Article https://doi.org/10.

1038/s41467-023-38237-7

Epitaxial growth of highly symmetrical

branched noble metal-semiconductor het-
erostructures with efficient plasmon-induced
hot-electron transfer

Received: 21 November 2022 Li Zhai 1,2,3,14, Sara T. Gebre4,14, Bo Chen 2,14, Dan Xu1,14, Junze Chen5,
Zijian Li 2, Yawei Liu4, Hua Yang2, Chongyi Ling 2, Yiyao Ge2, Wei Zhai2,
Accepted: 21 April 2023
Changsheng Chen6, Lu Ma7, Qinghua Zhang8, Xuefei Li1, Yujie Yan9,
1234567890():,;
1234567890():,;

Xinyu Huang 10, Lujiang Li2, Zhiqiang Guan2, Chen-Lei Tao1, Zhiqi Huang2,
Hongyi Wang2, Jinze Liang2, Ye Zhu6, Chun-Sing Lee 2, Peng Wang9,11,
Check for updates Chunfeng Zhang 10, Lin Gu12, Yonghua Du 7, Tianquan Lian 4 ,
Hua Zhang 2,3,13 & Xue-Jun Wu 1

Epitaxial growth is one of the most commonly used strategies to precisely

tailor heterostructures with well-defined compositions, morphologies, crystal
phases, and interfaces for various applications. However, as epitaxial growth
requires a small interfacial lattice mismatch between the components, it
remains a challenge for the epitaxial synthesis of heterostructures constructed
by materials with large lattice mismatch and/or different chemical bonding,
especially the noble metal-semiconductor heterostructures. Here, we develop
a noble metal-seeded epitaxial growth strategy to prepare highly symmetrical
noble metal-semiconductor branched heterostructures with desired spatial
configurations, i.e., twenty CdS (or CdSe) nanorods epitaxially grown on
twenty exposed (111) facets of Ag icosahedral nanocrystal, albeit a large lattice
mismatch (more than 40%). Importantly, a high quantum yield (QY) of
plasmon-induced hot-electron transferred from Ag to CdS was observed in
epitaxial Ag-CdS icosapods (18.1%). This work demonstrates that epitaxial
growth can be achieved in heterostructures composed of materials with large
lattice mismatches. The constructed epitaxial noble metal-semiconductor
interfaces could be an ideal platform for investigating the role of interfaces in
various physicochemical processes.

Heterostructures have attracted tremendous attention due to their energy conversion13,14, etc. To date, the seeded/templated epitaxial
physicochemical properties arising from their morphologies1–3, growth is the most commonly used strategy to precisely tailor the
interfaces4–6, and spatial arrangements of different components7,8. The hierarchical heterostructures15–19. Epitaxial growth of a secondary
precise control over their nanoscale structures is essential to under- material on a specific facet of the seed/template allows it to follow the
stand the structure-property correlation and enhance their perfor- crystallographic orientation of the seed/template17,20. Combined with
mance in various applications, such as electronics9,10, catalysis11,12, solar delicate control over the exposed facet and crystal phase of seeds/

A full list of affiliations appears at the end of the paper. e-mail: tlian@emory.edu; hua.zhang@cityu.edu.hk; xjwu@nju.edu.cn

Nature Communications | (2023)14:2538 1

Article https://doi.org/10.1038/s41467-023-38237-7

templates, this strategy has been successfully used to prepare various same rotational symmetry as the Ag icosahedral nanocrystal. As shown
hierarchical heterostructures with well-defined spatial structures21,22, in Supplementary Fig. 3a–f, the Ag icosahedral nanocrystal exhibits
interfaces23,24, crystal phases25–27, and programmable components28,29. twofold (C2), threefold (C3), and five-fold (C5) symmetrical Moiré pat-
However, the aforementioned seeded/templated epitaxial growth terns, which are consistent with the previously reported five-fold
is mainly used to prepare heterostructures composed of components multi-twinned icosahedral nanocrystals49. After the growth of CdS
with similar lattice structures and/or chemical bonding, e.g., metal- nanorods, the obtained Ag-CdS icosapods with C2, C3, C5 symmetrical
metal20,26,27, metal oxide-metal oxide30,31, metal chalcogenide-metal patterns (Supplementary Fig. 3g–i) have also been observed under
chalcogenide1,16,22,23,28, perovskite-perovskite29, and metal-organic fra- HAADF-STEM (Fig. 1e–g), which are consistent with the HRTEM images
mework (MOF)-MOF32,33 heterostructures due to the common (Supplementary Fig. 3j–l). In particular, when projected along its C3
requirement of similar structure (normally with lattice mismatch rotational axis (Fig. 1f), the Ag-CdS icosapod presents a sophisticated
smaller than 5%) for the epitaxial growth17,18,34. Therefore, it remains a star-like pattern with the observed eighteen branches radially exten-
great challenge for the epitaxial growth of heterostructures made of ded from the bright core, while the other two branches are unobserved
distinctly different materials, especially for the noble metal- since they are perpendicular to the projected plane. Moreover, a five-
semiconductor heterostructures, in which two materials could have fold pattern of the Ag-CdS icosapod with the C5 symmetry exhibits
a large lattice mismatch (normally larger than 20%) and different different contrast along the ten extending directions of branches, due
chemical bonding18,24,35–38. Commonly, the preparation of noble metal- to the overlapped images of two nanorods in the same direction when
semiconductor heterostructures relies on conventional non-epitaxial viewed along its C5 rotational axis (Fig. 1g). The HAADF-STEM and the
strategies including chemical deposition39,40, cation-exchange- corresponding energy-dispersive X-ray spectroscopy (EDS) elemental
facilitated non-epitaxial growth35,41, phase transfer42, photochemical mapping images (Fig. 1h) of a representative Ag-CdS icosapod reveal
deposition43, chemical extraction44,45, sol-gel method46, etc. Although that the Ag element is located in the core, while the Cd and S elements
the obtained non-epitaxial noble metal-semiconductor hetero- are homogeneously distributed over the twenty nanorods. X-ray
structures have shown great potential in diverse fields, such as photoelectron spectroscopy (XPS) was used to determine the valence
catalysis6,14,40,43,46, photovoltaic devices4,14,47 and sensors36,44,48, their state of components in Ag-CdS icosapods (Supplementary Fig. 4). In
wide application has been hindered due to difficulties in the precisely particular, the peaks located at 373.7 and 367.7 eV can be ascribed to
defined control over their structures and interfaces. Therefore, a reli- Ag 3d3/2 and Ag 3d5/2 of zerovalent Ag, respectively (Supplementary
able epitaxial growth method is urgently desired to rationally con- Fig. 4b). Combined XPS results with the Ag K-edge XANES spectra
struct hierarchical noble metal-semiconductor heterostructures with (Supplementary Fig. 5), it can be concluded that the Ag nanocrystals
precisely controlled architectures and well-defined interfaces. have not been sulfurized during the synthesis of Ag-CdS icosapods,
In this work, we report a controlled synthesis of highly symme- which might arise from the trioctylphosphine (TOP) used in the
trical branched noble metal-semiconductor heterostructures via a synthesis44,45,50. Furthermore, the X-ray diffraction (XRD) pattern of the
noble-metal-seeded epitaxial growth. The 1D II-VI semiconductor synthesized Ag-CdS icosapods can be well indexed to the wurtzite CdS
nanorods can be uniformly grown on the exposed (111) facets of zero- and fcc-Ag (Supplementary Fig. 6). The aforementioned results clearly
dimensional platonic Ag nanocrystals. The obtained heterostructure demonstrate the successful growth of CdS nanorods on the Ag icosa-
possesses the same crystallographic symmetry as the noble metal hedral nanocrystals, and the obtained Ag-CdS icosapods show the
seed, i.e., twenty CdS (or CdSe) nanorods radially grown on twenty uniform size as well as well-defined morphology with sophisticated
exposed (111) facets of the five-fold multi-twinned Ag icosahedron, symmetrical branched architecture. Ultraviolet photoelectron spec-
respectively, denoted as Ag-CdS (or Ag-CdSe) icosapods. Importantly, troscopy (UPS) was used to reveal the interfacial electronic structure of
the epitaxial growth of CdS (or CdSe) nanorods on Ag icosahedral Ag-CdS icosapods. As shown in Supplementary Fig. 7a, the cut-off
nanocrystals has been achieved albeit a large lattice mismatch between energies (Ecut-off) of Ag nanocrystals and Ag-CdS icosapods are deter-
Ag and CdS (or CdSe) (more than 40%). Impressively, the pump-probe mined to be 16.95 eV and 17.20 eV, respectively. The Fermi levels of Ag
transient absorption (TA) spectroscopy reveals the ultrafast plasmon nanocrystals and Ag-CdS icosapods are estimated to be −4.27 eV and
hot-electron transfer process in the epitaxial heterostructures with −4.02 eV relative to the vacuum level, respectively. Compared with the
high QY of the plasmon-induced hot-electron transferred from noble Ag nanocrystals, the shift (≈0.25 eV) of the Fermi level of Ag-CdS ico-
metal to semiconductor, i.e., 18.1% and 17.6% for the Ag-CdS and Ag- sapods demonstrates a Schottky barrier was formed between Ag and
CdSe icosapods, respectively. n-type semiconductor CdS in the Ag-CdS icosapods51. Since the Fermi
level, conduction band (CB), and valence band (VB) of CdS are mea-
Results sured to be −3.92 eV, −3.61 eV, and −6.22 eV, respectively (Supple-
Material synthesis and characterizations mentary Fig. 7b, c), the interfacial electronic structure of Ag-CdS
Figure 1a schematically illustrates the synthesis of Ag-CdS icosapods. icosapods can be schematically shown in Supplementary Fig. 7d.
Briefly, five-fold multi-twinned Ag icosahedral nanocrystals with a size Importantly, the synthesis of noble metal-semiconductor hetero-
of 13.3 ± 0.6 nm were first synthesized (Supplementary Fig. 1), which structures with tunable components can be achieved. By exchanging the
were then used as seeds for the epitaxial growth of CdS nanorods via a trioctylphosphine sulfide (TOPS) with trioctylphosphine selenium
hot injection method (see the Methods section). Scanning electron (TOPSe) as the chalcogen precursor in our experiment (see the “Meth-
microscope (SEM) (Fig. 1b), transmission electron microscopy (TEM) ods” section), branched Ag-CdSe icosapods have been obtained. The
(Fig. 1c) and high-angle annular dark-field scanning transmission TEM and HAADF-STEM images of Ag-CdSe icosapods (Supplementary
electron microscopy (HAADF-STEM, Fig. 1d) images show the high Fig. 8) and the EDS elemental mapping result (Supplementary Fig. 9) of a
purity and uniformity of the synthesized Ag-CdS icosapods with a size representative Ag-CdSe icosapod demonstrate the successful growth of
of 108.2 ± 5.5 nm, and the length and diameter of CdS nanorods are CdSe nanorods on the icosahedral Ag seeds. The XPS spectra of Ag-CdSe
around 50 and 7 nm, respectively (Supplementary Fig. 2). icosapods confirm the Ag core in Ag-CdSe icosapods maintained zer-
To reveal the detailed three-dimensional (3D) architecture of the ovalent (Supplementary Fig. 10), and the XRD spectrum matches well
obtained Ag-CdS icosapods, a series of HAADF-STEM images of a with the wurtzite CdSe and fcc-Ag (Supplementary Fig. 11).
representative Ag-CdS icosapod tilted with different angles were col-
lected. The 3D reconstruction shows that twenty CdS nanorods Epitaxial noble metal-semiconductor interfaces
radially extend from twenty exposed facets of the Ag icosahedron seed The interfacial structures between the icosahedral Ag seed and the
(Supplementary Movie 1), confirming that the Ag-CdS icosapod has the grown CdS and CdSe nanorods are very important for the formation of

Nature Communications | (2023)14:2538 2

Article https://doi.org/10.1038/s41467-023-38237-7

a Ag icosahedrons Ag-CdS icosapods

C2 C2

C3 C5 C3 C5

Epitaxial growth

b c d

500 nm 200 nm 200 nm

e f g

20 nm 20 nm 20 nm

50 nm
Ag Cd S Mixed

Fig. 1 | Synthesis and characterization of Ag-CdS icosapods. a Schematic illus- b–d SEM (b), TEM (c), and HAADF-STEM (d) images of Ag-CdS icosapods.
tration of the synthesis of Ag-CdS icosapods by using the pre-synthesized Ag e–g HAADF-STEM images of a representative Ag-CdS icosapod observed along its
nanocrystals, i.e., Ag icosahedral nanocrystals, as seeds for the epitaxial growth of C2 (e), C3 (f), and C5 (g) rotational axes, respectively. h HAADF-STEM and the
CdS nanorods. C2, C3, and C5 indicate the observed rotational axes of the sche- corresponding elemental mapping images of a representative Ag-CdS icosapod.
matically drawn Ag icosahedral nanocrystals (left) and Ag-CdS icosapods (right).

such sophisticated noble metal-semiconductor heterostructures. As Together with the aligned spots in the FFT pattern (Fig. 2b), it can be
shown in Supplementary Fig. 12a, although it is very difficult to use concluded that the CdS are epitaxially grown on Ag icosahedral
spherical aberration-corrected high-angle annular dark-field STEM (Cs- nanocrystals with an epitaxial relationship of Ag (111)∥CdS (002) and
HAADF-STEM) to clearly observe the detailed interfacial structure in Ag (220)∥CdS (100). Moreover, as shown in the Cs-HAADF-STEM image
the synthesized Ag-CdS icosapod due to its sophisticated architecture (Fig. 2a) and inverse FFT image (Fig. 2c), the nearest neighbor atoms on
with large size of 108.2 ± 5.5 nm (Fig. 1b–d and Supplementary Fig. 2), the Ag surface are determined to be S of CdS (001) plane since the S
the epitaxial growth of CdS rods on Ag with the epitaxial interface atoms have lower contrast than the Cd atoms under the STEM mode,
between the Ag (111) and CdS (002) planes, referred to as Ag (111)∥CdS thus CdS is grown along the [001] direction52 (Supplementary Fig. 14).
(002), can be confirmed by the fast Fourier transformation (FFT) Similarly, as shown in Fig. 2d–f, when CdSe was grown on the icosa-
pattern (Supplementary Fig. 12b). To identify the Ag-CdS interface, Ag- hedral Ag seeds at a short reaction time of 1 min (Supplementary
CdS icosapods extracted from the reactor at ~60 s were characterized Fig. 13b), instead of 3 min used for the preparation of aforementioned
by the Cs-HAADF-STEM (Supplementary Fig. 13a). As shown in Fig. 2a, Ag-CdSe icosapods (Supplementary Figs. 8 and 9), an epitaxial rela-
the measured lattice distances of 0.14, 0.24, 0.34 and 0.36 nm are well- tionship can also be found between Ag and CdSe. The epitaxial Ag-
matched with the interplanar distances of Ag (220), Ag (111), CdS (002), CdSe interface was determined to be Ag (111)∥CdSe (002) and Ag
and CdS (100) planes, respectively. Besides the sharp Ag-CdS interface (220)∥CdSe (100), and CdSe grows along the [001] direction with the
observed, it is also found that the Ag (111) planes are parallel to the CdS nearest neighbor Se atoms on the Ag surface (Fig. 2f). Based on the
(002) planes, and Ag (220) planes are parallel to CdS (100) planes. aforementioned results, the epitaxial interfaces of Ag-CdS and Ag-

Nature Communications | (2023)14:2538 3

Article https://doi.org/10.1038/s41467-023-38237-7

a b Ag (111) CdS (002) c

Ag (220) CdS (100) CdS (100) CdS (002)
0.36 nm 0.34 nm

CdS (100) CdS (002)

0.36 nm 0.34 nm

[001]

Ag (220) Ag (111)
Ag (220) Ag (111)
0.14 nm 0.24 nm
0.14 nm 0.24 nm
1 nm 5 nm-1 1 nm

d e Ag (111) CdSe (002) f

Ag (220) CdSe (100) CdSe (100) CdSe (002)
0.37 nm 0.35 nm
CdSe (100) CdSe (002)
0.37 nm 0.35 nm

[001]
[001]

[001]
Ag (220) Ag (111)
0.14 nm 0.24 nm
Ag (220) Ag (111)
0.14 nm 0.24 nm
1 nm 5 nm-1 1 nm

Fig. 2 | Characterization of epitaxial interfaces in Ag-CdS and Ag-CdSe het- (100), respectively. d Cs-HAADF-STEM image of the Ag-CdSe interface. e, f The FFT
erostructures obtained at a short reaction time of 1 min. a Cs-HAADF-STEM pattern (e) and inverse FFT image (f) of the Ag-CdSe interface. The silver square and
image of the Ag-CdS interface. b, c The FFT pattern (b) and inverse FFT image (c) of silver triangle in (e) are applied to mark the FFT spots indexed to Ag (111) and Ag
the Ag-CdS interface. The silver square and silver triangle in (b) are applied to mark (220), respectively. The red circle and red rhombus are applied to mark the FFT
the FFT spots indexed to Ag (111) and Ag (220), respectively. The yellow circle and spots indexed to CdSe (002) and CdSe (100), respectively. The silver, blue, yellow,
yellow rhombus are applied to mark the FFT spots indexed to CdS (002) and CdS red balls in (a), (c), (d), (f) represent the scheme of Ag, Cd, S, Se atoms, respectively.

CdSe can be schematically illustrated in Supplementary Fig. 15 and band gap of the CdS nanorods, the Ag LSPR bands in Ag-CdS ico-
Supplementary Data 1. Importantly, despite the large lattice mismatch sapods are selectively excited (Supplementary Fig. 17). The TA
(more than 40%) between Ag and CdS (or CdSe) (Supplementary spectrum of Ag-CdS icosapods, displayed as a two-dimensional
Table 1), which is much greater than the previously reported value for pseudo-color plot in Fig. 3c, shows the distinct bleaching bands
the epitaxial growth (normally <5%)17,18,30,36, the epitaxial interfaces centered at ~480 and ~550 nm that can be attributed to the bleach of
can still be achieved in our synthesized Ag-CdS and Ag-CdSe the CdS 1Σ band and Ag LSPR band, respectively. As the CdS was not
heterostructures. excited, the bleach signal of the CdS 1Σ band can be attributed to the
plasmon-induced hot-electron transfer from the Ag core to CdS
Ultrafast plasmon-induced charge transfer nanorods.
The successful preparation of noble metal-semiconductor with well- Moreover, as shown in Fig. 3a, when the plasmon-induced hot-
defined morphologies and epitaxial interfaces provides a possibility electron is transferred from the Ag to CdS or CdSe, it can generate a
to explore and understand their structure-property relationships. positive mid-IR TA signal due to the 1Σ-to-1П intraband transition, which
Here, we systematically studied the plasmon-induced charge trans- enables us to directly observe the plasmon-induced hot-electron
fer process in the obtained Ag-CdS and Ag-CdSe icosapods using transfer process and quantify its QY from Ag to CdS or CdSe53–55. As
pump-probe TA spectroscopy. As schematically shown in Fig. 3a, the shown in the averaged mid-IR TA spectra of Ag-CdS icosapods (Fig. 3d),
pump laser-excited plasmon in Ag, which decays into electron-hole positive mid-IR TA signals were observed after the excitation of the Ag
pairs, is followed by hot-electron transfer to semiconductor and/or LSPR band, further confirming the presence of plasmon-induced hot-
direct interfacial charge transfer transition. The UV-visible absorp- electron in the CB of CdS. The QY of plasmon-induced hot-electron
tion spectra of Ag icosahedral seeds, CdS nanorods (Supplementary transferred from Ag to CdS can be determined by the peak amplitude of
Fig. 16), and the obtained Ag-CdS icosapods are presented in Fig. 3b. the CdS intraband absorption signal according to Eq. 4, which is cal-
The strong peak for the blue curve located at 408 nm can be culated to be 18.1% (see “Methods”, Supplementary Figs. 18, 19 and
assigned to the localized surface plasmon resonance (LSPR) Supplementary Table 2 for details). This QY value is much higher than
absorption peak of the Ag icosahedral nanocrystals. After the that previously reported in nonepitaxial noble metal-semiconductor
growth of CdS on Ag, two peaks located at 462 and 548 nm appear heterostructures53,55–58, for instance, about 2.75% for Au-tipped CdS
(red curve in Fig. 3b). The peak at 462 nm, which is similar to the nanorods53. Fitting the time-resolved mid-IR absorption curve of Ag-CdS
absorption peak of CdS nanorods (465 nm, black curve in Fig. 3b), is icosapods (Fig. 3e) based on Eq. 6 yielded a formation time of
assigned to the 1Σ exciton transition of CdS between the CB and 18.0 ± 0.3 fs and a multi-exponential decay with a half-life of
valence band VB levels53, and the peak at 548 nm should be attrib- 1.9 ± 0.16 ps, which corresponds to plasmon-induced hot-electron
uted to the redshifted LSPR of Ag nanocrystals. We then performed transfer time and charge-recombination time, respectively (Supple-
pump-probe TA spectroscopy using the 550 nm laser to excite Ag- mentary Table 3). A similar result was obtained in the epitaxial Ag-CdSe
CdS icosapods. As the photon energy of 550 nm laser is below the icosapods. The measured QY and the time of plasmon-induced hot-

Nature Communications | (2023)14:2538 4

Article https://doi.org/10.1038/s41467-023-38237-7

a CB
1Π
e- e-
LSPR 1Ʃ

h+ h+
VB
Ag CdS (or CdSe)

ΔAbs (mOD) 10 5 0 -5 -10

b 1.0 c
Ag Ag-CdS icosapods
CdS 100 Excited by 550 nm laser
0.8 Ag CdS Ag-CdS
Ag-CdS
Abs (abs .u.)

Delay (ps)
0.6

0.4

0.2 2

0 0
400 450 500 550 600 650 700 750 450 500 550 600 650 700 750
Wavelength (nm) Wavelength (nm)
d 4 e 1.2

Normalized ΔAbs
1-2 ps 1.0
20-40 ps 1.0
100-200 ps
Normalized ΔAbs

3 0.5 Exp. data

600-900 ps 0.8
ΔAbs (mOD)

Fit

2 0.6 0
-0.5 0 0.5 1 1.5 2
Time (ps)
0.4
1
Ag-CdS icosapods
0.2
Fit
CdS nanorods
0 0
3280 3000 3320 3340 3360 0 5 10 15 20
Frequency (cm-1) Time (ps)

Fig. 3 | Plasmon-induced charge separation in Ag-CdS icosapods. a Schematic in milli-optical density unit (mOD). Gray shading is used to avoid the disturbance of
illustration of plasmon-induced hot-electron transfer in Ag-CdS icosapods after the 550 nm pump laser to the probed signal. d Averaged mid-IR TA spectra of Ag-
femtosecond laser pumping. The dashed ellipsis schematically illustrates the CdS icosapods from 1 to 2 ps (red curve), 20 to 40 ps (orange curve), 100 to 200 ps
plasmon-induced interfacial charge-transfer transition in Ag-CdS icosapods. b UV- (blue curve), and 600 to 900 ps (black curve) after 550 nm laser excitation. e Time-
visible absorption spectra of the colloidal solutions of Ag icosahedral nanocrystals, resolved mid-IR absorption spectrum of Ag-CdS icosapods. Gray curves represent
CdS nanorods, and Ag-CdS icosapods in toluene. Insets: photographs of the cor- the multiexponential fits as described in Eq. 6. A negligible intraband absorption
responding colloid solutions. c Two-dimensional pseudo-color plots of pump- signal (black squares) is observed in a control experiment conducted on CdS
visible probe TA spectra of the colloid solutions of Ag-CdS in toluene. X-axis: probe nanorods. Inset in e: zoom-in kinetics spectrum for the Ag-CdS icosapods. Source
wavelength; Y-axis: pump-probe delay; Color: change in absorbance (ΔAbs), shown data are provided as a Source data file.

electron transfer were determined to be 17.6% and 7.5 ± 7.2 fs when In summary, we have demonstrated a noble-metal-seeded epi-
epitaxial Ag-CdSe icosapods were pumped by a 700 nm laser, respec- taxial growth method to achieve the controlled synthesis of highly
tively (Supplementary Figs. 19–22; Supplementary Tables 3 and 4). The symmetrical noble metal-semiconductor heterostructures, i.e., Ag-CdS
plasmon-induced hot-electron transfer processes in the epitaxial Ag- and Ag-CdSe icosapods. Abrupt epitaxial Ag-CdS/CdSe interfaces are
CdS and Ag-CdSe icosapods are much faster than the conventional observed in the Ag-CdS and Ag-CdSe icosapods, which could be con-
indirect plasmon-induced hot-electron transfer process (commonly cluded as Ag (111)∥CdS/CdSe (002) and Ag (220)∥CdS/CdSe (100),
more than 100 fs53,55). On such a short timescale (<20 fs), the plasmon albeit a large lattice mismatch (more than 40%). Detailed pump-probe
directly decays into the electrons in the CB of semiconductors via the TA-IR spectroscopy studies reveal high QYs of plasmon-induced hot-
interfacial charge-transfer transition before rapid cooling of hot elec- electron transferred from noble metal to semiconductor in the Ag-CdS
trons induced by electron-electron scattering54,55,58–60. Unfortunately, (18.1%) and Ag-CdSe (17.6%) icosapods. This work demonstrates a
because the transient IR selectively probes the electrons injected into strategy to rationally design and controllably synthesize sophisticated
the CB of CdS, similar to the previous reports54, our study was not able hierarchical noble metal-semiconductor heterostructures. The con-
to provide direct information about the effect of the metal- structed epitaxial interface provides an ideal platform for investigating
semiconductor interfacial states on the plasmon hot electron transfer the role of the metal-semiconductor interface in various physico-
and recombination processes. chemical processes.

Nature Communications | (2023)14:2538 5

Article https://doi.org/10.1038/s41467-023-38237-7

Methods heated to 370 °C under N2 atmosphere, the solution becomes trans-

Materials parent. Then, 1 mL of TOP was injected into the flask, and the tem-
Silver nitrate (AgNO3, 99.9999% trace metals basis), cadmium oxide perature was kept at 360 °C for 15 min. 0.5 mL of TOPS stock was
(CdO, ≥99.99% trace metals basis), cadmium chloride (CdCl2, 99.99% injected into the flask at 370 °C swiftly, and the temperature was kept
trace metals basis), sulfur powder (S, ≥99.98% trace metals basis), at 360 °C for 3 min. Then, the reaction was stopped by removing the
selenium powder (Se, 99.99%), oleylamine (technical grade, 70%), heating mantle. After the solution was cooled down to 100 °C, 10 mL of
trioctylphosphine (97%), trioctylphosphine oxide (technical grade, toluene was injected into the reaction flask. Then, 5 mL of methanol
90%), oleic acid (technical grade, 90%) were purchased from Sigma- was added to the solution, and the product was collected by centrifuge
Aldrich. N-octadecylphosphonic acid (95%), n-hexylphosphonic acid at 5800 × g for 3 min. The obtained precipitate was washed with the
(95%) were bought from Alfa Aesar. Hexane (≥99.0%), toluene mixture of toluene and methanol (Vtoluene/Vmethanol = 5/1) and then
(≥99.0%), and methanol (≥99.9%) were purchased from Aladdin and dispersed into 50 mL of toluene. For the synthesis of CdSe nanorods,
used as received without any further purification. all the conditions are the same as the synthesis of CdS nanorods except
0.5 mL of TOPSe stock was injected into the flask at 380 °C and the
Synthesis of five-fold multi-twinned Ag icosahedral nanocrystals temperature was kept at 360 °C for 1 min.
Typically, after 85 mg of AgNO3 were mixed with 10 mL of oleylamine
in a 50 mL three-neck flask, the mixture was degassed at 60 °C for Characterization
30 min under vigorous magnetic stirring at 750 r.p.m. It was then Powder XRD patterns of the dried samples were recorded on Bruker
quickly heated up (≥15 °C min−1) to 220 °C under N2 atmosphere. After D8 diffractometer at a scanning rate of 2° min−1 using Cu Kα radiation
the temperature was maintained at 220 °C for 1 h, the solution was (λ = 1.5406 Å). XPS and UPS spectra were recorded on the ESCALAB
cooled down to 160 °C, and a mixture of 60 μL of acetate and 0.5 mL of 250Xi (Thermo Fisher Scientific) system. The XPS results were cali-
oleylamine was injected into the flask. After the temperature was kept brated with the reference C 1s peak located at 284.6 eV. The X-ray
at 160 °C for 10 min, the reaction was stopped by removing the heating absorption fine structure (XAFS) experiments under transmission
mantle. The flask was then cooled down to room temperature natu- mode were performed at the 7-BM of the National Synchrotron Light
rally. The obtained brownish product was precipitated by centrifuge at Source II (NSLS-II) which is using a channel-cut monochromator. The
5800 × g. for 2 min. The precipitate was then washed 3 times using the NSLS-II is rung at 400 mA electron beam current under top-off mode.
mixture of hexane and acetone (Vhexane/Vacetone = 5/1). Finally, the Samples for SEM and TEM characterizations were prepared by drop-
obtained Ag nanocrystals were re-dispersed in 2.5 mL of TOP as seeds ping nanomaterial dispersions in toluene on a silicon substrate and
for the synthesis of Ag-CdS icosapods and Ag-CdSe icosapods. amorphous carbon-coated copper grid, respectively. SEM images were
obtained using FESEM, JEOL JSM-7800F. TEM characterization was
Synthesis of Ag-CdS and Ag-CdSe icosapods performed with a JEOL JEM-2100 and FEI TF20 operated at 200 kV.
TOPS and TOPSe stock solutions were first prepared by dissolving Dark-field STEM, Cs-HAADF STEM, and elemental mapping images
18 mg of sulfur powder or 10 mg of selenium powder in 0.5 mL of TOP were obtained using FEI Titan G2 60-300 operated at 300 kV as well as
in a glove box filled with argon, respectively. For the synthesis of Ag- a JEM-ARM200F operated at 200 kV. The 3D reconstruction of Ag-CdS
CdS icosapods, 115 mg of CdO, 18 mg of CdCl2, 580 mg of icosapods was carried out using FEI Titan G2 60-300 operated
n-octadecylphosphonic acid, 160 mg of n-hexylphosphonic acid, and at 60 kV.
6 g of trioctylphosphine oxide were added into a 50 mL three-neck
flask and degassed at 120 °C for 30 min under vigorous magnetic Visible femtosecond transient absorption
stirring at 750 r.p.m. After it was heated to 360 °C under N2 atmo- The visible femtosecond transient absorption measurements were
sphere, the solution becomes transparent. Then, 1 mL of TOP was conducted in a Helios spectrometer (Ultrafast Systems LLC) with
injected into the flask, and the temperature was kept at 360 °C for pump and probe beams derived from a regenerative amplified Ti:
15 min. A mixture of 0.5 mL of the above-mentioned Ag seeds in TOP Sapphire laser system (Coherent Astrella, 35 fs, 4 mJ per pulse, and
and 0.5 mL of TOPS stock was injected into the flask at 360 °C swiftly. 1 kHz repetition rate). The 800 nm output pulse was split into two
After the injection of Ag seeds and TOPS, the temperature was kept at beams with a beam splitter. One beam passed through a tunable
350 °C for 6 min. Then the reaction was stopped by removing the optical parametric amplifier (OperA solo, Coherent) to generate a
heating mantle. After the solution was cooled down to 100 °C, 10 mL of tunable visible pump. During the measurement, the pump beam was
toluene was injected into the reaction flask. Then, 5 mL of methanol chopped by a synchronized chopper to 500 Hz. The other beam was
was added to the solution, and the product was collected by centrifuge attenuated and focused on a CaF2 window to generate the white light
at 5800 × g for 3 min. The obtained precipitate was washed with the continuum with a wavelength range from 350 nm to 800 nm, referred
mixture of toluene and methanol (Vtoluene/Vmethanol = 5/1) and then to as the probe beam. The probe beam was focused into a 1-mm path-
dispersed into 50 mL of toluene. For the synthesis of Ag-CdSe icosa- length quartz cuvette (Starna) containing the sample in toluene. The
pods, all the conditions are the same with the synthesis of Ag-CdS transmission of the probe was collected by a fiber optics-coupled
icosapods except the mixture of 0.5 mL of the above-mentioned Ag multichannel spectrometer with complementary metal-oxide-
seeds in TOP and 0.5 mL of TOPSe stock was injected into the flask at semiconductor (CMOS) sensors and detected at a frequency of 1 kHz
310 °C and the temperature was kept at 300 °C for 3 min. Note that the (Ultrafast systems, Helios). The delay between the pump and probe
narrow size distribution of Ag seeds (Supplementary Fig. 1) is critical pulses was controlled by a motorized delay stage. Samples in 1-mm
for synthesizing such delicate Ag-CdS and Ag-CdSe icosapods. cuvettes were used for all spectroscopy measurements and stirred
vigorously during the measurements. The data were analyzed with
Synthesis of CdS and CdSe nanorods surface Xplorer.
TOPS and TOPSe stock solutions were first prepared by dissolving
32 mg of sulfur powder or 40 mg of selenium powder in 0.5 mL of TOP Mid-IR femtosecond transient absorption
in a glove box filled with argon, respectively. For the synthesis of CdS Mid-IR TA experiments were conducted using a commercial Ti:Sap-
nanorods, 120 mg of CdO, 580 mg of n-octadecylphosphonic acid, phire regenerative amplifier (Astrella, Coherent) at 800 nm with a
160 mg of n-hexylphosphonic acid, and 6 g of trioctylphosphine oxide repetition rate of 1 kHz and pulse duration of 35 fs. An optical amplifier
were added into a 50 mL three-neck flask and degassed at 120 °C for (OperA solo, Coherent) pumped by the regenerative amplifier was
30 min under vigorous magnetic stirring at 750 r.p.m. After it was used to provide a pump beam with tunable wavelengths. Another

Nature Communications | (2023)14:2538 6

Article https://doi.org/10.1038/s41467-023-38237-7

ΔA
optical amplifier (OperA solo, Coherent) pumped by the regenerative Here, we define ΔS = power , a quantity to be obtained from the
amplifier was used to generate a tunable IR probe via difference fre- transient IR measurement. To improve the measurement accuracy and
quency generation (DFG) by signal and idler beams. The pump beam to check the linearity of the signal, ΔA as a function of excitation power
was chopped by a synchronized chopper to 500 Hz and focused on a was measured for all samples. As shown in Supplementary Fig. 18, the
1.5-mm path-length Harrick cell containing the sample in toluene. The TA signal amplitude increases linearly with the excitation power. ΔS
IR probe, which was also focused on a 1.5-mm path length Harrick cell can be obtained from the slope of the fitted linear line. For all samples,
containing the sample in toluene, overlapped with the temporally the OD is controlled at around 0.2–0.8. The values of epitaxial Ag-CdS
delayed pump beam controlled by a movable delay stage. The probe icosapods are listed in Supplementary Table 2. It should be noted that
intensity with and without the pump was analyzed by a Teledyne/ because the nanorods are directly grown on the Ag nanoparticles,
phasetech nitrogen-cooled 128 × 128 MCT detector. All measurements there may be slight differences in the diameter of the CdS nanorods
were carried out under ambient conditions. control sample compared to the diameter of the nanorods in the
heterostructure. This could affect the QY as the optical properties of
QY measurement of the plasmon-induced hot-electron transfer the nanorods are dependent upon their size (absorption cross sec-
in Ag-CdS heterostructures tion), therefore, the QYs for each sample were calculated to the best of
The QY of the plasmon-induced hot-electrons transferred from the Ag our ability using the independently synthesized CdS nanorods
core to the CdS under 550 nm excitation can be quantified by the mid- controls.
IR intraband absorption at around 3340 cm−1 of CB electrons in CdS.
The CB electron in CdS produces a broad IR absorption which can be QY measurement of the plasmon-induced hot-electron transfer
attributed to the transition from 1Σ to a higher level (such as 1П), as in Ag-CdSe icosapods
shown in Fig. 3a. To determine the plasmon-induced hot-electron The QY of the electrons transferred from the Ag core to the CdSe
transfer QY, the amplitude of IR signal per CdS CB electron was under 700 nm excitation can be quantified by the mid-IR intraband
quantified, denoted as S0. Hence, S0 can be calculated by Eq. 1: absorption at around 3340 cm−1 of CB electrons in CdSe. Similar to Ag-
CdS icosapods, the QY is calculated by the number of hot electrons
ΔAðCdS,400 nmÞ generated in Ag-CdSe icosapods per absorbed photons under 700 nm
S0 = ð1Þ
NðCdSCBelectronsÞ excitation. CdSe nanorods are excited by a 400 nm laser. Cd3P2
quantum dots are excited by 400 nm laser and 700 nm laser for cali-
where ΔA is the maximum signal amplitude obtained from the kinetics; bration. The QY of plasmon-induced hot-electron transfer in Ag-CdSe
N(CdS CB electrons) is the number of electrons in CdS CB band, which icosapods was shown in Eq. 5.
can be determined by the number of absorbed photons by CdS (every
ΔSðAgCdSe,700 nmÞ ΔSðCd3P2,400 nmÞ
photon creates one electron in CdS CB band under 400 nm excitation). 110ODðAgCdSe,700 nmÞ 110ODðCd3P2,400 nmÞ
QY = × ð5Þ
Therefore, N(CdS CB electrons) equals the pump photon flux in the ΔSðCdSe,400 nmÞ ΔSðCd3P2,700 nmÞ
110ODðCdSe,400 nmÞ 110ODðCd3P2,700 nmÞ
pump/probe overlap region times the sample absorption


power=hv
beam size
× ð1  10OD Þ , where hv is photon energy. For all samples, the OD is controlled at around 0.4–0.7. The values
The QY is calculated by the number of hot electrons generated in used in Eq. 4 is listed in Supplementary Table 4.
Ag-CdS per absorbed photons under 550 nm excitation, as shown in
Eq. 2. The multiexponential fit of the decay kinetics of Ag-CdS and Ag-
CdSe icospads
Nðhot electronsÞ
The normalized TA IR kinetics shown in Fig. 3e and Supplementary
QY = Fig. 23b are fitted by a single exponential rise and multi-exponential
Nðabsorbed photonsÞ
ΔAðAgCdS,550 nmÞ decay function according to Eq. 6:
S0
=

powerð550 nmÞ=hvð550 nmÞ
× 1  10ODðAgCdS,550 nmÞ " !  #
beam sizeð550 nmÞ
" # t X
3
t
ΔAðAgCdS,550 nmÞ  Δ AðCdS,400 nmÞ SðtÞ = IRF  exp  + Ai exp  ð6Þ
beam sizeð550 nmÞ powerð550 nmÞ=hvð550 nmÞ powerð400 nmÞ=hvð400 nmÞ Tf Ti
= × 1
beam size ð400 nmÞ ð1  10 ODðAgCdS,550 nmÞ
Þ ð1  1 0 ODðCdS,400 nm Þ
Þ
ð2Þ The instrument response function (IRF) of the measurement was
determined in a 1 mm silicon wafer and can be well modeled by a
Gaussian function with an FWHM of ~177 ± 25 fs. Tf is the time constant
In Eq. 2, beam size (550 nm) and beam size (400 nm) represent the of the signal rise. Ai and Ti (i = 1,2,3) are amplitude and time constants
overlap areas of pump and probe beams under 550 nm, and 400 nm of the multiexponential function decay function, respectively,
pump laser excitation, respectively, whose ratio can be determined in a P3
1 Ai = 1. The fitting parameters are listed in Supplementary Table 3.
calibration sample (Cd3P2 quantum dots in this work) under the same
conditions. The calibration sample is excited at both wavelengths and Reporting summary
is expected to have the same QY for generating CB electrons. Then Further information on research design is available in the Nature
beam sizeð550 nmÞ
beam sizeð400 nmÞ
can be described as follows: Portfolio Reporting Summary linked to this article.
ΔAðCd3P2,400 nmÞ  ΔAðCd3P2,550 nmÞ
beam sizeð550 nmÞ
=
powerð400 nmÞ=hvð400 nmÞ powerð550 nmÞ=hvð550 nmÞ
ð3Þ Data availability
beam sizeð400 nmÞ ð1  10ODðCd3P2,400 nmÞ Þ ð1  10 ODðCd3P2,550 nmÞ
Þ The raw data that support the results of this study are available from
the corresponding authors on request. Source data are provided with
Thus, the QY of plasmon-induced hot-electron in Ag-CdS icosa- this paper.
pods is given by Eq. 4.
References
ΔSðAgCdS,550 nmÞ ΔSðCd3P2,400 nmÞ 1. Borys, N. J., Walter, M. J., Huang, J., Talapin, D. V. & Lupton, J. M. The
110ODðAgCdS,550 nmÞ 110ODðCd3P2,400 nmÞ
QY = ΔSðCdS,400 nmÞ
× ΔSðCd3P2,550 nmÞ
ð4Þ role of particle morphology in interfacial energy transfer in CdSe/
110ODðCdS,400 nmÞ 110ODðCd3P2,550 nmÞ CdS heterostructure nanocrystals. Science 330, 1371–1374 (2010).

Nature Communications | (2023)14:2538 7

Article https://doi.org/10.1038/s41467-023-38237-7

2. Gu, J., Zhang, Y. W. & Tao, F. Shape control of bimetallic nanoca- 26. Lu, Q. P. et al. Crystal phase-based epitaxial growth of hybrid noble
talysts through well-designed colloidal chemistry approaches. metal nanostructures on 4H/fcc Au nanowires. Nat. Chem. 10,
Chem. Soc. Rev. 41, 8050–8065 (2012). 456–461 (2018).
3. Jiang, L. Z. et al. Facet engineering accelerates spillover hydro- 27. Liu, J. W. et al. Selective epitaxial growth of Rh nanorods on 2H/fcc
genation on highly diluted metal nanocatalysts. Nat. Nanotechnol. heterophase Au nanosheets to form 1D/2D Rh-Au heterostructures
15, 848–853 (2020). for highly efficient hydrogen evolution. J. Am. Chem. Soc. 143,
4. Clavero, C. Plasmon-induced hot-electron generation at nano- 4387–4396 (2021).
particle/metal-oxide interfaces for photovoltaic and photocatalytic 28. Zhang, Z. W. et al. Robust epitaxial growth of two-dimensional
devices. Nat. Photon. 8, 95–103 (2014). heterostructures, multiheterostructures, and superlattices. Science
5. Liu, M. X. et al. Lattice anchoring stabilizes solution-processed 357, 788–792 (2017).
semiconductors. Nature 570, 96–101 (2019). 29. Shi, E. Z. et al. Two-dimensional halide perovskite lateral epitaxial
6. Yuan, W. T. et al. In situ manipulation of the active Au-TiO2 interface heterostructures. Nature 580, 614–620 (2020).
with atomic precision during CO oxidation. Science 371, 30. Oh, M. H. et al. Design and synthesis of multigrain nanocrystals via
517–521 (2021). geometric misfit strain. Nature 577, 359–363 (2020).
7. Zhao, M. T. et al. Metal-organic frameworks as selectivity regulators 31. Forticaux, A., Hacialioglu, S., DeGrave, J. P., Dziedzic, R. & Jin, S.
for hydrogenation reactions. Nature 539, 76–80 (2016). Three-dimensional mesoscale heterostructures of ZnO nanowire
8. Cheng, K. et al. Maximizing noble metal utilization in solid catalysts arrays epitaxially grown on CuGaO2 nanoplates as individual
by control of nanoparticle location. Science 377, 204–208 (2022). diodes. ACS Nano 7, 8224–8232 (2013).
9. Kagan, C. R., Lifshitz, E., Sargent, E. H. & Talapin, D. V. Building 32. Yao, Y. et al. Heteroepitaxial growth of multiblock Ln-MOF micro-
devices from colloidal quantum dots. Science 353, aac5523 (2016). rods for photonic barcodes. Angew. Chem. Int. Ed. 58,
10. Liu, Y. et al. Van der Waals heterostructures and devices. Nat. Rev. 13803–13807 (2019).
Mater. 1, 16042 (2016). 33. Zhao, M. T. et al. Selective epitaxial growth of oriented hierarchical
11. Losch, P. et al. Colloidal nanocrystals for heterogeneous catalysis. metal-organic framework heterostructures. J. Am. Chem. Soc. 142,
Nano Today 24, 15–47 (2019). 8953–8961 (2020).
12. Vogt, C. & Weckhuysen, B. M. The concept of active site in het- 34. Grundmann, M. Formation of epitaxial domains: unified theory and
erogeneous catalysis. Nat. Rev. Chem. 6, 89–111 (2022). survey of experimental results. Phys. Status Solidi B 248,
13. Lu, Q. P., Yu, Y. F., Ma, Q. L., Chen, B. & Zhang, H. 2D transition- 805–824 (2011).
metal-dichalcogenide-nanosheet-based composites for photo- 35. Zhang, J. T., Tang, Y., Lee, K. & Min, O. Y. Nonepitaxial growth of
catalytic and electrocatalytic hydrogen evolution reactions. Adv. hybrid core-shell nanostructures with large lattice mismatches.
Mater. 28, 1917–1933 (2016). Science 327, 1634–1638 (2010).
14. Linic, S., Chavez, S. & Elias, R. Flow and extraction of energy and 36. Liu, J. & Zhang, J. T. Nanointerface chemistry: lattice-mismatch-
charge carriers in hybrid plasmonic nanostructures. Nat. Mater. 20, directed synthesis and application of hybrid nanocrystals. Chem.
916–924 (2021). Rev. 120, 2123–2170 (2020).
15. Lauhon, L. J., Gudiksen, M. S., Wang, C. L. & Lieber, C. M. Epitaxial 37. Palmstrom, C. J. Epitaxy of dissimilar materials. Annu. Rev. Mater.
core-shell and core-multishell nanowire heterostructures. Nature Sci. 25, 389–415 (1995).
420, 57–61 (2002). 38. Pendharkar, M. et al. Parity-preserving and magnetic field-resilient
16. Talapin, D. V. et al. Seeded growth of highly luminescent CdSe/CdS superconductivity in InSb nanowires with Sn shells. Science 372,
nanoheterostructures with rod and tetrapod morphologies. Nano 508–511 (2021).
Lett. 7, 2951–2959 (2007). 39. Mokari, T., Rothenberg, E., Popov, I., Costi, R. & Banin, U. Selective
17. Tan, C. L., Chen, J. Z., Wu, X.-J. & Zhang, H. Epitaxial growth of growth of metal tips onto semiconductor quantum rods and tetra-
hybrid nanostructures. Nat. Rev. Mater. 3, 17089 (2018). pods. Science 304, 1787–1790 (2004).
18. Kum, H. et al. Epitaxial growth and layer-transfer techniques for 40. Wolff, C. M. et al. All-in-one visible-light-driven water splitting by
heterogeneous integration of materials for electronic and photonic combining nanoparticulate and molecular co-catalysts on CdS
devices. Nat. Electro 2, 439–450 (2019). nanorods. Nat. Energy 3, 862–869 (2018).
19. Kelso, M. V., Mahenderkar, N. K., Chen, Q. Z., Tubbesing, J. Z. & 41. Fan, Z. X. et al. Synthesis of 4H/fcc-Au@metal sulfide core shell
Switzer, J. A. Spin coating epitaxial films. Science 364, nanoribbons. J. Am. Chem. Soc. 137, 10910–10913 (2015).
166–169 (2019). 42. Yang, J. & Ying, J. Y. A general phase-transfer protocol for metal ions
20. Habas, S. E., Lee, H., Radmilovic, V., Somorjai, G. A. & Yang, P. and its application in nanocrystal synthesis. Nat. Mater. 8,
Shaping binary metal nanocrystals through epitaxial seeded 683–689 (2009).
growth. Nat. Mater. 6, 692–697 (2007). 43. Ben-Shahar, Y. et al. Optimal metal domain size for photocatalysis
21. Miszta, K. et al. Hierarchical self-assembly of suspended branched with hybrid semiconductor-metal nanorods. Nat. Commun. 7,
colloidal nanocrystals into superlattice structures. Nat. Mater. 10, 10413 (2016).
872–876 (2011). 44. Chen, J. Z. et al. Ag@MoS2 core-shell heterostructure as SERS
22. Wu, X.-J. et al. Controlled growth of high-density CdS and CdSe platform to reveal the hydrogen evolution active sites of single-
nanorod arrays on selective facets of two-dimensional semi- layer MoS2. J. Am. Chem. Soc. 142, 7161–7167 (2020).
conductor nanoplates. Nat. Chem. 8, 470–475 (2016). 45. Han, S. K., Gu, C., Gong, M. & Yu, S. H. A Trialkylphosphine-driven
23. Li, M. Y. et al. Epitaxial growth of a monolayer WSe2-MoS2 lateral p-n chemical transformation route to Ag- and Bi-based chalcogenides.
junction with an atomically sharp interface. Science 349, J. Am. Chem. Soc. 137, 5390–5396 (2015).
524–528 (2015). 46. Atta, S., Pennington, A. M., Celik, F. E. & Fabris, L. TiO2 on gold
24. Kanne, T. et al. Epitaxial Pb on InAs nanowires for quantum devices. nanostars enhances photocatalytic water reduction in the near-
Nat. Nanotechnol. 16, 776–781 (2021). infrared regime. Chem 4, 2140–2153 (2018).
25. Chen, Y. et al. Phase engineering of nanomaterials. Nat. Rev. Chem. 47. Jang, Y. H. et al. Plasmonic solar cells: from rational design to
4, 243–256 (2020). mechanism overview. Chem. Rev. 116, 14982–15034 (2016).

Nature Communications | (2023)14:2538 8

Article https://doi.org/10.1038/s41467-023-38237-7

48. Knight, M. W., Sobhani, H., Nordlander, P. & Halas, N. J. Photo- Synchrotron Light Source II, a U.S. Department of Energy (DOE) Office of
detection with active optical antennas. Science 332, Science User Facility operated for the DOE Office of Science by Broo-
702–704 (2011). khaven National Laboratory under Contract No. DE-SC0012704.
49. Wang, H. L., Zhou, S., Gilroy, K. D., Cai, Z. S. & Xia, Y. N. Icosahedral
nanocrystals of noble metals: synthesis and applications. Nano Author contributions
Today 15, 121–144 (2017). H.Z. and X.-J.W. proposed the research direction. H.Z., X.-J.W., and T.L.
50. Sines, I. T. & Schaak, R. E. Phase-selective chemical extraction of supervised the project. L.Z., D.X., and J.C. carried out the synthesis of the
selenium and sulfur from nanoscale metal chalcogenides: a general materials. L.Z., Z.L., and D.X. performed HRTEM measurements. B.C.,
strategy for synthesis, purification, and phase targeting. J. Am. L.Z., Y.Y., C.C., Q.Z., Y.Z., P.W., and L.G. collected dark-field STEM and
Chem. Soc. 133, 1294–1297 (2011). EDS elemental mapping data. L.Z., S.T.G., Y.L., X.H., and C.Z. performed
51. Wager, J. F. & Kuhn, K. Device physics modeling of surfaces and steady-state and TA measurements. C.L. built the 3D model of the
interfaces from an induced gap state perspective. Crit. Rev. Solid interface. H.Y., W.Z., X.L., C.-L.T., Z.G., and C.-S.L. contributed to XPS,
State Mater. Sci. 42, 373–415 (2017). SEM, and XRD characterization. L.M. and Y.D. analyzed synchrotron data.
52. Bladt, E. et al. Atomic structure of wurtzite CdSe (core)/CdS (giant Y.G., L.L., Z.H., H.W., Y.D., and J.L. discussed the experimental results.
shell) nanobullets related to epitaxy and growth. J. Am. Chem. Soc. L.Z., S.T.G., Y.L., T.L., X.-J.W., and H.Z. wrote the manuscript. All authors
138, 14288–14293 (2016). checked the manuscript and agreed with its content.
53. Wu, K. F., Rodriguez-Cordoba, W. E., Yang, Y. & Lian, T. Q. Plasmon-
induced hot electron transfer from the Au tip to CdS rod in CdS-Au Competing interests
nanoheterostructures. Nano Lett. 13, 5255–5263 (2013). The authors declare no competing interests.
54. Wu, K., Chen, J., McBride, J. R. & Lian, T. Efficient hot-electron
transfer by a plasmon-induced interfacial charge-transfer transi- Additional information
tion. Science 349, 632–635 (2015). Supplementary information The online version contains
55. Ratchford, D. C. Plasmon-induced charge transfer: challenges and supplementary material available at
outlook. ACS Nano 13, 13610–13614 (2019). https://doi.org/10.1038/s41467-023-38237-7.
56. Tagliabue, G. et al. Ultrafast hot-hole injection modifies hot-electron
dynamics in Au/p-GaN heterostructures. Nat. Mater. 19, Correspondence and requests for materials should be addressed to
1312–1318 (2020). Tianquan Lian, Hua Zhang or Xue-Jun Wu.
57. Cresi, J. S. P. et al. Ultrafast dynamics of plasmon-mediated charge
transfer in Ag@CeO2 studied by free electron laser time-resolved X- Peer review information Nature Communications thanks the anon-
ray absorption spectroscopy. Nano Lett. 21, 1729–1734 (2021). ymous reviewers for their contribution to the peer review of this
58. Tan, S. et al. Plasmonic coupling at a metal/semiconductor inter- work. Peer reviewer reports are available.
face. Nat. Photon. 11, 806–812 (2017).
59. Long, R. & Prezhdo, O. V. Instantaneous generation of charge- Reprints and permissions information is available at
separated state on TiO2 surface sensitized with plasmonic nano- http://www.nature.com/reprints
particles. J. Am. Chem. Soc. 136, 4343–4354 (2014).
60. Tan, S. et al. Ultrafast plasmon-enhanced hot electron generation at Publisher’s note Springer Nature remains neutral with regard to jur-
Ag nanocluster/Graphite heterojunctions. J. Am. Chem. Soc. 139, isdictional claims in published maps and institutional affiliations.
6160–6168 (2017).
Open Access This article is licensed under a Creative Commons
Acknowledgements Attribution 4.0 International License, which permits use, sharing,
This work was supported by the National Natural Science Foundation of adaptation, distribution and reproduction in any medium or format, as
China (21871129 and 22271142), the Fundamental Research Funds for the long as you give appropriate credit to the original author(s) and the
Central Universities (020514380204), “Innovation & Entrepreneurship source, provide a link to the Creative Commons license, and indicate if
Talents Plan” of Jiangsu Province, and start-up funds from Nanjing Uni- changes were made. The images or other third party material in this
versity. H.Z. thanks the support from ITC via Hong Kong Branch of article are included in the article’s Creative Commons license, unless
National Precious Metals Material Engineering Research Center (NPMM), indicated otherwise in a credit line to the material. If material is not
and the Start-Up Grant (Project No. 9380100), and the grants (Project included in the article’s Creative Commons license and your intended
Nos. 9678272 and 1886921) from the City University of Hong Kong. T.L. use is not permitted by statutory regulation or exceeds the permitted
acknowledges the support from the U.S. Department of Energy, Office use, you will need to obtain permission directly from the copyright
of Science, Office of Basic Energy Sciences, Solar Photochemistry Pro- holder. To view a copy of this license, visit http://creativecommons.org/
gram under Award Number (DE-SC0008798). S.T.G. acknowledges licenses/by/4.0/.
support from an AGEP supplement to NSF award number CHE-
2004080. L.G. thanks the Beijing Natural Science Foundation (Z190010), © The Author(s) 2023
National Natural Science Foundation of China (51991344, 52072400,
52025025). This research used 7-BM and 8-BM of the National

1
State Key Laboratory of Coordination Chemistry, School of Chemistry and Chemical Engineering, Nanjing University, Nanjing 210023, China. 2Department of
Chemistry, City University of Hong Kong, Hong Kong, China. 3Hong Kong Branch of National Precious Metals Material Engineering Research Center (NPMM),
City University of Hong Kong, Hong Kong, China. 4Department of Chemistry, Emory University, Atlanta, GA 30322, USA. 5College of Materials Science and
Engineering, Sichuan University, Chengdu, Sichuan 610065, China. 6Department of Applied Physics, The Hong Kong Polytechnic University, Hung Hom, Hong
Kong. 7National Synchrotron Light Source II, Brookhaven National Laboratory, Upton, NY 11973, USA. 8Beijing National Laboratory for Condensed Matter
Physics, Institute of Physics, Chinese Academy of Sciences, Beijing 100190, China. 9National Laboratory of Solid State Microstructures, Jiangsu Key
Laboratory of Artificial Functional Materials, College of Engineering and Applied Sciences and Collaborative Innovation Center of Advanced Microstructures,

Nature Communications | (2023)14:2538 9

Article https://doi.org/10.1038/s41467-023-38237-7

Nanjing University, Nanjing 210093, China. 10National Laboratory of Solid State Microstructures, School of Physics, and Collaborative Innovation Center of
Advanced Microstructures, Nanjing University, Nanjing 210093, China. 11Department of Physics, University of Warwick, Coventry CV4 7AL, UK. 12Beijing
National Center for Electron Microscopy and Laboratory of Advanced Materials, Department of Materials Science and Engineering, Tsinghua University,
Beijing 100084, China. 13Shenzhen Research Institute, City University of Hong Kong, Shenzhen 518057, China. 14These authors contributed equally: Li Zhai,
Sara T. Gebre, Bo Chen, Dan Xu. e-mail: tlian@emory.edu; hua.zhang@cityu.edu.hk; xjwu@nju.edu.cn

Nature Communications | (2023)14:2538 10