A reactivity-controlled epitaxial growth strategy for synthesizing large nanocrystals

Large ZnSe nanocrystals are expected to be promising blue-light emitters with an emission peak of 455–475 nm, which is important for the construction of display apparatus. The final size of ZnSe nanocrystals via one-step injection can be varied by the reactivity of the Zn and Se precursors; however, it has a limit of <5 nm. To describe the key factors in determining the final size of ZnSe nanocrystals, we proposed a nuclei number-considered LaMer model based on the Maxwell–Boltzmann distribution of crystal embryos. As a result, a general strategy of reactivity-controlled epitaxial growth was developed to synthesize large ZnSe nanocrystals through sequential injection of high-reactivity and low-reactivity Zn and Se precursors. The resultant ZnSe nanocrystals achieved pure blue emission between 455 and 470 nm. We further fabricated stable, large ZnSe/ZnS core–shell nanocrystals with photoluminescence quantum yields up to approximately 60%. Moreover, the reactivity-controlled epitaxial growth strategy is versatile and could be used to synthesize large ZnSe, CdSe and PbSe nanocrystals with average sizes up to 35 nm, 76 nm and 87 nm, respectively. The control of quantum-confined and classical effects in these large semiconductor nanocrystals will open up new directions for fundamental research and application exploration. Synthesizing Se-based nanocrystals with large diameters remains challenging. Here, a reactivity-controlled epitaxial growth strategy was demonstrated to synthesize nanocrystals of ZnSe, CdSe and PbSe with average diameters of 35 nm, 76 nm and 87 nm, respectively. The large ZnSe nanocrystals emitted pure blue light, which is important for display technology.

Large ZnSe nanocrystals are expected to be promising blue-light emitters with an emission peak of 455-475 nm, which is important for the construction of display apparatus. The final size of ZnSe nanocrystals via one-step injection can be varied by the reactivity of the Zn and Se precursors; however, it has a limit of <5 nm. To describe the key factors in determining the final size of ZnSe nanocrystals, we proposed a nuclei number-considered LaMer model based on the Maxwell-Boltzmann distribution of crystal embryos. As a result, a general strategy of reactivitycontrolled epitaxial growth was developed to synthesize large ZnSe nanocrystals through sequential injection of high-reactivity and lowreactivity Zn and Se precursors. The resultant ZnSe nanocrystals achieved pure blue emission between 455 and 470 nm. We further fabricated stable, large ZnSe/ZnS core-shell nanocrystals with photoluminescence quantum yields up to approximately 60%. Moreover, the reactivity-controlled epitaxial growth strategy is versatile and could be used to synthesize large ZnSe, CdSe and PbSe nanocrystals with average sizes up to 35 nm, 76 nm and 87 nm, respectively. The control of quantum-confined and classical effects in these large semiconductor nanocrystals will open up new directions for fundamental research and application exploration.
Semiconductor nanocrystals, also known as quantum dots, have been intensively investigated due to their well-known size-dependent effects, resulting in a large family of functional nanomaterials with tunable optical properties for many cutting-edge technologies 1,2 . In the field of display technology, blue emitters with emission peaks of 455-475 nm are highly desirable for the construction of electroluminescence-based display apparatus, which will promote the industrialization of printable display technology 3 . ZnSe nanocrystals are considered to be potential blue-light emitters in view of their suitable bandgap (2.7 eV) [4][5][6][7] . To avoid blue shifts due to quantum confinement effects, the fabrication of large ZnSe nanocrystals is of vital importance to tune their emission into the pure blue region.
Over more than 30 years, researchers have thoroughly investigated nanocrystal synthesis, resulting in the state-of-the-art CdSe nanocrystals and the corresponding nucleation and growth mechanisms [8][9][10][11][12][13][14][15] . Nevertheless, there are few reports on CdSe nanocrystals with an average size over 10 nm 16 . Even when applying a multiplestep injection method, the average size of giant CdSe-based core-shell nanocrystals was limited to 20 nm [17][18][19] . Increasing a nanocrystal's size leads to a decrease in the surface chemical potential due to the reduction of surface to volume ratio 13 . Under high concentration or multiplestep injections, monomers tend to grow epitaxially on crystal planes with higher chemical potential 20,21 , resulting in the formation of anisotropic nanocrystals, such as nanorods 22 , nanosheets 23 or branched nanocrystals 24 . Similar anisotropic structures have been also observed in ZnSe-based nanocrystals [25][26][27][28] . However, it remains a great challenge to fabricate large nanocrystals with isotropic shapes. Article https://doi.org/10.1038/s44160-022-00210-5 prepared as Zn precursors by dissolving zinc acetate (Zn(OAc) 2 ) into a mixed solvent of OLA, OA and ODE. The Se precursors were prepared by dissolving Se powder into diphenylphosphine (DPP), TOP or ODE, respectively. The reactivity is mainly determined by the bonds between cationic and coordinated functional groups 32 , which can be described by comparing the nucleation temperature 33 and the reaction time 34 .
In this work, the nucleation temperature was determined by monitoring the appearance of an exciton absorption peak in ultravioletvisible (UV-vis) absorption spectra, which can be validated by the transmission electron microscopy (TEM) observations ( Supplementary  Fig. 1). Supplementary Figures 2-4 show the UV-vis absorption spectra of aliquots taken from the reaction between OA-[Zn(OAc) 2 ]-OLA complexes and Se precursors at indicated temperatures. The results show the nucleation temperature is strongly correlated with the Se precursors with a sequence of Se-DPP (140-185 °C), Se-TOP (230-270 °C) and Se-ODE (250-280 °C) ( Supplementary Fig. 5). The reactivity of Se precursors was further investigated by applying nuclear magnetic resonance (NMR) spectroscopy measurements. The chemical shift (Δδ) between Se-DPP and free DPP (Δδ = 47.66 ppm) is smaller than that of the shifting between Se-TOP and free TOP (Δδ = 67.11 ppm), indicating the weaker coordination binding of Se-P in Se-DPP (Supplementary Fig. 6). Based on the above results, the reactivity of Se precursors has a sequence of Se-DPP ≫ Se-TOP > Se-ODE. With the OA:OLA ratio increasing, the nucleation temperature between OA-[Zn(OAc) 2 ]-OLA complexes and Se-DPP can be varied from 140 °C In this work, we investigated the influence of the reactivity of Zn and Se precursors on the nucleation and growth of ZnSe nanocrystals via a hot-injection method. Based on the experimental results and previous theoretical models, we developed a nuclei number-considered LaMer model to illustrate the correlation between the final size of nanocrystals and the total number of nuclei. A general strategy of reactivity-controlled epitaxial growth (RCEG) was proposed to synthesize large-diameter ZnSe nanocrystals with an average size over 35 nm. The resultant large ZnSe nanocrystals show pure blue emission between 455 and 470 nm with narrow photoluminescence (PL) fullwidth at half-maximum (FWHM; 16-25 nm). After coating with a ZnS shell, we obtained nearly spherical ZnSe/ZnS core-shell nanocrystals with PL quantum yield (PLQY) of 60%. The RCEG strategy could also be extended to synthesize large CdSe and PbSe nanocrystals.

Reactivity control of ZnSe nanocrystal synthesis
Considering ZnSe nanocrystals synthesis, most of the previous studies adapted the reaction routes from CdSe nanocrystals, using zinc carboxylates (such as zinc oleate or zinc stearate) and selenium (Se) coordinates (such as Se-trioctylphosphine (TOP) or Se-1-octadecene (ODE)) as precursors [3][4][5][6][7][25][26][27][28][29][30]    We further studied the effects of the precursor's reactivity on the nucleation and growth of ZnSe nanocrystals via hot-injection method. Figure 1a-e shows the evolution of UV-vis absorption and PL spectra of ZnSe nanocrystals with reaction time under different Zn and Se precursors. Both the absorption and PL emission spectra gradually red shifted, then approached a maximum value with the reaction prolonging. Figure 1f plots the FWHM and PL peaks of the final products. The use of lower reactivity precursors can result in a narrower FWHM and shorter PL peak wavelength. As is well known, the PL peak of nanocrystal ensembles depends on their average size, whereas the FWHM of PL emission is mainly correlated with the size distribution and defect-related broadening 8,10 . The TEM results show that higher reactivity results in a larger final average size and broader size distribution ( Supplementary Fig. 10). To track the size evolution with prolonged reaction time, we derived the size-dependent optical bandgaps by plotting the PL and absorption (Abs) peaks versus diameters (Fig. 1g, the corresponding TEM results and PL spectra are shown in Supplementary Fig. 11). Figure 1h shows the size evolution of ZnSe nanocrystals obtained from different precursors. Notably, the resultant ZnSe nuclei underwent an exponential growth process at the beginning, then approached a maximum diameter at the end. Based on the above results, the use of higher reactivity precursors can induce the formation of larger ZnSe nanocrystals. Even when using highly reactive Se-DPP and OA-[Zn(OAc) 2 ]-OLA (OA:OLA = 0.2), the final average size of the resultant ZnSe nanocrystals was limited to around 5 nm, with a PL peak at 425 nm. Here, we try to understand the limitation of the final size of ZnSe nanocrystals from the viewpoint of formation mechanism.

Insights into nucleation and growth of monodisperse nanocrystals
As schematically shown in Fig. 2a, the formation of colloidal nanocrystals contains at least three stages, including monomer formation, nucleation and growth 36 . If the conversion rate from precursor to monomer is very fast, the nucleation process can be accomplished in a very short time period, which is denoted as 'burst nucleation'. If the nucleation process coexists with the conversion from precursor to monomer in a much longer time period, the nucleation and growth process cannot be effectively separated, resulting in a 'continuous nucleation' event 37,38 . According to previous reports, both 'burst nucleation' and 'continuous nucleation' are effective to explain the fabrication of monodisperse nanocrystals 11,38 .
To clarify the nucleation process of ZnSe nanocrystals, we determined the conversion rate of [ZnSe] units (monomers) from precursors and the concentration of ZnSe nanocrystals. Figure 2b plots the time-resolved absorbance at 300 nm of the aliquots taken from the reaction between OA-[Zn(OAc) 2 ]-OLA (OA:OA = 0.2, 1.0) and Se-DPP. The results suggest that Zn and Se precursors completely converted into [ZnSe] units from 3 seconds after injection. As shown in Fig. 2c, the concentration of ZnSe nanocrystals was kept at a fixed value (the timeresolved absorption spectra as shown in Supplementary Fig. 12). These results indicate that the synthesis of ZnSe nanocrystals experienced a 'burst nucleation' process. In addition, the comparison between these two reactions implies that the use of lower reactivity precursors can induce more nanocrystals. The growth mechanism can be derived by monitoring the evolution of size and size distribution with prolonged reaction time [38][39][40] . Supplementary Fig. 13 presents the growth rate (dr/dt) versus diameter of nanocrystals, which was obtained by fitting the experimental data in Fig. 1h. The observed size narrowing can be assigned to the domination of diffusion-controlled growth process (Fig. 2d, the simulated evolution of size distribution is shown in Supplementary Fig. 14, details are described in Supplementary Section 3). Herein, we adapted a modified LaMer model that was developed from classical CdSe nanocrystals synthesis to further illustrate the nucleation and growth of ZnSe nanocrystals (the LaMer picture is shown in Fig. 2e) (refs. 41-45). Because the nuclei concentration remains unchanged during the growth of ZnSe nanocrystals, the final size (R f ) of resultant ZnSe nanocrystals is mainly determined by the ratio between the total amount of monomers and the total number of nuclei (n nuclei ), which can be estimated by equation (1), where M total is the total amount of monomers and φ is the proportion of atomic volume in unit cell.
To gain deep insights into the influence of precursor reactivity on final size of ZnSe nanocrystals, we first analysed the key factors in determining n nuclei . Although many previous works have derived the equation of nucleation rate (dN/dt) from chemical dynamics, it is still a great challenge to calculate the n nuclei due to the difficulty in determining the nucleation time 46,47 . According to the experimental results and previous theoretical models, we developed a nuclei number-considered LaMer model based on the Maxwell-Boltzmann distribution of crystal embryos. In the initial stage of nucleation, the aggregation of monomers induces the formation of crystal embryos. Nuclei are crystal embryos with sizes exceeding the critical nucleation radius (r*). From the viewpoint of thermodynamics, the size distribution of crystal embryos follows a Maxwell-Boltzmann statistic as shown in equation (2) In equation (2), ΔN r is the number of crystal embryos of size r, N m is the total number of monomers, A is the pre-exponential factor, V m is the molar volume of the crystal, γ is the specific surface energy, R is the gas constant, T is the absolute temperature, ω is the degree of supersaturation and k is the Boltzmann constant. Figure 2f shows the general profile of Maxwell-Boltzmann size distribution of crystal embryos at a constant temperature at which the nucleation occurred. Hence, the n nuclei can be estimated by the integral of the size distribution function  (r* to ∞ ). Based on the above discussion, the value of r* is the key factor to determine n nuclei and the final size of nanocrystals in one-step injection synthesis. From the classical Gibbs equation, r* can be derived as equation (3) (ref. 11). In nanocrystal synthesis, r* can also be described as equation (4) from the modified Gibbs-Thompson equation 49 . The r* found in equation (4) is one-third larger than the value in equation (3), indicating the Gibbs-based equations are not self-consistent. The difference can be attributed to neglect of the size and ligand effects on surface energy. In detail, the value of γ of a nanoparticle is strongly correlated with its size, surface ligands and solvent medium 50,51 . Basically, ligands 'push' on the surface of nuclei to stabilize it by moderating the surface energy. The binding affinity between ligands and as-formed nuclei strongly affects the surface energy of the nucleus.
Based on the proposed model above, the strong binding affinity between OA and ZnSe nuclei induced the decreasing of nuclei surface energy, resulting in the decrease of r*. As a result, the n nuclei increased as the OA concentration increased (Supplementary Table 1). Therefore, according to equation (1), the increase in the total number of ZnSe nuclei accounts for the final size of ZnSe nanocrystals decreasing. As discussed above, except for γ, ω also determined the value of r*. We further investigated the influence of ω (precursor concentration) on as-formed n nuclei . As shown in Supplementary Fig. 15, the final size of resultant ZnSe nanocrystals decreased as precursor concentration increased. The increase of precursor concentration resulted in the increase of ω, thus inducing the decrease of r*. As a result, the decrease of the ZnSe nanocrystals' size with increasing precursor concentration can be explained by the increase of n nuclei . It is concluded that the nucleation process in ZnSe nanocrystals synthesis can be well described by the nuclei number-considered LaMer model. To understand the challenges for fabricating large nanocrystals, we discussed a mathematical model of diffusion-controlled growth (detailed in Supplementary Section 4; Supplementary Fig. 16). It is assumed that a nanocrystal is grown layer by layer, and all monomers in the diffusion sphere satisfy the monolayer growth. Figure 2g illustrates the calculation result of evolution of the diffusion radius (R dif ) with nanocrystal radius (R) during one-step growth. R dif increases dramatically when reaching the critical value (diffusion spheres are tangential), indicating that further growth of the nanocrystal becomes extremely difficult (as indicated by the change of the ratio of R dif to R, as shown in Supplementary Fig. 17b). The size evolution of ZnSe nanocrystals with prolonged time can be predicted by considering the monomer concentration in solution (C L ) over time ( Supplementary Fig. 18a), which is consistent with our experimental results and the above discussions. Therefore, we conclude that the final size of ZnSe nanocrystals obtained from hot-injection synthesis has a limited value (depending on n nuclei ). Although n nuclei can be varied by tuning the reactivity or concentration of Zn and Se precursors, the scope of this variation is limited. The size limitation can theoretically be broken by epitaxial growth, such as through continuous injection of the precursors (Supplementary Fig. 19a). However, to our knowledge, there is no existing reaction system to obtain large-diameter ZnSe nanocrystals with expected size (over 10 nm) for pure blue light-emitting applications (Supplementary Table 2

RCEG of large ZnSe nanocrystals
Epitaxial growth is a typical methodology to grow core-shell nanocrystals, which can be realized via continuous injection of precursors into small seeds [52][53][54] . In comparison with small nanocrystals, large nanocrystals are difficult to grow due to their lower surface reactivity 38 . With size increasing, the fraction of edge (weaker binding) sites decreases, which is unfavourable for monomer adsorption. In that case, the surface reaction will limit the growth of large nanocrystals 55 . In this work, we developed a versatile strategy of RCEG for fabricating large spherical ZnSe nanocrystals with pure blue emission by sequential injection of high-and low-reactivity Zn and Se precursors at a high temperature (300 °C). The secondary nucleation during epitaxial growth was suppressed by adding lower reactivity precursors and choosing a befitting replenishment rate of precursors, and nearly spherical large-diameter ZnSe nanocrystals were obtained. In a typical synthesis, high-reactivity Zn and Se precursors are employed for fabricating ZnSe seeds, whereas lower reactivity Zn and Se precursors are added for further epitaxial growth into large ZnSe nanocrystals.  Figure  3b,c shows the evolution of the UV-vis absorption and PL spectra with the RCEG synthesis prolonging. Excitingly, the PL peaks of these ZnSe nanocrystals shifted from 425 nm to 470 nm during the RCEG process, and a sample with a FWHM of 25 nm at 460 nm was obtained. As shown in Fig. 3d-

The fabrication of large ZnSe and ZnS core-shell nanocrystals
To enhance the PL efficiency and photostability of large ZnSe nanocrystals, ZnS (bandgap, 3.7 eV) was chosen as a shell coating to passivate the surface defects of ZnSe nanocrystals. In our work, we used ZnSe nanocrystals with an average diameter of approximately 9 nm, PL peak of 455 nm, FWHM of 22 nm and PLQY of 23% for shell coating. The ZnS shell coating was performed by successively injecting sulfur (S)-TOP (to form the first layer of ZnS shell, marked as ZnS1) and preprepared Zn-S precursors (to form the second layer of ZnS shell, marked as ZnS2). Figure 4a shows the evolution of the absorption and PL spectra after the ZnS shell growth on the ZnSe cores. The absorbance at 365 nm remained the same and the absorbance at 300 nm increased after the epitaxial growth of ZnS on the surface of ZnSe cores, indicating that ZnS was successfully coated on ZnSe cores. Figure 4b shows the evolution of PLQY, PL peak and FWHM during the ZnS shell growth. The PLQY reached 60% as the shell thickness increased to approximately four monolayers and then started to decrease, indicating the occurrence of interfacial strain-related defects 6 .
As compared in time-resolved PL decaying profiles of nanocrystal dispersions with ZnSe core, and ZnSe/ZnS1 and ZnSe/ZnS2 core-shell nanocrystals ( Supplementary Fig. 21), average lifetimes (τ av ) were determined by the fitting of a bi-exponential function to be 26.7, 53.5 and 75.0 ns, respectively. The prolonged lifetime is mainly attributed to the reduced non-radiative defect-assisted recombination due to surface passivation. Upon growth of the ZnS shell, three major XRD peaks shifted to higher angles due to the smaller ZnS lattice constant compared with ZnSe (Fig. 4c). Figure 4d-f provides the TEM, highresolution TEM and corresponding fast Fourier Transform images of ZnSe, ZnSe/ZnS1 and ZnSe/ZnS2, respectively. Both lattice structure and fast Fourier Transform correspond to the zinc blende structure. Such large ZnSe/ZnS core-shell nanocrystals show pure blue emission (approximately 455 nm) and narrow ensemble PL width (approximately 22 nm), which are ideal emitter properties for display applications.

RCEG of large CdSe and PbSe nanocrystals
To verify the generality of the RCEG strategy for large nanocrystal synthesis, we further fabricated large CdSe and PbSe nanocrystals. First, monodispersed small CdSe and PbSe seeds (<5 nm, the TEM images are shown in Supplementary Figs. 22a and 23a) were obtained through a hot-injection method. Then, epitaxial growth of CdSe and PbSe seeds was achieved by continuous injection of moderate-reactivity cationic and anionic precursors. Note that during the nucleation and growth process, after using our precise reactivity regulating strategy of precursors, secondary nucleation can be effectively suppressed. Figure 5 shows the TEM images of CdSe and PbSe nanocrystals that were

Conclusions
In summary, we investigated the reactivity of Zn and Se precursors to illustrate the key factors in determining the final size of ZnSe nanocrystals via a hot-injection method. It was found that the final size can be varied by using different precursors and ligands. We further illustrated that the final size of nanocrystal synthesis has a critical value that is mainly determined by the total number of nuclei. A nuclei number-considered LaMer model based on the Maxwell-Boltzmann distribution of crystal embryos was developed to describe the influence of precursors and ligands on the critical nucleation radius, which provide insights into the correlations between the critical nucleation radius and total number of nuclei. Based on this understanding, we designed a general RCEG strategy to synthesize large-diameter ZnSe, CdSe and PbSe nanocrystals with average sizes over 35 nm, 76 nm and 87 nm, respectively. The large ZnSe nanocrystals show strong PL emission peaking up to 470 nm. We further fabricated efficient and stable large ZnSe/ZnS core-shell nanocrystals with QYs up to 60%. We believe that the control of quantum-confined and classical effects in these large semiconductor nanocrystals provides new opportunities for research and applications.

Preparation of Se-DPP, Se-TOP and Se-ODE stock solution
The [0.4 M] Se-DPP stock solution was prepared as follows: 0.48 mmol of Se powder was added to 1.2 ml of DPP in a N 2 filled glove box. The mixture was heated to 80 °C and became a transparent faint yellow solution that was kept in the glove box at room temperature for further use. The preparation of [0.4 M] Se-TOP stock solution is the same as above, except that DPP is replaced by TOP. The [0.4 M] Se-ODE stock solution was prepared as follows: a mixture of 4 mmol of Se powder and 10 ml of ODE solvent was kept under vacuum at 120 °C for 40 min, then heated to 220 °C under N 2 for 30 min. Then the solution was heated to 240 °C for 60 min until it became a transparent brown solution, before being cooled down to 100 °C for further use.

The nucleation temperature of ZnSe nanocrystal synthesis
A mixture of 0.4 mmol of Zn(OAc) 2 , x ml of OA (x = 0, 0.2, 0.4, 0.6, 0.8 and 1.0), 1 ml of OLA and 10-x ml of liquid paraffin was placed in a 50 ml threeneck flask, then stirred (900 r.p.m.) and heated to 120 °C under vacuum, and kept at this temperature for 40 min to obtain OA-[Zn(OAc) 2 ]-OLA precursors. After the atmosphere was changed to N 2 , 0.5 ml [0.4 M] Se-X stock solution (X represents DPP, TOP and ODE) was injected into the above Zn precursor solution at the same temperature. Subsequently, the mixed solution was heated at intervals of 5 °C (with a heating rate of 5 °C min −1 ) and kept for 10 min at that temperature. The nucleation temperature of ZnSe nanocrystals was determined by monitoring the timeresolved optical absorption spectra of aliquots at different temperatures.

Synthesis of small ZnSe nanocrystals
For synthesis of small ZnSe nanocrystals, 0.4 mmol of Zn(OAc) 2 , 0.2 ml of OA, 1 ml of OLA and 10 ml of liquid paraffin were placed in a 50 ml three-neck flask, then stirred (900 r.p.m.) and heated to 120 °C under vacuum, and kept at this temperature for 40 min. After the atmosphere was changed to N 2 , the mixed solution was heated to 280 °C, at which temperature 0.5 ml of Se-DPP was quickly injected. Subsequently, the temperature was increased to 300 °C (with a heating rate of 5 °C min −1 ) and reacted at this temperature for 30 min to obtain ZnSe seeds.

Epitaxial growth of large ZnSe nanocrystals
To obtain large ZnSe nanocrystals, the Zn-Se precursor was continuously injected into the preformed ZnSe nanocrystals at a rate of 3.6 ml h −1 at 300 °C. The Zn-Se precursor was prepared by mixing lowreactivity Zn (OA:OLA = 1) and Se-ODE stock solution. Zn stock solution was prepared by mixing 8 mmol of Zn(OAc) 2 , 3 ml of OA, 3 ml of OLA and 14 ml of ODE solvent in a three-neck flask by stirring, and the mixture was heated to 120 °C under vacuum and kept at that temperature for 40 min. After the atmosphere was changed to N 2 , the mixed solution was heated to 160 °C until it became a transparent faint yellow solution, then cooled down to 100 °C. Next, 10 ml [0.4 M] Se-ODE stock solution was added in the above Zn stock solution to obtain Zn-Se precursor.

Synthesis of ZnSe and ZnS core-shell nanocrystals
For the fabrication of Zn-S precursor, 8 mmol of Zn acetate, 8 mmol (1.6 ml) of octanethiol, 3 ml of OLA and 35.4 ml of ODE were mixed in a three-neck flask with vigorous stirring. The mixture was then heated to 120 °C under vacuum and kept at that temperature for 40 min to obtain a colourless transparent solution. To facilitate the growth of the first ZnS shell, 0.2 M TOP-S was added to the ZnSe core solution and further reacted for 1.5 h at 300 °C to form the first layer of ZnS shell, marked as ZnSe/ZnS1. After that, the solution was cooled to 280 °C for the second shell coating. The above Zn-S precursor was continually injected (6 ml h −1 ) into the reaction medium to form the second layer of ZnS shell, marked as ZnSe/ZnS2. Here, we balanced the precursor concentrations to control the reaction rates for the growth of the core and each shell, and maintained the metal precursor excess during the reaction.

The synthesis of large CdSe nanocrystals
For synthesis of CdSe seeds, 0.4 mmol of cadmium oxide, 0.5 ml of OA, 1 ml of OLA and 10 ml of ODE were placed in a three-neck flask, then stirred (900 r.p.m.) and heated to 120 °C under vacuum, and kept at this temperature for 40 min. After the atmosphere was changed to N 2 , the mixed solution was heated to 280 °C. Subsequently, 0.5 ml of 0.4 M Se-TOP was quickly injected into the mixed solution and further reacted for 30 min to produce CdSe seeds. For the synthesis of large CdSe nanocrystals, the RCEG method was used, Cd-Se precursor was continuously injected into the original CdSe seeds solution at a rate of 3.6 ml h −1 at 260 °C.

The synthesis of large PbSe nanocrystals
For the synthesis of PbSe seeds, 0.4 mmol of lead oxide, 1 ml of OA and 10 ml of ODE were placed in a three-neck flask, then stirred (900 r.p.m.) and heated to 120 °C under vacuum, and kept at that temperature for 40 min. After the atmosphere was changed to N 2 , the mixed solution was heated to 220 °C. Subsequently, 0.5 ml of 0.4 M Se-TOP was quickly injected into the mixed solution, and the temperature was kept at 220 °C and further reacted for 30 min. For the synthesis of large PbSe nanocrystals, the RCEG method was used, Pb-Se precursor was continuously injected into the original PbSe seeds solution at a rate of 3.6 ml h −1 at 200 °C.

Purification of nanocrystals
The raw nanocrystal solution was dissolved in hexane with a volume ratio of 1:1, which was precipitated by adding ethanol with volume ratio of 1:2 and then centrifuged at 7104g for 5 min. The precipitate was redispersed in hexane and the resultant nanocrystal products were purified by this method three times before further characterization. Article https://doi.org/10.1038/s44160-022-00210-5

Material characterizations
All TEM images were acquired on a FEI Talos F200S field-emission transmission electron microscope (FEI Co.) operated at 200 kV. A BRUKER D8 advance X-ray diffractometer equipped with a Cu Kα radiation source was used to record the XRD patterns. An AVANCE NEO 400 MHz Digital NMR Spectrometer (Bruker Corporation) was used to record the 31 P NMR spectra. The samples were dissolved in deuterated chloroform.

Spectroscopic measurements
Steady-state UV-vis absorption spectra were measured by using a UV-6100 spectrophotometer (Shanghai Mapada Instruments Co., Ltd.). PL spectra were measured by using a F-380 fluorescence spectrometer (Tianjin Gangdong Science and Technology Development Co., Ltd.). TR-PL spectra were obtained with a FLS1000 fluorescence lifetime spectrometer (Edinburgh). The PLQY of the samples was measured under excitation at 365 nm in comparison with 9,10-diphenylanthracene in cyclohexane as the reference, using equation (5). Usually, 9,10-diphenylanthracene with a high absolute PLQY of 95% is used as the reference for PLQY testing.
In equation (5), QY S and QY R are the QY of sample and reference, respectively. I S and I R are the integral area of the spectrum of sample and reference, respectively. A S and A R are the absorbance of sample and reference, respectively. n S and n R are the refractive index of sample and reference solvent, respectively.