Article https://doi.org/10.

1038/s41467-025-57304-9

Highly efficient and eco-friendly green

quantum dot light-emitting diodes through
interfacial potential grading
Received: 31 October 2024 Yuhe Bi 1, Jianhui Sun2, Sheng Cao 1
, Qiuyan Li1, Jinju Zheng3, Xi Yuan4,
Yunjun Wang 5 , Bingsuo Zou 1,6 & Jialong Zhao 1
Accepted: 18 February 2025

As next-generation display technologies, eco-friendly colloidal quantum dot

Check for updates light-emitting diodes have drawn great attention due to their excellent lumi-
1234567890():,;
1234567890():,;

nescence properties, along with their rapid development. However, practical

applications of eco-friendly quantum dot light-emitting device remain chal-
lenging, primarily due to the inferior performance of green device, which still
lag behind their red and blue counterparts. Herein, we present efficient green
device based on interfacial potential-graded ZnSeTe quantum dots. Our find-
ings show that this potential-graded structure alleviates interfacial lattice
mismatch and strain, reducing structural deformation and misfit defects. The
smoothed interfacial potential suppresses the nonradiative recombination
processes, particularly Auger recombination revealed by excitation-intensity
dependent ultrafast transient absorption kinetics. Consequently, the inter-
facial potential-graded quantum dots demonstrate highly efficient green
quantum dot light-emitting diodes, with a peak external quantum efficiency of
21.7% at 520 nm and a corresponding current efficiency of 75.7 cd A−1.

Quantum dot light-emitting diodes (QLEDs) are commonly seen as 442 h at 650 cd m−2 for T50 (the time when the luminance decreases to
crucial contenders for next-generation display and lighting 50% of its original value)17,19. Notably, recent advancements in green
technologies1–6, owing to their exceptional luminescence properties InP-based QLEDs have demonstrated a high EQE of 26.68% and a T95
and low production costs. Over the past decade, significant progress lifetime of 1241 h (at 1000 cd m−2)18. These developments, particularly
has been made in cadmium (Cd) selenide and lead (Pb) halide those led by Shen et al., represent a significant step forward. However,
perovskite-based QLEDs7–16. Recently, heavy-metal-free indium phos- despite these advancements, most reported green InP and ZnSeTe
phide (InP) and zinc selenide-telluride (ZnSeTe) quantum dot (QD)- QLEDs continue to lag behind their red and blue counterparts20,23–28.
based QLEDs have emerged as a more sustainable alternative for dis- For instance, the maximum EQE of green ZnSeTe-based QLEDs at just
play applications and have demonstrated promising performance17–22. 10.1%, with a T50 lifetime of only 52 h at 1000 cd m−2. This disparity in
For instance, red InP-based QLEDs and blue ZnSeTe-based QLEDs have performance has impeded the balanced development and practical
reached external quantum efficiency (EQE) of 21.4% and 20.2%, utilization of heavy-metal-free red, green, and blue QLEDs. Since the
respectively, having long lifetimes—615 h at 1000 cd m−2 for T95 (the human eye is highly sensitive to the green spectral region29, resolving
time when the luminance decreases to 95% of its original value) and this performance gap to develop efficient and bright green QLEDs is

1
School of Physical Science and Technology, State Key Laboratory of Featured Metal Materials and Life-cycle Safety for Composite Structures, Guangxi
University, Nanning, China. 2College of Physics Science and Technology, Heilongjiang University, Harbin, China. 3Institute of Micro/Nano Materials and
Devices, Ningbo University of Technology, Ningbo, China. 4College of Information Technology, Jilin Engineering Research Center of Optoelectronic Materials
and Devices, Jilin Normal University, Siping, China. 5Suzhou Xingshuo Nanotech Co. Ltd. (Mesolight), Suzhou, China. 6School of Resources, Environment and
Materials, Guangxi Key Laboratory of Processing for Non-Ferrous Metals and Featured Materials, Guangxi University, Nanning, China.
e-mail: [email protected]; [email protected]; [email protected]; [email protected]

Nature Communications | (2025)16:1945 1

Article https://doi.org/10.1038/s41467-025-57304-9

essential for the technological advancement and to fully realize its QDs. Next, we grow the ZnSe0.9Te0.1 IPG shell at 280 °C (Stage 2, core/
potential in display and lighting applications. IPG). Finally, the QDs are sequentially coated with a ZnSe shell (Stage 3)
Research on heavy-metal-free green QLEDs currently focuses on and a ZnS shell (Stage 4), as shown in Fig. 1b. For the control QDs, we
QDs based on either InP or ZnSeTe18,20,23,27,28. InP QDs encounter sub- directly coat the ZnSe0.8Te0.2 QD cores with a ZnSe shell of the same
stantial hurdles because of their narrow bulk bandgap (1.35 eV) and thickness instead of IPG shell, while keeping the other shells the same,
large excitonic Bohr radius (10.0 nm)30. Smaller cores are required for to ensure that size-related factors are consistent between samples
green emission, and the synthesis of high-quality green QDs with core/ (detailed synthesis methodology is shown in the experimental
shell structures is thus complicated31,32. Additionally, InP has been section).
classified as a Group 2A carcinogen by the World Health X-ray diffraction (XRD) patterns (Fig. 1c, d) show that all diffrac-
Organization33, raising concerns about its safety and suitability as an tion peaks match the standard blende (ZB) ZnSe pattern (JCPDS No. 37-
eco-friendly emitter in QLEDs. In contrast, ZnSeTe QDs show potential 1463), confirming that both QD types maintained a ZB crystal structure
as eco-friendly green emitters due to their band-gap-bowing without the emergence of the wurtzite (WZ) structure. This is likely
properties27,28,34, enabling precise tuning of emission wavelengths due to the effective elimination of stacking faults in the ZnSe shell
through alloy composition and quasi-type II heterostructure. However, using HF during synthesis19. As the ZnSe or ZnS shell layers are applied,
achieving green emission in ZnSeTe QDs requires a significant amount both the IPG QDs and the control QDs exhibited improved crystal-
of Te, which results in a larger lattice mismatch at the core/shell lization, as indicated by the sharpening of diffraction peaks, a slight red
interface compared to blue ZnSeTe QDs. This leads to a substantial shift, and the narrowing of the peak line widths.
interfacial potential difference between core and shell28,35, increasing To verify the IPG structure of the green ZnSeTe alloy QDs, high-
Auger recombination rates and drastically reducing device efficiency, angle annular dark-field (HAADF) scanning transmission electron
particularly at high brightness levels. Currently, there is no effective microscopy (STEM) and energy dispersive spectroscopy (EDS) ele-
solution to mitigate Auger recombination in green ZnSeTe QLEDs26–28. mental mapping are used to examine the morphology and composi-
While progress has been made in suppressing Auger recombination in tion distribution of QDs (Fig. 1e, f). The IPG QDs exhibit a well-defined
Cd-based and InP-based QDs by smoothing interfacial potential31,32,36,37, cubic shape, while the control QDs display an irregular shape
the relationship between interfacial potential and the performance of (Fig. 1g, h), likely due to greater structure deformation caused by
ZnSeTe QDs in QLEDs remains unclear. These unresolved challenges increased lattice strain between the core and ZnSe shell38. EDS analysis
keep continue to hinder the development of high-performance green indicate that the Te distribution in the control QDs is more inhomo-
ZnSeTe QLEDs. geneous compared to the IPG QDs (full EDS data available in Supple-
In this work, we report highly efficient, eco-friendly green QLEDs mentary Figs. 3 and 4). In traditional ZnSeTe/ZnSe core/shell structure,
enabled by tailoring the interfacial potential structure of ZnSeTe QDs. the heavy Te alloying in the ZnSeTe core increases lattice strain and
The interfacial potential is controlled by the growth of graded Te interlayer mismatch34,35. The inhomogeneous Te distribution in the
distribution core/shell structure, resulting in an interfacial potential- control QDs can be attributed to the increased interlayer mismatch,
graded (IPG) structure, referred to as IPG QDs. We demonstrate that which enhances compressive stress on the core as the shell thickness
the IPG QDs possess a compositional gradient alloyed heterostructure, increases, leading to a partially aggregated distribution of Te within
which significantly enhances their luminescent properties. Lattice the core. Furthermore, element line scanning of the Se and Te atomic
vibration analysis of the longitudinal optical (LO) phonon mode con- fractions directly reveal the interface component changes and core/
firms that compressive strain in the QD core is reduced with the IPG shell structure. As shown in Fig. 1i, j, the Te distribution in the control
configuration. Furthermore, ultrafast exciton dynamic analysis reveals QDs exhibit a steep transition, whereas in the IPG QDs, the Te dis-
a notable reduction in nonradiative Auger recombination in IPG QDs tribution shows a gradual gradient along the radius. These results
compared to control QDs. As a result, the IPG QDs achieve a photo- confirm that the IPG shell is epitaxially grown on the ZnSeTe core,
luminescence (PL) quantum yield (QY) of up to 95%, with a narrow resulting in the formation of a graded interfacial potential in the target
linewidth of 38 nm. Green ZnSeTe QLEDs based on IPG QDs demon- IPG QDs.
strate a peak EQE of 21.7%, a peak current efficiency of 75.7 cd A−1, and a
T50 lifetime of 99.4 h at 1000 cd m−2. This work successfully addresses Optical property and growth process characterization
the performance limitations in green ZnSeTe QLEDs and provides To investigate the effects of IPG structure on luminescence perfor-
deep insights into the relationship between interfacial confinement mance, we record the absorption and PL spectra at each
potential and QLEDs performance, advancing the development of synthesis stage (Fig. 2a, b), along with the PL QY of QDs (Fig. 2c, d).
high-performance, eco-friendly green QLEDs. Due to the quasi-type II heterostructure of ZnSeTe/ZnSe, electro-
ns can delocalize across the shell22,34. As a result, the PL peak shifts
Results from 463 nm (for the cores) to 515 nm (for the control QDs) and
Design and structure characterization 519 nm (for the IPG QDs) with the coating of ZnSe/ZnS shells
To achieve green light emission from ZnSeTe-based QDs, we tailor the (Fig. 2a, b). The 4 nm red shift observed in IPG QDs compared to the
Te content in the ZnSeTe QD cores to a Te/Se ratio of 1:4, leveraging control QDs is attributed to carrier wavefunction delocalization,
the bandgap bowing properties of ternary alloys. Previous studies have which is induced by the IPG structure and the effect of Te in the
shown that increasing the Te content enhances the lattice mismatch shell. In the control QDs, the hole wavefunction is strongly
between the ZnSeTe core and the ZnSe shell, which amplifies the confined within the core. However, the introduction of the IPG leads
potential distribution at the core-shell interface34,35. This leads to a to a stronger leakage of the carrier wave function outside the core.
steep interfacial potential that induces Auger recombination, reducing In the IPG QDs, the hole wavefunction becomes more
luminescence efficiency31. To address this issue, we introduce an IPG delocalized due to the gradient potential, as shown in the inset of
shell between the ZnSeTe core and the ZnSe shell (as demonstrated in Fig. 2e. This delocalization results in a lower confinement of the hole,
Supplementary Figs. 1 and 2). The designed QD structure and the causing the red shift in the PL peak31. Additionally, the presence of Te
corresponding interfacial potential are depicted in Fig. 1a, while the in the shell further affects the band structure, contributing to the
synthesis schematic of the IPG QDs is shown in Fig. 1b. redshift.
We first synthesize the green ZnSe0.8Te0.2 QD cores at 240 °C The QDs grow in size from stage 1 (cores, 3.1 nm) to stage 4
(referred to Stage 1), and then passivate the surface defects with HF/ (8.7 nm for control QDs, 8.8 nm for IPG QDs). Transmission electron
ZnCl2, using a method successfully demonstrated in InP and ZnSeTe microscopy (TEM) images of QDs at each stage (size distribution of

Nature Communications | (2025)16:1945 2

Article https://doi.org/10.1038/s41467-025-57304-9

Fig. 1 | Schematic illustration and structure characterization. a Schematic scale bar is 5 nm. g, h HAADF STEM image of signal control and IPG QD. The scale
illustration of the IPG QDs and interfacial potential structure. b Synthesis schematic bar is 2 nm. i, j Smoothed EDS line scanning profiles of the corresponding signal
of core/IPG/ZnSe/ZnS QDs. c, d XRD patterns of control and IPG QDs at each stage. control and IPG QD along dotted arrow.
e, f HAADF STEM image and EDS elemental mapping of control and IPG QDs. The

QDs at all stages is shown in Supplementary Figs. 5 and 6) are pre- the PL decreases from 53 nm (cores) to 38 nm (IPG QDs) and 47 nm
sented alongside the corresponding PL spectra. At stage 2, both the (control QDs). It has been reported that the large lattice mismatch
IPG and ZnSe shells are 1.0 nm thick for IPG QDs and control QDs. For between the core and shell generates increased lattice strain, applying
stages 3 and 4, the thickness of ZnSe and ZnS shells is measured to be compressive stress to the core38,39. This stress induces misfit defects
1.0 nm and 0.8 nm, respectively, for both types of QDs. TEM images and structure deformation at the interface, ultimately degrading
reveal that the core/IPG QDs in stage 2 exhibit a uniformly distributed luminescence performance. Therefore, the significant improvement in
spherical shape, while the core/ZnSe QDs become increasingly irre- PL QY and the reduction in FWHM in IPG QDs can be attributed to the
gular in shape, especially after the ZnSe (stage 3) and ZnS (stage 4) reduced interfacial strain and the suppression of interfacial defects.
shell coatings (Fig. 2a, b). This intensified structure deformation is To further validate the role of the IPG structure in enhancing
attributed to the large interfacial potential and lattice strain between luminescence, we perform time-resolved PL spectroscopy (Fig. 2e, f)
the ZnSe shell and ZnSeTe core, where heavy-Te alloying significantly on the QDs. The PL dynamics are analyzed using a triexponential
narrows the optical band gap (Eg)28,35. As the shells are applied, the PL function, with the fitting results provided in Supplementary Table 1.
QY of IPG QDs increases dramatically from less than 6% at stage 1 to We observe that the average lifetime (τav) increases from 21.6 ns in
95% at stage 4, while the PL QY of control QDs only reaches 70% stage 1 (cores) to 58.3 ns in stage 4 (control QDs) and 70.3 ns in stage 4
(Fig. 2c, d). Concurrently, the full width at half-maximum (FWHM) of (IPG QDs). This increase is attributed to the enhanced exciton

Nature Communications | (2025)16:1945 3

Article https://doi.org/10.1038/s41467-025-57304-9

Fig. 2 | Luminescence and structure characteristics for control and IPG QDs. electron wavefunction and localized hole wavefunction for IPG QDs (red solid line)
a, b PL and absorption spectra of control and IPG QDs at each stage. The corre- and control QDs (gray dotted line)), and IPG QDs (inset: schematic energy level
sponding TEM images are shown on the right side. The scale bar is 10 nm. c,d The diagram of radiative and non-radiative transition pathways, where τ1, τ2, and τ3
FWHM and PL QY of control and IPG QDs at each stage. e, f Time-resolved PL represents the lifetime of defect-assisted recombination, band-edge recombination
spectra of control QDs (inset: schematic energy level diagram of delocalized and localized state recombination of IPG QDs, respectively) at each stage.

delocalization and the suppression of nonradiative recombination due transversal optical (TO) phonons (~200 cm−1), first-order longitudinal
to the successive growth of multiple shell layers. optical (1LO) phonons (~240 cm−1), and surface optical (SO) phonons
The PL dynamic is strongly influenced by the Te content, which (~220 cm−1)49. The 1LO peaks show a low-frequency shift (~10 cm−1)
can lead to localized hole states at the top of valence band and affect compared to binary ZnSe (1LO, ~252 cm−1), which is attributed to the
the recombination processes. The fast decay components are likely effect of ZnTe alloying (1LO, ~210 cm−1). After shell coating, the 1LO
associated with nonradiative defect-assisted recombination40–44, while peak shifts to a higher frequency in both IPG QDs (from 237.6 cm−1 to
the slow decay components are related to the recombination of loca- 240.2 cm−1) and control QDs (from 237.6 cm−1 to 244.1 cm−1). The rela-
lized hole states45–48 (as shown in inset of Fig. 2f). Specifically, the tive intensity of 2LO to 1LO (I2LO/I1LO) ratio increases from 0.52 for the
amplitude of the fast decay components in the cores is 0.43, core to 0.74 for the IPG QDs and 0.65 for the control QDs as the QDs
decreasing to 0.26 for control QDs and 0.17 for IPG QDs as additional grow in size (Supplementary Fig. 7)50. The trend is attributed to the
shell layers are coated. The reduction in the fast decay component confinement of electron and hole wavefunctions. The higher I2LO/I1LO
amplitude represents a decrease in nonradiative recombination, which ratio observed in IPG QDs can be explained by the stronger leakage of
is inversely proportional to the PL QY22,40,43. The coating of ZnSe/ZnS carrier wavefunction due to the gradient interfacial potential, as dis-
shell effectively passivates interfacial defects and reduces nonradiative cussed in the previous study51. Additionally, the linewidth (Γ) of the 1LO
recombination pathways, which directly contributes to the dramatic peak narrows from 31.0 cm−1 to 20.2 cm−1 in control QDs and to
improvement in PLQY from less than 6% at Stage 1 to 95% at Stage 4. 15.3 cm−1 in IPG QDs. The more pronounced 1LO peak shift
The suppression of the fast decay component in IPG QDs indicates that (Δω = 6.9 cm−1) and less linewidth narrowing (ΔΓ = 10.8 cm−1) observed
defect-assisted nonradiative recombination is significantly minimized, in control QDs suggest that they experience more significant com-
likely due to minimized interfacial strain in the IPG structure. pressive strain compared to IPG QDs.
To quantify the strain on the ZnSeTe cores, we analyze the relative
Lattice strain and ultrafast carrier dynamics 1LO phonon shifts (Δωω ) of core/IPG QDs and core/ZnSe QDs to estimate
To further explore the effects of interfacial potential on lattice strain of relative changes in lattice constant (Δa
a ) using the Grüneisen parameter
ZnSeTe alloy QDs, we conduct Raman spectroscopy measurements38. (γ)38,39:
Figure 3a, b shows the Raman spectra of both IPG and control QDs,
revealing asymmetric fundamental Raman peak around 240 cm−1.

These spectra are deconvoluted into three Lorentzian peaks (fitting Δω Δa
γ
= 1+3
1 ð1Þ
results are shown in Supplementary Table 2), corresponding to ω a

Nature Communications | (2025)16:1945 4

Article https://doi.org/10.1038/s41467-025-57304-9

Fig. 3 | Lattice strain variations and recombination dynamics. a, b Raman right side. The scale bar is 2 nm. d, e TA spectra at selected timescales for IPG and
spectra of IPG and control QDs at different stages. The purple dotted line refers to control QDs. f, g PB1 bleach dynamics of IPG and control QDs at different excitation
TO mode, the black dotted line refers to SO mode and the red/blue dotted line densities. h The bleach amplitude of PB1 as a function of excitation density.
refers to the 1LO mode of IPG/control QDs. c, Raman shift and 1LO linewidth versus i, Average lifetime of PB1 dynamics as a function of excitation density for IPG and
different stages of QDs, relative change in lattice constants versus different stages control QDs. The schematic diagram depicts the Auger recombination under high
of core, core/IPG, core/ZnSe QDs. The inverse Fourier transform (FFT) images of excitation density.
core/IPG QDs (top) and core/ZnSe QDs (bottom) along [111] are also shown on the

The γ = 0.85 is taken in these calculations for both IPG and 3.02 eV) and 471 nm (PB1, 2.63 eV). PB2 is attributed to carrier filling in
control QDs49. This analysis reveals that both IPG and ZnSe shells the ZnSe shell transition, while PB1 corresponds to state-filling of band-
introduced compressive strain on the ZnSeTe core, with the ZnSe edge in core34. At a high excitation density of 4.5 mJ cm–2, the 1S exciton
shell inducing a larger compressive strain (Δaa = 0.65%) compared to bleach (PB1) in control QDs recovers by about 90% within 0.5 ns,
IPG shell (Δaa = 0.16%), as shown in Fig. 3c. The lattice constant, whereas in IPG QDs it recovers by around 60%, indicating that Auger
derived from the mean interplanar spacing, is 5.71 Å for core/IPG QDs recombination occurs more rapidly in control QDs.
(d = 3.29 Å) and 5.84 Å for core/ZnSe QDs (d = 3.37 Å) along the [111] Further investigation using excitation-intensity-dependent TA
crystallographic direction (Fig. 3c and Supplementary Fig. 8). Addi- measurements (varying the excitation density from 0.1 to 4.5 mJ cm–2,
tionally, the lattice mismatch between the ZnSeTe core and the IPG Supplementary Fig. 9) focuses on PB1 for detailed kinetic analysis
shell is calculated to be 1.5%, while the mismatch with the ZnSe shell (Fig. 3f, g). All dynamics are well-fitted using a triexponential function,
is significantly higher at 3.9%. These findings suggest that the IPG where τ1, τ2, and τ3 represent multiexciton, biexciton, and single-
structure effectively alleviates interfacial potential, reduces com- exciton lifetimes, respectively. The fast component (τ1) is primarily
pressive strain, and minimizes misfit defects, thereby enhancing the associated with Auger recombination40,43,45,47,48,52. For IPG QDs, the
luminescent performance of the QDs, which aligns well with the PL average lifetime decreases from 495.5 ps (at 0.1 mJ cm–2) to 282.7 ps (at
characterization results. 4.5 mJ cm–2), while control QDs exhibit a faster decrease, from 405.2 ps
To investigate excited-state dynamics and Auger recombination to 159.1 ps (Supplementary Tables 3 and 4). Both QD types exhibit
in IPG QDs, we perform ultrafast transient absorption (TA) spectro- similar τ1 amplitudes up to an excitation intensity of 1.6 mJ cm–2,
scopy. The TA spectra for both IPG and control QDs (Fig. 3d, e) reveal beyond which control QDs exhibit larger τ1 amplitudes, indicating
two distinct photoinduced bleaching (PB) features at ~410 nm (PB2, faster Auger recombination. We fit the 1S absorption changes at the

Nature Communications | (2025)16:1945 5

Article https://doi.org/10.1038/s41467-025-57304-9

different excitation densities using the formula53,54: polystyrene sulfonate (PEDOT:PSS) (45 nm)/poly((9,9-dioctyl-
fluorenyl-2,7-diyl)-alt-(9-(2-ethylhexyl)-carbazole-3,6-diyl)) (PF8Cz)
   
n1s = 1
e
jp σ a 1 + j p σ a =2 ð2Þ (40 nm)/ZnSeTe-based core/shell QDs (25 nm)/ZnMgO nanocrystal
(60 nm)/Al (100 nm), as depicted in Fig. 4a. The energy levels of QDs
 
where n1s is the average occupation number of the 1S electron state, are determined using ultraviolet photoemission spectroscopy (UPS)
the jp is excitation density and σa is the absorption cross-section of a and absorption spectra (Supplementary Fig. 10), while the energy band
QD. As shown in Fig. 3h, the change in absorption (ΔA) initially diagram of QLEDs is provided in Supplementary Fig. 11. The large
increases linearly before saturating at densities above 1.6 mJ cm−2. This potential barrier and high electron mobility of ZnMgO lead to electron
behavior aligns with previous studies of ZnSeTe, CuInS2, and CdSe accumulation at the interface2,10,24. The introduction of the IPG con-
core/shell QDs40,52,54 and supports our earlier findings. figuration enhances carrier injection into the IPG QDs compared to
To quantitatively analyze Auger recombination, we examine the control QDs, as shown by capacitance-voltage curves (Supplementary
relationship between excitation density and average lifetime. It is Fig. 12). As the applied voltage increases, carriers are injected and
widely acknowledged that when the densities of photoexcited elec- accumulate in the QLEDs, which is reflected by a slight initial increase
trons and holes are identical, the rate equation for the photocarriers in capacitance followed by a rapid rise around 2.2 V. The peak capa-
can be simplified to: 1=τ = κ 1 + κ 2 n + κ 3 n2 , where κ 1 , κ 2 , and κ 3 repre- citance of IPG QLEDs occurs at a lower voltage and shows a smaller
sent the nonradiative single-carrier-trapping rate, radiative bimole- value than that of control QLEDs, indicating a lower carrier injection
cular recombination coefficient, and nonradiative Auger barrier and charge accumulation.
recombination coefficient42,43, respectively. Global fitting is carried out The current density-voltage-luminance (J-V-L) characteristics
to simulate the relationship between average lifetime (τ) and excita- (Fig. 4b) demonstrate that QLEDs based on IPG QDs exhibit a low
tion density (jp). The fitting results (Fig. 3i) show that the average turn-on voltage of 2.2 V (driving voltage at a luminance of 1 cd m–2),
lifetime remains constant under low excitation densities but decreases and a peak brightness of 52,040 cd m–2 at a low driving voltage of
rapidly above 1.6 mJ cm–2, confirming the onset of Auger recombina- 5.4 V. This high brightness at such a low driving voltage far exceeds
tion at high excitation densities. The calculated Auger recombination the brightness of the most heavy-metal-free green QLEDs (such as
coefficient for IPG QDs is 9.7 × 10–5 ps–1 mJ–2 cm4, which is significantly InP-based QLEDs and carbon QLEDs)20,23–26,28. In contrast, the QLEDs
lower than the 2.2 × 10–4 ps–1 mJ–2 cm4 for control QDs (Supplementary based on control QDs show a lower peak brightness of 35,000 cd m–2.
Table 5). These results demonstrate that the smoothed interfacial As the driving voltage increased from 2.0 V to 6.0 V, the EL intensity
potential in IPG QDs effectively suppresses Auger recombination, of the IPG QLEDs rose rapidly and remained stable, without any shift
indicating a key factor in their enhanced luminescence performance. in the EL spectra at 520 nm, even under high bias voltage (Fig. 4c,
Supplementary Fig. 13a). The slight discrepancy between EL and PL
Green ZnSeTe-based QLED performance peak can be primarily attributed to two factors: Förster resonance
To verify that reducing defect-assisted nonradiative recombination energy transfer (FRET) and the electric field-induced Stark effect.
and Auger recombination in QDs through a smoothed interfacial Given that the EL peak consistently remains at 520 nm across various
potential enhances electroluminescence (EL) performance, we fabri- voltages with no observable spectral shifts, this discrepancy in PL
cate a series of QLEDs. The device structure used in this work is as peak can be attributed to FRET, as observed in the QDs film PL
follows: indium tin oxide (ITO)/poly-(ethylene dioxythiophene): (Supplementary Fig. 13b).

Fig. 4 | Device structure and performance of QLEDs. a Device architecture of the (inset: summary of the reported ZnSeTe green QLEDs performance on the basis of
green QLED. b Current density-voltage-luminance characteristics of the green maximum EQE and the corresponding luminance). e Histograms of the peak EQEs
QLEDs based on control and IPG QDs. c Two-dimensional (2D) map of drive- for control and IPG QDs, each measured with over 60 devices. f Operational sta-
voltage-dependent EL spectra of IPG QLED. The photo of operating device is shown bility of QLEDs under constant current of 1.22 mA for IPG and control QLEDs.
in inset with a pixel size of 4 mm2. d Corresponding EQE-luminance characteristics

Nature Communications | (2025)16:1945 6

Article https://doi.org/10.1038/s41467-025-57304-9

Notably, the QLEDs based on IPG QDs achieve a maximum EQE of and 140 mL of TOA. This mixture was heated to 120 °C and under-
21.7% at a high luminance of 5963 cd m–2 (Fig. 4d), with a corre- went degassing for 1 h. Subsequently, it was heated to 300 °C and
sponding current efficiency (CE) of 75.7 cd A−1 (Supplementary Fig. 14). kept boiling for 20 min. Selenium precursor: 500 mmol of selenium
In comparison, the QLEDs based on control QDs show a maximum EQE powder was blended with 250 mL of TOP and stirred continuously
of 13.0% at a luminance of 1290 cd m–2. This high EQE at such high until a clear solution was achieved. Sulfur precursor: 500 mmol of
luminance reflects the effect of suppression in Auger recombination, sulfur powder was blended with 250 mL of TOP and stirred con-
enabled by the smoothed interfacial potential in the IPG QDs. This tinuously until a clear solution was achieved. Zn(OLA)2 precursor: a
performance represents the highest EQE for green ZnSeTe-based 1000 mL flask was filled with a mixture of 290 mmol of Zn(AC)2,
QLEDs to the best of our knowledge (in the inset of Fig. 4d and Sup- 120 mL of OLA, and 500 mL of ODE. The mixture was then heated to
plementary Table 6)26–28. 120 °C and degassed for 1 h. ZnCl2-OLA precursor: ZnCl2 (25 mmol)
The reproducibility of these devices is demonstrated by the dis- and OLA (50 mL) were loaded into a 100 mL flask, heated to 120 °C,
tribution of EQE values obtained from over 60 devices, as shown in and degassed for 1 h.
Fig. 4e. The average EQE for the control QLEDs is 10.5%, whereas QLEDs
based on IPG QDs achieve an average EQE of 18.6%. This demonstrates Synthesis of IPG and control QDs
consistent performance improvement in the IPG QLEDs, highlighting For the synthesis of core, 0.256 g of PA (1 mmol), 2.5 mL of OLA, and
the effectiveness of the IPG design in enhancing efficiency across 10 mL of TOA were added into a 100 mL three-neck flask. The mixture
multiple devices. Since the Auger process significantly impacts device was heated to 120 °C and degassed for 2 h. The temperature was then
operational stability18,42, the operational stability of QLEDs is evaluated raised to 240 °C under a N2 flow. At this point, DPP (2 mmol), TOP-Se
at high luminance. The lifetimes of QLEDs are shown in Fig. 4f. The T50 (2 M, 0.48 mmol), TOP-Te (0.1 M, 0.12 mmol), and DEZ (1 M, 1.2 mmol)
lifetime is measured at a constant current of 1.22 mA. The QLEDs based were injected into the flask, and the reaction was maintained for
on IPG QDs exhibit a T50 lifetime of 18.6 h at an initial luminance of 30 min to grow ZnSeTe cores. Subsequently, 0.1 mL ZnCl2-OLA (0.5 M)
n
2538 cd m–2, corresponding to 99.4 h @ 1000 cd m–2 ( L0 × T 50 = and 1.0 mL of diluted HF (1 wt%) were injected into the flask with
constant, where n = 1.8 is acceleration factor). In contrast, the control carefully controlled N2 flow.
QLED exhibit a much shorter lifetime, with T50 being 1.9 h at For the growth of IPG shell, after synthesizing of ZnSeTe cores, the
2941 cd m–2 (corresponding to T50 = 13.3 h at 1000 cd m–2). These temperature was increased to 280 °C, and 4 mL of Zn(OA)2 was added
results demonstrate that the suppression of Auger recombination to the flask. Then, over a period of 0.7 h, 0.2 mL Se precursor and
through the engineering of the interfacial potential plays a critical role 0.45 mL Te precursor were injected using a syringe pump to grow the
in enhancing both the efficiency and stability of QLEDs. The substantial IPG shell. For the control QDs, the same initial steps were followed,
improvement in both performance and operational lifetime indicates with the temperature raised to 280 °C, and 4 mL of Zn(OA)2 added.
the great potential of IPG QDs for practical, long-lasting applications in However, only 0.2 mL of Se precursor was injected over 0.7 h using a
high-performance displays and lighting technologies. syringe pump, without the addition of Te, to form the standard core/
shell structure.
Discussion Growth of ZnSe/ZnS shell: After the formation of IPG or control
In conclusion, we have successfully developed efficient ZnSeTe green shell, the temperature was increased to 310 °C. At this temperature,
QLEDs by controlling the interfacial potential. The proposed IPG 10 mL Zn(OA)2 and 2.0 mL Se precursor were injected over 1 h. Fol-
structure design mitigates the lattice strain in ZnSeTe QDs and reduces lowing this, the temperature was maintained at 310 °C, and 10 mL of
their nonradiative Auger recombination rate. Time-dependent and Zn(OA)2 and 2.0 mL of S precursor were injected over another 1 h to
excitation-density-dependent TA kinetics reveal that Auger recombi- coat the ZnS shell. The reaction was then cooled to 250 °C, and 20 mL
nation in IPG QDs is reduced by an order of magnitude compared to of Zn(OLA)2 and 4 mL of S precursor were injected over 0.5 h, com-
that in control QDs with an abrupt interfacial potential structure. As a pleting the shell coating process. After the reaction, the flask was
result, the IPG QDs achieve an impressive PL QY of up to 95%, with a cooled to room temperature. The resulting QDs were dispersed in
narrow linewidth of 38 nm. Furthermore, the QLEDs based on IPG QDs heptane and centrifuged at 1791 × g for 5 min. The supernatant was
exhibit a high EQE of 21.7% and a brightness of 52,040 cd m–2, along- purified with ethanol and centrifuged again at 1791 × g for 5 min for
side a long T50 lifetime of 99.4 h at 1000 cd m–2. This work represents a twice. The supernatant solution was discarded, and the precipitate was
significant advancement toward the use of eco-friendly QDs-based dispersed in 10 mL octane for further characterization.
QLEDs for display and lighting applications.
Fabrication of green QLEDs
Methods The QLEDs were constructed based on the indium tin oxide (ITO)/
Materials PEDOT:PSS/PF8Cz/QDs/ZnMgO/Al structure. Except for the Al cath-
Zinc chloride (ZnCl2, 99.99%), zinc acetate (Zn(OAc)2, 99.99%), sulfur ode deposited by thermal evaporation under vacuum, the remaining
powder (S, 99.999%), selenium powder (Se, 99.999%), diethylzinc layers were spin-coated onto ITO glasses. The ITO-coated glasses
(DEZ; 1.0 M in ODE), trioctylphosphine (TOP, 97%), tellurium (99.99%), underwent successive ultrasonic cleaning in washing water, deio-
tri-n-octylamine (TOA, 90%), oleylamine (OLA, 90%), oleic acid (OA, nized water, acetone, and isopropanol, each for 30 min, and then 15-
90%), 1-octadecene (ODE, 90%), heptane (90%), octane (99%), Palmitic min ultraviolet ozone cleaning. The PEDOT:PSS (4083) solution was
acid (PA, 90%), 1-dodecanethiol (DDT, 98%), diphenylphosphine (DPP), spin-coated onto ITO glasses at 3000 rpm for 30 s and annealed at
HF (48%), tris[2-(diphenylphosphino)ethyl]phosphine were purchased 130 °C for 20 min in air. Subsequently, the PF8Cz solution (8 mg mL−1
from Aladdin. Poly-(ethylene dioxythiophene): polystyrene sulfonate dissolved in trimethylbenzene) was spin-coated at 3000 rpm for 30 s
(PEDOT:PSS), poly((9,9-dioctylfluorenyl-2,7-diyl)-alt-(9-(2-ethylhexyl)- and annealed at 130 °C for 20 min in a glove box. Next, the QDs
carbazole-3,6-diyl)) (PF8Cz), and zinc magnesium oxide (ZnMgO) solution (16 mg mL−1 in octane) was spin-coated at 3000 rpm for 30 s
nanocrystals which were purchased from Suzhou Xingshuo Nanotech and annealed at 80 °C for 5 min in the glove box. After that, the
Co., Ltd All chemicals were used without any purification. ZnMgO solution (20 mg mL⁻1 in ethanol) was spin-coated at
3000 rpm for 30 s, and the resulting films were annealed at 80 °C for
Preparation of precursors 10 min. After that, the devices were moved to a vacuum evaporator
Zn(OA)2 precursor: a 1000 mL flask was charged with a mixture with a vacuum level of 5 × 10−4 Pa to deposit a 100-nm-thick Al
consisting of 500 mmol of Zn(AC)2, 310 mL of OA, 140 mL of TOP, cathode at a speed of 5 nm s−1. Some devices were fabricated using

Nature Communications | (2025)16:1945 7

Article https://doi.org/10.1038/s41467-025-57304-9

 
Mesolight Inc’s “Hands-Liberated Diode (HLD-V1.0) Automated QD. Thus, the n1s can be expressed as:
Integration System” for reproducibility verification, which can auto-  
 
matically process PEDOTS:PSS/PF8Cz/QDs/ZnO layers (spin-coating n1s = 1
e
jp σ a 1 + j p σ a =2 ð4Þ
& annealing). In the glove box, all the devices were encapsulated
using commercially available ultraviolet-curable resin for subsequent The changes in 1S absorption and excitation density were fitted
characterization. using Eq. (4).

PL and structure characterization QLED characterization

PL spectra, PL QY, and TRPL of QDs were characterized using a A Keithley 2400 source meter and a spectrometer (Ocean Optics,
Horiba Fluorolog®-3 system equipped with a Synapse PLUS CCD QE65000) were utilized to measure the current density-voltage (J–V)
detection system, an F-3029 integrating sphere accessory, the characteristics and EL spectra of the QLEDs. A calibrated spectral
Quanta-φ, and a 450 W Xenon lamp along with associated acces- scanning colorimeter (Photo Research 655) was employed to measure
sories. For all samples, the excitation wavelength was set to 450 nm. the luminance of the QLEDs. The EQE of the QLEDs was calculated
Absorption spectra were measured with a UV-vis spectrophotometer according to the formula:
(PerkinElmer Instruments, Lambda 750). The QD films were prepared R
by spin-coating the QD solution on a cleaned glass at 2000 rpm for πLe IðλÞλdλ
η= R ð5Þ
15 s. X-ray diffraction measurements were performed over a 2θ range KmhcJ IðλÞV ðλÞdλ
of 20°–70°, using a Cu Kα (λ = 1.5405 Å) rotating anode with a
SMARTLAB 3 kW X-ray diffractometer (Rigaku Corporation). TEM where e represents the electron charge, h stands for the Planck con-
images were acquired using a FEI Talos F200X G2 microscope stant, c is the velocity of light, and Km = 683 lm W−1 is the maximum
operating at 200 kV. HAADF STEM images, EDS elemental mapping, luminous efficacy. J refers to the current density, I(λ) refers to the
and line scan images of QDs were measured using a Thermo Scientific relative electroluminescence intensity at wavelength λ, V(λ) is
Spectra 300 microscope. The VBM of QDs was estimated using a UV the normalized photonic spectral response function, and L represents
photoelectron spectrometer (Thermofisher ESCALAB 250Xi) with a the luminance.
He I (21.2 eV) photon source. XPS data were obtained using a Thermo
Scientific K-Alpha system with a monochromatic Al Kα1 X-ray source Data availability
(hν = 1486.6 eV, where h is the Planck constant and ν is the frequency The data that support the findings of this study are available from the
of the X-ray photons). corresponding author upon request.

Raman characterization References

Raman spectra of the QDs were performed using a confocal Raman 1. Shirasaki, Y., Supran, G. J., Bawendi, M. G. & Bulović, V. Emergence
spectral system (alpha 300 R) with a 532 nm excitation wavelength and of colloidal quantum-dot light-emitting technologies. Nat. Photon.
a 50× objective lens at room temperature. The Raman spectra were 7, 13–23 (2012).
calibrated using a standard Si wafer prior to the measurements. The 2. Dai, X. et al. Solution-processed, high-performance light-emitting
prepared samples were mounted on piezoelectric stages, and the laser diodes based on quantum dots. Nature 515, 96–99 (2014).
light was focused on the sample using a lens with a numerical aperture 3. Yoo, J. et al. Highly efficient printed quantum dot light-emitting
of 0.82, producing a spot ~1 μm in diameter. The emitted light was diodes through ultrahigh-definition double-layer transfer printing.
collected by the same objective and scattered by a 0.75 m spectro- Nat. Photon. 18, 1105–1112 (2024).
meter, equipped with 150 and 1800 lines/mm gratings and a CCD 4. Vasilopoulou, M. et al. Efficient colloidal quantum dot light-emitting
camera. diodes operating in the second near-infrared biological window.
Nat. Photon. 14, 50–56 (2019).
TA characterization and analyses 5. Dai, X., Deng, Y., Peng, X. & Jin, Y. Quantum‐dot light‐emitting
Femtosecond pump-probe setup. A Ti: Sapphire laser (Coherent Libra) diodes for large‐area displays: towards the dawn of commerciali-
was employed to generate femtosecond pulses at 800 nm. These zation. Adv. Mater. 29, 1607022 (2017).
pulses were directed towards an optical parametric amplifier (OperA 6. Meng, T. et al. Ultrahigh-resolution quantum-dot light-emitting
Solo from Coherent), yielding femtosecond pulses (~50 fs, 1 kHz) with diodes. Nat. Photon. 16, 297–303 (2022).
wavelengths spanning from 400 to 2500 nm and a maximum average 7. Li, M. et al. Ultrabright and stable top-emitting quantum-dot light-
power of around 500 mW. The idler light served as the pump light. The emitting diodes with negligible angular color shift. Nat. Commun.
probe light, a white-light continuum, was generated by focusing the 15, 5161 (2024).
800 nm laser beam onto a sapphire plate. The electronic signal from 8. Xu, H. et al. Dipole–dipole-interaction-assisted self-assembly of
the detector was amplified by a lock-in amplifier (Stanford Research quantum dots for highly efficient light-emitting diodes. Nat. Photon.
Systems: SR830) and synchronized with the laser’s trigger frequency. 18, 186–191 (2024).
TA fitting analysis. The QD populations were modeled using a 9. Gao, Y. et al. Minimizing heat generation in quantum dot light-
Poisson distribution, given that pump wavelength saturation was emitting diodes by increasing quasi-Fermi-level splitting. Nat.
negligible. The average QD population, hN i, was described by: Nanotechnol. 18, 1168–1174 (2023).
PðNÞ = hN iN e
hN i =N!. The change in 1S absorption (ΔA)
 is proportional
  10. Deng, Y. et al. Solution-processed green and blue quantum-dot
to the population of the 1S electron state: 4A / n1s , which n1s light-emitting diodes with eliminated charge leakage. Nat. Photon.
represents the average occupation number of the 1S electron state. 16, 505–511 (2022).
Considering
  the twofold spin degeneracy for the 1S electron state, they 11. Shen, H. et al. Visible quantum dot light-emitting diodes with
n1s can be expressed as: simultaneous high brightness and efficiency. Nat. Photon. 13,
192–197 (2019).
  
n1s = 1
e
hN i 1 + hN i=2 ð3Þ 12. Xing, S. et al. Energy-efficient perovskite LEDs with rec 2020
compliance. ACS Energy Lett. 9, 3643–3651 (2024).
here, the hN i is proportional to the pump fluence (jp) and can be 13. Kim, J. S. et al. Ultra-bright, efficient and stable perovskite light-
expressed as: hN i= jpσa, where σa is the absorption cross-section of the emitting diodes. Nature 611, 688–694 (2022).

Nature Communications | (2025)16:1945 8

Article https://doi.org/10.1038/s41467-025-57304-9

14. Yuan, S. et al. Efficient blue electroluminescence from reduced- 38. Zhang, J., Li, C., Li, J. & Peng, X. Synthesis of CdSe/ZnSe core/shell
dimensional perovskites. Nat. Photon. 18, 425–431 (2024). and CdSe/ZnSe/ZnS core/shell/shell nanocrystals: surface-ligand
15. Yuan, F. et al. Bright and stable near-infrared lead-free perovskite strain and CdSe–ZnSe lattice strain. Chem. Mater. 35,
light-emitting diodes. Nat. Photon. 18, 170–176 (2024). 7049–7059 (2023).
16. Zhao, J. et al. Efficient CdSe/CdS quantum dot light-emitting diodes 39. Van Avermaet, H. et al. Full-spectrum InP-based quantum dots with
using a thermally polymerized hole transport layer. Nano Lett. 6, near-unity photoluminescence quantum efficiency. ACS Nano 16,
463–467 (2006). 9701–9712 (2022).
17. Won, Y.-H. et al. Highly efficient and stable InP/ZnSe/ZnS quantum 40. Huang, Z. et al. Deciphering ultrafast carrier dynamics of eco-
dot light-emitting diodes. Nature 575, 634–638 (2019). friendly ZnSeTe-based quantum dots: toward high-quality blue-
18. Bian, Y. et al. Efficient green InP-based QD-LED by controlling green emitters. J. Phys. Chem. Lett. 12, 11931–11938 (2021).
electron injection and leakage. Nature 635, 854–859 (2024). 41. Kim, S. et al. Efficient blue-light-emitting Cd-free colloidal quantum
19. Kim, T. et al. Efficient and stable blue quantum dot light-emitting well and its application in electroluminescent devices. Chem.
diode. Nature 586, 385–389 (2020). Mater. 32, 5200–5207 (2020).
20. Yu, P. et al. Highly efficient green InP-based quantum dot light- 42. Jiang, Y. et al. Reducing the impact of Auger recombination in
emitting diodes regulated by inner alloyed shell component. Light. quasi-2D perovskite light-emitting diodes. Nat. Commun. 12,
Sci. Appl. 11, 162 (2022). 336 (2021).
21. Bi, Y. et al. Reducing emission linewidth of pure‐blue ZnSeTe 43. Yamada, Y., Yasuda, H., Tayagaki, T. & Kanemitsu, Y. Temperature
quantum dots through shell engineering toward high color purity dependence of photoluminescence spectra of nondoped and
light‐emitting diodes. Small 19, 2303247 (2023). electron-doped SrTiO3: crossover from auger recombination to
22. Lee, S.-H. et al. Heterostructural tailoring of blue ZnSeTe quantum single-carrier trapping. Phys. Rev. Lett. 102, 247401 (2009).
dots toward high-color purity and high-efficiency electro- 44. Wu, B. et al. Excited-state regulation in eco-friendly ZnSeTe-based
luminescence. Chem. Eng. J. 429, 132464 (2022). quantum dots by cooling engineering. Sci. China Mater. 65,
23. Chao, W.-C. et al. High efficiency green InP quantum dot light- 1569–1576 (2022).
emitting diodes by balancing electron and hole mobility. Commun. 45. Cai, W. et al. Emission mechanism of bright and eco‐friendly
Mater. 2, 96 (2021). ZnSeTe quantum dots. Adv. Opt. Mater. 12, 2301970 (2023).
24. Moon, H. et al. Composition-tailored ZnMgO nanoparticles for 46. Gu, Y. et al. Zn−Se−Te multilayers with submonolayer quantities of
electron transport layers of highly efficient and bright InP-based Te: type-II quantum structures and isoelectronic centers. Phys.
quantum dot light emitting diodes. Chem. Commun. 55, Rev. B 71, 045340 (2005).
13299–13302 (2019). 47. Chang, J. H. et al. Impact of morphological inhomogeneity on
25. Liu, P. et al. Green InP/ZnSeS/ZnS core multi‐shelled quantum dots excitonic states in highly mismatched alloy ZnSe1–XTeX nanocrys-
synthesized with aminophosphine for effective display applica- tals. J. Phys. Chem. Lett. 13, 11464–11472 (2022).
tions. Adv. Funct. Mater. 31, 2008453 (2021). 48. Imran, M. et al. Molecular‐additive‐assisted tellurium homogeniza-
26. Heo, H. S. et al. Enhanced fluorescence in green ZnSeTe quantum tion in ZnSeTe quantum dots. Adv. Mater. 35, 2303528 (2023).
dots using gradient layer technique. Adv. Opt. Mater. 13, 49. Boldt, K. Raman spectroscopy of colloidal semiconductor nano-
2402215 (2025). crystals. Nano Futures 6, 012003 (2022).
27. Lee, S.-H. et al. ZnSeTe quantum dots as an alternative to InP and 50. Lu, S. H., Chen, T. F., Wang, A. J., Wu, Z. L. & Wang, Y. S. Lattice and
their high-efficiency electroluminescence. Chem. Mater. 32, optical property evolution of ultra-small ZnS quantum dots grown
5768–5775 (2020). from a single-source precursor. Appl. Surf. Sci. 299, 116–122 (2014).
28. Yoon, S.-Y. et al. Highly emissive green ZnSeTe quantum dots: 51. Zhang, Q. et al. Exciton-phonon coupling in individual ZnTe
effects of core size on their optical properties and comparison with nanorods studied by resonant Raman spectroscopy. Phys. Rev. B
InP counterparts. ACS Energy Lett. 8, 1131–1140 (2023). 85, 085418 (2012).
29. Pust, P., Schmidt, P. J. & Schnick, W. A revolution in lighting. Nat. 52. Zhu, H., Song, N., Rodríguez-Córdoba, W. & Lian, T. Wave function
Mater. 14, 454–458 (2015). engineering for efficient extraction of up to nineteen electrons from
30. Chen, B., Li, D. & Wang, F. InP quantum dots: synthesis and lighting one CdSe/CdS quasi-type II quantum dot. J. Am. Chem. Soc. 134,
applications. Small 16, 2002454 (2020). 4250–4257 (2012).
31. Wu, K., Lim, J. & Klimov, V. I. Superposition principle in auger 53. Nanda, J. et al. Absorption cross sections and Auger recombination
recombination of charged and neutral multicarrier states in semi- lifetimes in inverted core-shell nanocrystals: implications for lasing
conductor quantum dots. ACS Nano 11, 8437–8447 (2017). performance. J. Appl. Phys. 99, 034309 (2006).
32. Cragg, G. E. & Efros, A. L. Suppression of auger processes in con- 54. Sun, J. et al. Ultrafast carrier dynamics in CuInS2 quantum dots.
fined structures. Nano Lett. 10, 313–317 (2009). Appl. Phys. Lett. 104, 023118 (2014).
33. World Health Organization IARC. Monographs on the evaluation of
the carcinogenic risk of chemicals to humans: Cobalt in hard metals Acknowledgements
and cobalt sulfate, Gallium Arsenide, Indium Phosphide and This work was supported by National Key Research and Development
Vanadium Pentoxide. 86, (IARC, 2006). Program of China (Project No. 2022YFB3602800), Guangxi Science and
34. Huang, Z. et al. Broadband tunable optical gain from ecofriendly Technology Major Project (No. AA23073018), National Natural Science
semiconductor quantum dots with near-half-exciton threshold. Foundation of China (62475054, 62165001 and 12174075), Guangxi
Nano Lett. 23, 4032–4038 (2023). Natural Science Foundation (2022GXNSFFA035032), and the Guangxi
35. Kim, Y.-H. et al. Compositional and heterostructural tuning in red- Hundred-Talent Program. The Center for Instrumental Analysis of
emissive ternary ZnSeTe quantum dots for display applications. Guangxi University is acknowledged for providing research facilities and
ACS Appl. Nano Mater. 6, 19947–19954 (2023). resources for the experiments.
36. Wang, X. et al. Non-blinking semiconductor nanocrystals. Nature
459, 686–689 (2009). Author contributions
37. Lee, Y. et al. Effectual interface and defect engineering for Auger J.Z. and Y.W. supervised the research. S.C. and Y.B. conceptualized and
recombination suppression in bright InP/ZnSeS/ZnS quantum dots. designed the experiments. Y.B. performed most of the experiments and
ACS Appl. Mater. Interfaces 14, 12479–12487 (2022). analyzed the data. J.S. conducted the TA measurements. B.Z., X.Y., J.Z.,

Nature Communications | (2025)16:1945 9

Article https://doi.org/10.1038/s41467-025-57304-9

and Q.L. provided technical support and advice. S.C., Y.B., and J.Z. wrote Publisher’s note Springer Nature remains neutral with regard to jur-
and revised the manuscript. All authors contributed to the analysis and isdictional claims in published maps and institutional affiliations.
discussion of the experimental results.
Open Access This article is licensed under a Creative Commons
Competing interests Attribution-NonCommercial-NoDerivatives 4.0 International License,
The authors declare no competing interests. which permits any non-commercial use, sharing, distribution and
reproduction in any medium or format, as long as you give appropriate
Additional information credit to the original author(s) and the source, provide a link to the
Supplementary information The online version contains Creative Commons licence, and indicate if you modified the licensed
supplementary material available at material. You do not have permission under this licence to share adapted
https://doi.org/10.1038/s41467-025-57304-9. material derived from this article or parts of it. The images or other third
party material in this article are included in the article’s Creative
Correspondence and requests for materials should be addressed to Commons licence, unless indicated otherwise in a credit line to the
Sheng Cao, Yunjun Wang, Bingsuo Zou or Jialong Zhao. material. If material is not included in the article’s Creative Commons
licence and your intended use is not permitted by statutory regulation or
Peer review information Nature Communications thanks Dongho Kim exceeds the permitted use, you will need to obtain permission directly
and the other anonymous reviewer(s) for their contribution to the peer from the copyright holder. To view a copy of this licence, visit http://
review of this work. A peer review file is available. creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions information is available at © The Author(s) 2025

http://www.nature.com/reprints

Nature Communications | (2025)16:1945 10