomic scale.
Small Methods 2022, 6, 2200171

© 2022 Wiley-VCH GmbH
2200171 (4 of 8)
www.advancedsciencenews.com www.small-methods.com
clusters and then gradually grew larger. The diusion FFT pat-
terns indicate the amorphous nature of the clusters was main-
tained at this stage. The average sizes of the clusters marked
with yellow dashed curves in Figure2f–h are measured to be 1.0,
1.8, and 2.2nm, respectively, indicating a rapid growth process
of the clusters in the early stage of the reaction. This process cor-
responds to the aggregation and Ostwald ripening mechanism.
In Figure 2i–k, the lattice fringes gradually grew clear, which
means the nucleation and crystallization took place. The meas-
ured average size of clusters reached 2.0nm. Most of the clusters
remained similar sizes and shapes compared with their amor-
phous counterparts. In Figure2i, most clusters had sizes close
to 2.0nm that are labeled as “1,” and a small number of larger
clusters (a typical one labeled as “2”). During this stage, dierent
ways of nucleation and crystallization processes between 1 and
2 were captured. Cluster 1 with a size close to 2.0nm entirely
crystallized into a nanocrystal, while cluster 2 with a cashew-
like shape crystallized step-by-step from one end to the entire
structure. As shown in the FFT pattern of Figure2i–k, the Pt
(111) and (200) diraction spots gradually appeared and became
stronger. These evolutions of FFT patterns indicate the crystal-
linity of the particles increased gradually and more stable crystal
planes formed during the e-beam irradiation.
Figure2m presents a typical HRTEM image of ripening Pt
nanocrystals. The size and morphology of the nanocrystals in
the ripening process changed at a limited rate. Little migration
of the nanocrystals was observed by comparing the distribution
of the Pt nanocrystals in Figure 2k,m. As illustrated in
Movie S1 (Supporting Information), the coalescence and orien-
tation attachment of Pt nanocrystals were also observed. As a
typical HRTEM image shown in Figure S3 (Supporting Infor-
mation), the nanocrystal is identified to be Pt along [110] zone
axis with exposed {111}, {200}, and {220} facets.
Dierent from the classical nucleation and growth process
previously reported,[19] our results (Figure2) indicate a three-
step formation and growth process of Pt nanocrystals. At the
first stage, sub-nanometer amorphous clusters formed by the
decomposition of H2PtCl6 precursors, then they grew into
larger ones with a size of ≈2.0nm by the migration and coales-
cence. At the second stage, the amorphous PtClx clusters were
reduced into Pt nanocrystals. At the final stage, the ripening of
the nanocrystals was observed. Similar reaction processes were
reproduced under the doses of 14 and 71 e Å–2s–1 (Figures S4
and S5, Supporting Information).
In the classical nucleation theory, the existence of a critical
size has been widely accepted, i.e., when a nucleus radius is
larger than a critical size, the crystal nucleus tends to grow,
whereas the crystal nucleus dissolves when its radius is smaller
than the critical size.[3a,20] Owing to the complexity of the nucle-
ation environment, however, some experiments are dierent
from the classical nucleation theory.[5a] Our in situ observa-
tion indicated that the critical size for the crystallization of Pt
clusters is ≈2.0nm.
2.2. Growth Kinetics from Amorphous Clusters to Nanocrystals
In our experiment, the electron beam acts as the source
of in situ observation and the external trigger of reaction
simultaneously. The high-energy electron beam might intro-
duce additional influence on the reaction process, including
growth rate, crystallization size, etc. Therefore, we performed
quantitative statistical analysis on the cluster size during the
growth under three dierent dose rates to evaluate the growth
kinetics of Pt clusters (Figure3). Distinct lattice fringes with a
d-spacing of ≈0.23nm were observed during the crystallization
stage, suggesting the formation of Pt {111} planes (Figure3a–c).
As indicated before, most of the cluster radius was measured to
be 1.8–2.2nm at this stage. To quantitatively descript the rela-
tionship between the cluster size and the irradiation dose, we
measured the average radius of more than 100 clusters for each
point (Figure3d). The dierence at the starting points of the
average size is due to the fact that the clusters with sizes of less
than 0.5nm are dicult to be distinguished. During the in situ
observation, we observed sub-nano amorphous clusters under
the dose rates of 71 and 118 e Å–2s
–1 at the beginning of the
experiment. Under the lower dose rate (14 e Å–2s–1), however,
it cost ≈250 s for the amorphous clusters to be distinguished,
implying an appropriate dose of electron beam irradiation
for the decomposition of the precursor. As the reaction pro-
gressed, the statistical average size increased with irradiation
time and the growth rate slowed down at the appropriate time
for each irradiation dose rate. Then the growth of the cluster
can be divided into two stages. For each stage, the average size
was increased nearly linearly. It is interesting that the intersec-
tion points of the linear fitting for each dose rate were found
to be almost similar (green zone in Figure3d). We denoted the
median value of the green zone (≈2.0nm) as critical size, which
might correspond to dierent growth boundaries. At the begin-
ning of the observation, no lattice fringes were distinguished,
indicating the amorphous nature of the clusters at this stage.
With the reaction proceeds, the lattice fringes of the clusters
started to be clear, indicating the initiation of crystallization. It
is interesting to see that most cluster sizes are about 2.0nm,
similar to the critical size. Then the critical size can be regarded
as the boundary between the first two stages of the growth of Pt
nanocrystals. From the linear fitting of the data before and after
the critical size, the growth rates of the clusters were extracted
and marked by the solid lines in Figure 3d. The growth
rates of Pt nanocrystals in the first stage were 1.64, 2.35, and
2.75 Å min–1 under the dose rates of 14, 71, and 118 e Å–2s–1,
respectively. Then the growth rates decreased to 0.48, 0.68, and
0.76 Å min–1 in the second stage, respectively. After crystalliza-
tion, the growth rate decreased to 27–30% under all the dose
rates conditions.
The classical nucleation theory describes a one-step nuclea-
tion process: monomers assemble by overcoming the surface
and volume free-energy barrier through density fluctuation.[3b]
Because of the limited characterization technique, the early
stage of cluster growth was rarely investigated. With the atomic
resolution in situ observation, we found that there were two
evolution stages before classical nucleation take place, including
amorphous cluster growth and crystallized process. As the
results of statistical analysis and in situ observation, we con-
sider that the overall process took place in three steps and the
details can be confirmed. During the first stage, a great number
of monomers from the decomposition of the precursor merged,
resulting in the fast growth of the amorphous clusters. When
Small Methods 2022, 6, 2200171

© 2022 Wiley-VCH GmbH
2200171 (5 of 8)
www.advancedsciencenews.com www.small-methods.com
the average size of clusters reached the critical size (≈2.0nm),
the growth rate of clusters slowed down significantly while the
crystallization process begins, indicating the second stage of
the reaction. The growth rate at the first stage was about four
times fast than the second stage, the observation of coalescence
and aggregation of nanocrystals, indicating the reaction enters
into the third stage. During this stage, nanocrystals aggregate
into larger ones, as in previous reports.[21] For most of the amor-
phous clusters, it is dicult to reveal the evolution details from
amorphous to crystal structure due to the rapid crystallization
process around the critical size. The crystallization process may
slow down with the increase of cluster size. Then we choose a
few amorphous clusters with sizes apparently larger than the
critical size for the investigation of crystallization dynamics.
2.3. Crystallization Dynamics of Amorphous Clusters
Most studies on crystal nucleation at atomic scale mainly use in
situ HRTEM in liquid-cell.[22] Density fluctuation of precursor in
solution and nanocrystal growth via the decomposition, solidifica-
tion, and crystallization processes have been observed.[13] In our
work, the crystallization process of amorphous clusters ≈4.0nm
was studied using in situ HRTEM and high-resolution STEM
(HRSTEM) in ETEM. Figure4a reveals a typical amorphous
cluster with a size of ≈3.7nm, marked by the yellow dashed curve.
The dispersive diraction ring in the corresponding FFT pattern
demonstrated the amorphous nature of this cluster. Figure4b–f
shows the crystallization process of the amorphous cluster at a
dose rate of 118 e Å–2s–1. It was found that the cluster shrunk and
rounded gradually, like the molten droplet. The low-pass inverse
FFT method was used to enhance the mass-thickness contrast
of the HRTEM images, as shown in Figure S6 (Supporting
Information). This method transforms the HRTEM image into
a digital bright-field image. Benefit from this method, a deeper-
contrast region with a size of ≈1.6nm was found to appear inside
the amorphous cluster at 80 s, as indicated by the green dashed
curve in Figure4c. Then ≈1.7nm Pt nanocrystal formed in the
green region at 130 s, as shown in Figure4d. Lattice fringes with
a d-spacing of 0.22nm were well resolved, indicating the forma-
tion of Pt crystal nuclei. As the reaction proceeds, the 2D lattice
fringes of the Pt crystal with similar nuclei size to the previous
one can be distinguished (Figure 4e, f). Movie S2 (Supporting
Information) provides the in situ observation details of the
dynamic process of crystallization of amorphous clusters.
Figure4g shows a typical HAADF-STEM image of an amor-
phous cluster with a size of ≈4.1 nm. The time-sequential
HAADF images with normalized intensity during the reac-
tion are shown in Figure 4h–p. The bright intensity in
HAADF-STEM reflects the thickness of the atom columns. A
high-intensity region ≈1.6nm in size was observed within the
amorphous cluster (Figure4h). The size of the high-intensity
Figure 3. Typical HRTEM images of clusters from amorphous to crystal structure and statistical cluster sizes under dierent dose rates versus reaction
time. a–c) HRTEM images of Pt nanoclusters under doses of 14, 71, 118 e Å–2s–1, respectively. d) Statistics of the average cluster size versus irradiation
time under the dose rates of 14, 71, and 118 e Å–2s–1, respectively. The solid curves indicated the fitting lines. Two growth rates for each dose rate were
found with a cross-point at a similar average size of ≈2.0nm. Each point is based on the statistical results of 100 clusters’ sizes and has conformed to
a normal distribution. Data expressed as the mean ± standard deviation values.
Small Methods 2022, 6, 2200171

© 2022 Wiley-VCH GmbH
2200171 (6 of 8)
www.advancedsciencenews.com www.small-methods.com
region in Figure4h is similar to that of the nuclei in Figure4c,
but smaller than the critical size (≈2.0 nm) in Figure 3.
Dierent from the crystallization process in Figure 4a–f, the
high-intensity region in Figure4h went through a dispersion
and re-aggregation process in ≈60 s, as shown in Figure4h–p.
The re-aggregation high-intensity region grew to ≈2.0 nm, as
shown in Figure 4p. The evolution of the HAADF intensity
implied the density fluctuation within a larger cluster that we
have discussed in Figures S7 and S8 (Supporting Information).
Further HRTEM investigation indicated that the Pt nanocrystal
formed in this region.
Based on the above investigation, the nucleation and growth
process of Pt nanocrystals can be divided into three stages:
i) the formation and growth of amorphous clusters, ii) crys-
tallization and growth of clusters, and iii) the ripening of
nanocrystals, which was schematically illustrated in Figure5.
At the first stage, the weak ionic bond between Pt and Cl ions
in the precursor was broken by the electron beam, and PtClx
(x<6, decreased as the reaction proceeded) intermediate amor-
phous clusters formed (Figure 5a,b). Then the clusters grew
rapidly by merging neighboring PtClx clusters until the average
size reaches the critical size of ≈2.0 nm (Figure 5c). Most of
the amorphous clusters with the critical size crystalized very
quickly (<0.5 s), while a few of them aggregated into lager
amorphous clusters. The nuclei with size ≈1.6 nm from the
center and density fluctuation were observed in the amorphous
cluster with a size ≈4 nm (Figure 5d). The third stage of the
nanocrystal growth by aggregation and coalescence is similar to
previous work, which has been extensively discussed.[23] Com-
pared with the classic LaMer's theory, our proposed process
makes it explicit that the aggregation of PtClx forms amorphous
clusters, and only when the clusters get large enough can they
crystalize and form nanocrystals. But in LaMer's model, there
is no discussion about crystal structure evolution from the
monomer aggregation to the nucleation.
This three-stage nucleation and growth behaviors of Pt
nanocrystals are supposed to be common in solid-phase reac-
tions and thus our work shall provide a deeper understanding
of the solid-phase nucleation mechanism. Combined with
the kinetic analysis of the process from amorphous clusters
to nanocrystals, it can also be instructive to the synthesis of
nanoparticles. Meantime, the in situ techniques with the high
spatial and temporal resolution are demonstrated to be eec-
tive for investigating and controlling the dynamic process of
nanocrystal growth. When combined with multi-reaction envi-
ronments, such as mechanics, electrics, and lighting, the in situ
observation could provide more information on the reaction
path and modulate the nucleation process of nanomaterials.
Figure 5. Schematic of the entire process of the early stage of nucleation process of H2PtCl6 into Pt nanocrystals. a) Atomic structure of H2PtCl6.
b) Amorphous intermediate clusters. c) Further grow into larger clusters with critical size 2.0nm. d) Nucleation and crystallization processes of Pt
nanoclusters with dierent sizes.
Figure 4. Structure evolution of a larger amorphous cluster. a–f ) Time-sequential TEM images of a typical larger amorphous cluster marked by a yellow
dashed curve showing the initial nucleation and crystallization process from 0 to 180 s. A deep-contrast area indicated by the green dashed curve
implied the location of the initial crystallization taking place between 80 and 180 s. Scale bar 1nm. g) A typical larger cluster STEM-HAADF image with
3.7nm in size. h–p) Time-sequential STEM images with the normalization of intensity show the density fluctuation in (g).
Small Methods 2022, 6, 2200171

© 2022 Wiley-VCH GmbH
2200171 (7 of 8)
www.advancedsciencenews.com www.small-methods.com
3. Conclusion
In summary, the e-beam was used to reduce the solid-phase
H2PtCl6 to Pt nanocrystals. The early two stages of crystal
nucleation and growth, i.e., formation and growth of amor-
phous clusters, crystallization, and growth of clusters, were
found before the nanocrystal aggregation and coalescence
stage. The in situ HRTEM observation and quantitative anal-
ysis revealed the existence of a critical size ≈2.0 nm, which
separate the first and second stage. The H2PtCl6 decomposes
into tiny PtClx amorphous clusters (<0.5nm) and grows larger
until reaching the critical size by aggregation or merging at
the first stage. Then most of the clusters crystallized quickly
(<0.5 s) and grow at the second stage. The growth rates of
the clusters were found to be dependent on the growth stage
and e-beam dose rate. The growth rate at the second stage
is slowed down by 27–30% compared with that at the first
stage. A few amorphous clusters with size apparently larger
than the critical size with density fluctuation inside and crys-
tallization from the center were found. Our work provides a
new pathway of nucleation and growth of metal nanocrystal,
which will benefit the design, fabrication, and application of
nanomaterials.
4. Experimental Section
Sample Preparation: In a typical experiment, H2PtCl6 (AR, Pt ≥ 37.5%, 1g)
was dissolved in ethanol (50 mL) and dispersed onto the SiNx film
of a TEM grid. After drying for 24 h in a vacuum, the H2PtCl6 crystal
precipitated on the SiNx film.
In Situ Observation: The experiments were carried out in an FEI Titan
G2 80–300 ETEM operated at 300 kV with an objective lens spherical
aberration corrector. The vacuum of the experimental environment was
in an order of 10−7Pa at room temperature. The integration time of TEM
images was adjusted from 0.5 to 10 s, depending on dierent dose
rates. The crystal of H2PtCl6 along the (001) direction was chosen for
the subject. Dark field STEM images were recorded by an ADF detector
and the collection semi-angle between 50 and 200 mrad. The probe
size of the e-beam was less than 0.5nm, which scanned a frame with
1024 × 1024 pixels in 8.05 s. An individual frame of HAADF image was
acquired with a pixel size of 84.1 pm.
Statistical Analysis: OriginPro Learning Edition software was used, and
the data are expressed as the mean ± standard deviation values. Each
point of Figure3d is based on the statistical mean of 100 clusters sizes
and conformed to a normal distribution. The quantified average size for
three dose rates was tested by a two-sided t-test. A p-value <0.05 was
considered statistically significant.
Supporting Information
Supporting Information is available from the Wiley Online Library or
from the author.
Acknowledgements
This work was supported by the National Key Research and Development
Program of China (No. 2018YFA0703700), the National Natural Science
Foundation of China (Nos. 51971025, 12034002 and 11974041), Beijing
Natural Science Foundation (Grant No. 2212034), and the Fundamental
Research Funds for the Central Universities (FRF-TP-18-075A1).
Conflict of Interest
The authors declare no conflict of interest.
Data Availability Statement
The data that support the findings of this study are available from the
corresponding author upon reasonable request.
Keywords
amorphous cluster, growth, in situ transmission electron microscopy,
nanocrystals, nucleation
Received: February 7, 2022
Revised: March 5, 2022
Published online: March 24, 2022
[1] a) Z. Ou, Z. Wang, B. Luo, E. Luijten, Q. Chen, Nat. Mater.
2020, 19, 450; b) M. H. Bistervels, M. Kamp, H. Schoenmakers,
A. M. Brouwer, W. L. Noorduin, Adv. Mater. 2021, 34, 2107843; c)
K.Cao, S. T.Skowron, C. T.Stoppiello, J.Biskupek, A. N.Khlobystov,
U.Kaiser, Angew Chem Int Ed Engl 2020, 59, 22922; d) X.Zhang,
S. Han, B. Zhu, G. Zhang, X. Li, Y.Gao, Z. Wu, B. Yang, Y. Liu,
W. Baaziz, O. Ersen, M. Gu, J. T.Miller, W.Liu, Nat. Catal. 2020,
3, 411; e) C.Chen, Y. Kang, Z.Huo, Z.Zhu, W. Huang, H. L. Xin,
J. D.Snyder, D.Li, J. A.Herron, M.Mavrikakis, M.Chi, K. L.More,
Y. Li, N. M.Markovic, G. A.Somorjai, P. Yang, V. R.Stamenkovic,
Science 2014, 343, 1339.
[2] a) Y.Wang, J. L.Li, W. X.Shi, Z. M. Zhang, S.Guo, R.Si, M.Liu,
H. C. Zhou, S. Yao, C. H. An, T. B. Lu, Adv. Energy Mater. 2020,
10, 2002138; b) N. de Jonge, F. M.Ross, Nat. Nanotechnol. 2011,
6, 695; c) T. J. Woehl, J. E. Evans, I. Arslan, W. D. Ristenpart,
N. D.Browning, ACS Nano 2012, 6, 8599; d) E.Arslan Irmak, P.Liu,
S.Bals, S.Van Aert, Small Methods 2021, 5, 2101150; e) R. M.Wang,
O.Dmitrieva, M.Farle, G.Dumpich, H. Q.Ye, H.Poppa, R.Kilaas,
C.Kisielowski, Phys. Rev. Lett. 2008, 100, 017205.
[3] a) V. K. LaMer, R. H. Dinegar, J. Am. Chem. Soc. 1950, 72, 4847;
b) S.Jeon, T.Heo, S. Y.Hwang, J.Ciston, K. C.Bustillo, B. W.Reed,
J. Ham, S. Kang, S. Kim, J. Lim, K. Lim, J. S. Kim, M. H. Kang,
R. S.Bloom, S.Hong, K.Kim, A.Zettl, W. Y.Kim, P.Ercius, J. Park,
W. C.Lee, Science 2021, 371, 498.
[4] a) J. De Yoreo, Nat. Mater. 2013, 12, 284; b) D.Sun, C. Shang,
Z.Liu, X.Gong, Chin. Phys. Lett. 2017, 34, 026402.
[5] a) R. P. Sear, Int. Mater. Rev. 2013, 57, 328; b) Y. Sun, H. Zhao,
D. Zhou, Y. Zhu, H. Ye, Y. A.Moe, R. Wang, Nano Res. 2019, 12,
947; c) J.Zhang, Y. Jiang, Q. Fan, M.Qu, N. He, J. Deng, Y.Sun,
J.Cheng, H. G.Liao, S. G.Sun, Small Methods 2021, 5, 2001234.
[6] K. Cao, J. Biskupek, C. T. Stoppiello, R. L. McSweeney,
T. W.Chamberlain, Z. Liu, K. Suenaga, S. T.Skowron, E. Besley,
A. N.Khlobystov, U.Kaiser, Nat. Chem. 2020, 12, 921.
[7] J.Wu, W.Gao, J.Wen, D. J.Miller, P.Lu, J. M.Zuo, H. Yang, Nano
Lett. 2015, 15, 2711.
[8] a) B.He, Y.Zhang, X.Liu, L. Chen, ChemCatChem 2020, 12, 1853;
b) R.Wang, Nat. Catal. 2020, 3, 333; c) F.Yang, H.Zhao, X.Wang,
X.Liu, Q.Liu, X.Liu, C.Jin, R.Wang, Y.Li, J. Am. Chem. Soc. 2019,
141, 5871.
[9] M.Sleutel, J.Lutsko, A. E.Van Driessche, M. A.Duran-Olivencia,
D.Maes, Nat. Commun. 2014, 5, 5598.
[10] a) J. A.Sellberg, C.Huang, T. A.McQueen, N. D.Loh, H.Laksmono,
D. Schlesinger, R. G. Sierra, D. Nordlund, C. Y. Hampton,
Small Methods 2022, 6, 2200171

© 2022 Wiley-VCH GmbH
2200171 (8 of 8)
www.advancedsciencenews.com www.small-methods.com
D. Starodub, D. P. DePonte, M. Beye, C. Chen, A. V. Martin,
A. Barty, K. T. Wikfeldt, T. M. Weiss, C. Caronna, J. Feldkamp,
L. B. Skinner, M. M. Seibert, M. Messerschmidt, G. J. Williams,
S.Boutet, L. G.Pettersson, M. J.Bogan, A.Nilsson, Nature 2014,
510, 381; b) M. K.Bera, M. R.Antonio, J. Am. Chem. Soc. 2016, 138,
7282.
[11] a) S.-M. Bak, Z. Shadike, R. Lin, X. Yu, X.-Q. Yang, NPG Asia
Mater 2018, 10, 563; b) F. Lin, Y. Liu, X. Yu, L.Cheng, A. Singer,
O. G. Shpyrko, H. L. Xin, N. Tamura, C. Tian, T. C. Weng,
X. Q.Yang, Y. S.Meng, D.Nordlund, W.Yang, M. M.Doe, Chem.
Rev. 2017, 117, 13123.
[12] D. J. Banner, E. Firlar, P. Rehak, A. H. Phakatkar, T. Foroozan,
J. K. Osborn, L. V.Sorokina, S. Narayanan, T.Tahseen, Y.Baggia,
P. Král, T. Shokuhfar, R. Shahbazian-Yassar, Adv. Funct. Mater.
2021, 31, 2007736.
[13] N. D.Loh, S.Sen, M.Bosman, S. F. Tan, J.Zhong, C. A. Nijhuis,
P. Kral, P. Matsudaira, U. Mirsaidov, Nat. Chem. 2017, 9,
77.
[14] W.Gao, P.Tieu, C.Addiego, Y.Ma, J.Wu, X.Pan, Sci. Adv. 2019, 5,
eaau9590.
[15] L. Fei, X. Gan, S. M. Ng, H. Wang, M. Xu, W. Lu, Y. Zhou,
C. W.Leung, C.-L.Mak, Y.Wang, ACS Nano 2018, 13, 681.
[16] F.Rau, U.Klement, K.-J.Range, Zeitschrift fuer Kristallographie 1995, 210,
606.
[17] J. W.Gibbs, Am. J. Sci 1878, s3-16, 441.
[18] M. Tang, W.Yuan, Y.Ou, G. Li, R. You, S.Li, H. Yang, Z.Zhang,
Y.Wang, ACS Catal. 2020, 10, 14419.
[19] a) T. W.Hansen, A. T.Delariva, S. R.Challa, A. K.Datye, Acc. Chem.
Res. 2013, 46, 1720; b) Y.Ma, W.Gao, H. Shan, W.Chen, W.Shang,
P. Tao, C. Song, C. Addiego, T.Deng, X. Pan, J. Wu, Adv. Mater.
2017, 29, 1703460.
[20] H.Reiss, J. Chem. Phys. 1951, 19, 482.
[21] W.Yuan, D.Zhang, Y.Ou, K.Fang, B.Zhu, H.Yang, T. W.Hansen,
J. B.Wagner, Z.Zhang, Y. Gao, Y. Wang, Angew Chem Int Ed Engl
2018, 57, 16827.
[22] H. G. Liao, D. Zherebetskyy, H. Xin, C. Czarnik, P. Ercius,
H.Elmlund, M.Pan, L. W.Wang, H.Zheng, Science 2014, 345, 916.
[23] Y. Jiang, Y. Wang, Y. Y.Zhang, Z. F. Zhang, W. T.Yuan, C. H.Sun,
X.Wei, C. N.Brodsky, C. K.Tsung, J. X.Li, X. F.Zhang, S. X.Mao,
S. B.Zhang, Z.Zhang, Nano Res. 2014, 7, 308.
Small Methods 2022, 6, 2200171