1. Introduction

Colloidal quantum dots have emerged as a prominent research topic in nanomaterials science research due to their exceptional optoelectronic properties and wide-ranging applications, leading to their recognition with the 2023 Nobel Prize in Chemistry [1–7]. Among various QD materials, InP-based QDs exhibit great potential for replacing traditional CdSe-based QDs owing to their environment-friendly, high photoluminescence quantum yields (PLQYs) and good stability [8–10]. However, challenges such as expensive and highly toxic raw materials, stringent preparation conditions and non-classical nucleation models impede the further optimization of InP QD synthesis. Moreover, the performance of InP-based QDs still falls short compared to that of CdSe counterparts, particularly concerning emission linewidths [10–13]. Achieving high-quality InP QDs with narrow fluorescence emission poses considerable challenges both in terms of intrinsic principles and rational synthetic routes.

Over the past few years, silyl-phosphorus-based InP QDs have achieved remarkable progress by reaching emission linewidths of 33 nm for green (535 nm) and 35 nm for red (630 nm) emissions [9,14,15]. Nevertheless, several drawbacks associated with of silyl-phosphorus including high cost, toxicity and flammability hinder simple synthesis methods, limiting their further applications. Fortunately, amino-phosphorus chemistry offers an alternative approach that overcomes many limitations associated with silyl-phosphorus derivatives. Extensive studies have been conducted on controllable synthesis strategies for amino-phosphorus-based InP QDs in recent years [16–19]. However, a common challenge encountered is the relatively wide full width at half maximum (FWHM) observed in synthesized amino-phosphine-derived InP QDs. The synthesis of InP QDs using amino-phosphine involves intricate precursor conversion processes and necessitates the disproportionation reaction to achieve the desired P valency. Consequently, this gives rise to multiple competing reactions during the complex nucleation process, leading to broader emission linewidths of amino-phosphine-based InP QDs [18–21]. Owen et al. investigated the nucleation kinetics of various amino-phosphine derivatives at different reaction temperatures and highlighted a continuous nucleation process in these reactions, which deviates from the conventional understanding of explosive nucleation [21]. Therefore, it is imperative to explore novel models that diverge from classical nucleation for controllable synthesis of InP QDs.

Additionally, growing a stable inorganic shell can not only modify the surface defect states of cores, thereby enhancing the optical performance of QDs but also improve color purity. ZnSe, ZnS, and ZnSeS materials are commonly chosen as passive shells for InP QDs due to their matched lattice and large bandgaps [9,15,19,22–25]. However, during the shelling process, Zn ions tend to diffuse into the InP core and negatively impact the radiative recombination in InP QDs [26]. Talapin et al. proposed that Zn impurities in InP introduce disorder in electronic states of QDs leading to broadened emission [27]. Numerous experimental studies reported on this issue and demonstrated that suppressing the diffusion of Zn through low-temperature shell growth and surface modification with Se minimizes the emission linewidth of InP QDs [23,28–30]. Yang et al. suggested utilizing a quasi-ZnSe shell instead of a bulk ZnSe shell as an outer shielding layer to inhibit the cation exchange between core and shell [29,30]. These investigations have proposed various methods to suppress the Zn²⁺ diffusion, thus narrowing the fluorescence emission of the InP-based QDs. Meanwhile, the majority of current research efforts are dedicated to the design of rational synthetic routes, with limited attention being paid to the intrinsic principles involving exciton-phonon coupling.

In this study, we propose a multistage heating method to adjust the nucleation process of InP cores by raising the temperature at the late nucleation stage, thereby reducing and minimizing the formation of small-sized QDs. The higher temperature promotes diffusion-controlled-like growth by facilitating the conversion of amino-phosphine precursors. Besides, we also employ the multistage heating method in the shelling process to suppress the Zn ion exchange with the InP core, resulting in narrowed emission linewidths caused by exciton-phonon coupling. Overall, our multistage heating strategy for core nucleation and shell growth effectively optimizes the size distribution and reduces exciton-phonon coupling. Compared to classical synthesis, the multistage heating strategy can reduce the emission linewidth of nucleation and shelling by 13.2% and 30.9% respectively. Our champion InP/ZnSe/ZnS QDs exhibit narrow spectral linewidths of 41.5 nm (PL peak: 630 nm) and high PLQYs of 88%. This study presents a viable approach for achieving narrow spectral linewidths in InP QDs and offers valuable insights for other semiconductor nanocrystals requiring narrow emission linewidths.

2. Experiment section

Chemicals: Indium chloride (InCl₃, 99.99%), Zinc chloride (ZnCl₂, 99.99%) were purchased from Aladdin. Selenium powder (Se, 200 mesh, 99.999%), Sulfur powder (S, 99.998%), (diethylamino)phosphine ((DMA)₃P, 97%), 1-octadecene (ODE, 90%), Oleic acid (OA, 90%) and Oleyl-amine (OAm, 98%) were purchased from Sigma-Aldrich. Tri-octyl-phosphine (90%, TOP) and Vaseline was purchased from Macklin. All organic solvents were purchased from Sinopharm Reagents. ODE, OAm and Vaseline are degassed at 120 °C for 2 h before use.

Synthesis of InP/ZnSe/ZnS QDs: 0.4 mmol InCl₃, 5 mmol ZnCl₂ and 8 ml OAm were placed in a 100 ml three-neck flask. The temperature is raised to 130°C, and after 40 minutes of degassing, it is filled with Ar gas. Repeat this process 3 times. And then, the temperature was raised to 185°C In Ar environment, and 1.2 ml of the P precursor was injected (1 ml of (DMA)₃P mixed with 2 ml of OAm). It is cooled to 175°C to keep warm. After holding for 20 minutes, raise the temperature to 185-195-205°C and maintain the holding time at each temperature for 20 minutes. The temperature was raised to 240°C and the Se precursor was dropwised at a rate of 3 ml/h (4 mmol Se dispersed into 0.5 ml TOP, 4.5 ml ODE and 5 ml Vaseline). After 30 minutes, the temperature was then raised to 260°C and kept for 30 minutes and 2 ml of the Se precursor was added. Further increase the temperature to 280°C, add 3 ml of the Se precursor, and then add 3 ml of the Zn precursor to the solution (dissolve 10 mmol of zinc acetate in 20 mmol oleic acid and 15 ml ODE, degassing and heating to 280°C until completely dissolved). Keep the temperature at 280°C for 45 minutes. Then the reaction temperature was raised to 320°C and the ZnSe shell continued to grow for 60 minutes. The S precursor is then added at a rate of 2 ml/h (4 mmol of S powder dissolved in 2 ml TOP). After 20 minutes, cool down to 300°C and continue to add 1 ml S precursor. Reduce the reaction temperature to room temperature after 20 minutes.

Measurements: The absorption and PL were collected on Horiba Duetta. Transmission electron microscopy (TEM) and High-resolution transmission electron microscopy (HRTEM) images were tested by Thermo Scientific Talos F200X G2. X-ray diffraction (XRD) patterns were recorded on a Bruker D8 Venture X-ray diffractometer operated with Cu Kα radiation. Horiba Fluoromax spectrofluorometer with an integrating sphere (Quanta-φ) was used to measure the PLQYs. The temperature resolved PL spectra were recorded on Edinburgh FLS980 of Edinburgh Instruments. The Raman spectra were measured by HORIBA HR Evolution.

3. Results and discussion

Since 1950, LaMer et al. have proposed a model for particle formation that provides unique insights into the formation of monodisperse particles [31–33]. Over the years, both theoretically and experimentally, it has been proven that the explosively nucleated synthesis of particles can effectively achieve monodispersity through rapid accumulation of monomer concentration [32,34,35]. As illustrated in Fig. 1, during Stage I, the precursor is introduced into the reaction system and converted into monomers. When the monomer concentration surpasses its saturation level, it enters Stage II—nucleation stage. After significant consumption of monomers through nucleation process, the concentration becomes insufficient to sustain nucleation, leading to Stage III—growth step of cores. Depending on the abundance of monomers in system, size distribution focusing or maturation processes occur.

 

Fig. 1. The LaMer model of nanocrystal nucleation and growth.

Download Full Size | PDF

However, the synthesis of InP using amino-phosphine does not meet the ideal nucleation conditions envisioned by the LaMer model. After the addition of amino-phosphine, it undergoes a stepwise amination and disproportionation process before generating the desired valence state of phosphorus, which then combines with indium to form the monomers [18,20,21]. The entire process is severely limited by the reaction rate, and even after injecting the phosphorus source and maintaining the reaction temperature for 60 minutes, unreacted amino-phosphine still remains [20,21]. Therefore, the synthesis of InP based on amino-phosphine shows a continuous nucleation process, and the nucleation interval can even last up to 30 minutes. Although large-sized QDs exhibit significantly slower growth rates compared to small-sized ones due to Superfocusing effects, their size distribution is not as poor as initially expected [36]. However, the results are still unsatisfactory. Based on our understanding of the characteristics of the amino-phosphine precursor nucleation reaction, we propose employing a multistage heating nucleation method to reduce both time and spatial dimensions associated with this process while promoting core size-focusing effects. Indeed, increasing temperature facilitates the evolution of the current equilibrium system to a new balance of larger average size. Among them, the growth rate of small-size QDs is much higher than that of the large one, so that their size can “catch up” with the large-size counterparts in the system and improve the size uniformity. In addition, the elevated temperature also accelerates the monomer conversion process in phosphine precursors, leading to diffusion-limited growth tendencies. This phenomenon contributes positively towards reducing size distribution.

Choosing the appropriate temperature to match the nucleation and growth process of QDs is crucial achieving uniform size, as different nucleation reaction temperatures are associated with the conversion rate of P precursors, monomers supersaturation and critical size. Fig. 2 illustrates the time-dependent evolution of the absorption spectra for InP cores prepared at various reaction temperatures. The absorption peak of InP undergoes a redshift with increasing holding time, indicating that larger InP QDs can be prepared at higher reaction temperatures. Due to the high density of defect states on the InP surface, the fluorescence emission peak of the bare Core of InP is weak. Therefore, analysis of size distribution in InP nuclei is performed using absorption spectra. The FWHM of the absorption peak provides insights into the size distribution of QDs to some extent. To calculate FWHM for comparison purposes, we employ both peak-to-valley ratio (P/V) and derivative methods based on absorption spectrum data. By identifying the second derivative's peak position in the absorption spectrum, we confirm the first exciton absorption peak's location and fit its corresponding FWHM value (i.e., FWHM_2nd). Additionally, Gaussian function parameters derived from two known extrema can estimate another measure of width at half maximum (i.e., FWHM_1st). It should be noted that slight numerical differences may arise due to variations in calculation methods employed. It is interesting to observe inconsistent patterns between FWHM_1st and FWHM_2nd values obtained at different temperatures. Specifically, according to FWHM_1st results, QDs prepared at 175°C exhibit narrower peak widths compared to other samples tested; conversely, analysis based on FWHM_2nd indicates that samples synthesized at 215°C possess narrower peak widths instead. The results of FWHM_2nd show that the sample at this discrepancy reflects the disparity in the nucleation mechanism between aminophosphine-based InP and the conventional mode of nucleation. The calculation of FWHM_2nd does not take into account the contribution of small core size observed in the absorption spectrum, owing to its continuous nucleation process lasting up to thirty minutes. Moreover, the variation trend of P/V is akin to that of FWHM_1st, indicating a certain degree of overall dimensional uniformity in cores. It can be observed that QDs synthesized at 175°C exhibit lower FWHM_1st, making it a more suitable nucleation temperature choice. Conversely, lower temperatures result in significantly reduced conversion rates for the P source, leading to an elongated nucleation window. Higher temperatures cause faster consumption of the P source and deteriorate size distribution due to maturation towards the end of holding time. At 175°C, growth occurs at a relatively optimal rate.

 

Fig. 2. (a) the absorption spectra of InP cores at different reaction temperatures. (b) variation curves of exciton absorption peaks with reaction time at different temperatures. (c) the peak/valley ratio at different temperature. (d) and (e) are the FWHM calculated from the first/second derivative of the absorption spectrum (FWHM_1st and FWHM_2nd) at different temperature. The insert diagram is the calculation method.

Download Full Size | PDF

The absorption spectra of InP synthesized under different temperatures (Fig. 3) clearly demonstrate the differences between the conventional method and the multistage heating method. In the traditional method, irrespective of the reaction temperature, there is minimal variation in the P/V ratio within the absorption spectra throughout the reaction process, and no distinct absorption peaks are observed. In contrast, with regards to the multistage heating technique, a pronounced excitonic absorption peak becomes evident following an increase in temperature. This signifies a substantial improvement in the size distribution of the InP core at this stage. Such improvement can be attributed to continuous nucleation characteristic of traditional synthesis methods that consistently yield a fraction of relatively smaller-sized QDs. By simply controlling temperature, it is possible to mitigate formation of these smaller-sized QDs. This offers two advantages. Firstly, reduction in smaller-sized QDs itself contributes to narrowing the size distribution. Secondly, monomers saved from this fraction can be utilized for targeted growth.

 

Fig. 3. (a) the schematic illustration of different synthetic routes. (b) the absorption spectra of InP cores with multistage heating. (c) the variation curves of exciton absorption peaks with multistage heating. (d) the peak/valley ratio curves of different heating moment. The moment when the temperature starts to heat up is recorded as 0 min. The blue line is the reference curve for the reaction at 175°C.

Download Full Size | PDF

Figure 3(b, c) and (d) show the absorption of the InP cores prepared using multistage heating method. Based on previous discussion, we chose 175°C as the initial nucleation temperature. After 20 min incubation, the temperature was increased in steps of 10°C, with 20 min incubation at each temperature point. With the temperature increase, accompanied by a faster redshift, the absorption peaks became more pronounced, indicating significant size-focused growth. Of course, after heating up, large-size QDs will grow further. But the growth rate of small-size QDs is still much higher than that of large-size QDs, so the monomers that should generate small-size cores are mostly consumed in size focused growth after heating up, thus promoting the size uniformity of InP cores. Subsequently, we investigated the applicability of the remaining incubation time at the initial temperature of 175°C using the multistage heating strategy, and labeled these samples as C₅, C₂₀, C₄₀ and C₆₀. As the reaction temperature increased, the absorption peaks showed a redshift, and with longer incubation times, the shift gradually decreased. When the incubation time was 5 min, the absorption peak exhibited a larger and faster redshift due to the sufficient precursors in the system. In contrast, only a slight redshift was observed in C₆₀, indicating a lower abundance of precursors in the solution. Moreover, the effect of the multistage heating strategy was closely related to the incubation time, as indicated by the changes in P/V ratio. As shown in Figure 3(d), the P/V values of C₅ and C₆₀ were not good, indicating that both too short and too long incubation times failed to achieve the best size-focused growth effect. The broadening of size distribution is inevitable if the limited monomers in the solution cannot be effectively utilized for size-focused growth. After prolonged incubation, there were very few remaining monomers available for growth, resulting in minimal effectiveness of the multistage heating strategy. On the other hand, if the temperature is increased during the initial incubation stage, although it promotes size-focused growth, the faster monomer conversion rate significantly increases the supersaturation of monomers in the solution, thereby facilitating nucleation. Therefore, the P/V ratio of C₅ is inferior to that of C₂₀.The presence of a distinct excitonic absorption peak in the multistage heating samples implies a more uniform and narrow size distribution of the InP QDs. This suggests that the multistage heating method enables enhanced control over the nucleation and growth processes, leading to increased homogeneity in the size of the InP cores.

Considering the lattice compatibility between the core-shell materials and the limiting ability of the exciton wave function, the ZnSe and ZnS layers were selected to grow sequentially outside the InP core, as shown in Fig. 4. Due to the dangling bonds on the bare core surface, the synthesized InP has low radiate fluorescence with PLQY of about 0.1%. To modify defect states on InP surface, eight layers (Fig. S1&S2, about 2.8 nm) of ZnSe were grown using a multistage heating method resulting in continuous redshifts in PL peak and significant improvements in PLQYs. Generally, higher-temperature growth leads to higher-quality shell layers and better compression of the fluorescent linewidth. However, considering the ion exchange effect of Zn, the sample S_MH synthesized using our multistage heating method (240°C -260°C -280°C -320°C) exhibited superior performance. Firstly, approximately 1-2 layers of ZnSe were grown at 240°C, reducing the diffusion of Zn ions through a lower reaction temperature, followed by gradual heating to grow higher-quality ZnSe layers. Since the CB of InP is lower than that of ZnSe, InP/ZnSe forms a quasi-type II band structure, which is conducive to confining electrons and holes in the core to form radiative recombination. As shell thickness increases, the FWHM decreases to 43 nm. To further enhance the stability and PLQYs of the QDs, ZnS layers were grown on the outside of the InP/ZnSe structure. After growing approximately four layers of ZnS (Fig. S2, about 1.2 nm), the PL peak redshifted to 630 nm, and the linewidth further decreased to 41.5 nm, resulting in an improved PLQY of approximately 85%. The average linewidth of the QDs prepared under the combinations of different methods are shown in Figure S3. Furthermore, XRD analysis revealed that the peak positions of the InP QDs matched the PDF reference card and exhibited good crystallinity. After the growth of ZnSe and ZnS on the outer layers, the diffraction peaks shifted towards larger angles, indicating lattice strain on the InP core after the growth of ZnSe and ZnS with smaller lattice constants.

 

Fig. 4. (a) Scheme of the synthesis of highly luminescent InP/ZnSe/ZnS QDs. (b) Evolution of absorption spectra (dotted line) and PL spectra (solid line) during QDs synthesis. The inset shows photographs of the QDs under UV lamp (left) and ambient light (right). (c) The PL and FWHM as functions of the monolayers of the shells. (d) The XRD patterns of QDs for different shell layers.

Download Full Size | PDF

We effectively reduced the size distribution of the core using a multistage heating nucleation method, thereby compressing the inhomogeneous broadening of the fluorescence emission. To minimize the linewidth of the InP QDs, we need to consider the mechanism of linewidth broadening and minimize the interactions between excitons and phonons. Considering the adverse effects of Zn ion exchange on the emission linewidth of InP observed in many previous studies, this research intends to investigate the impact of Zn ion diffusion in InP core on exciton-phonon interactions. Temperature plays a crucial role in Zn ion exchange, and thus, we prepared different samples by controlling the doping amount of Zn ions through varying the reaction temperature during the growth of ZnSe. We grew ZnSe at 240°C, 260°C, and 280°C, denoted as S₂₄₀, S₂₆₀, and S₂₈₀, respectively. The fluorescence peaks of these samples were in the range of 629-633 nm, with linewidths ranging from 132-147 meV. Subsequently, we investigated the exciton-phonon coupling strength through temperature-dependent photoluminescence (PL) spectroscopy, as shown in Fig. 5 and Fig. 6, covering a temperature range from 80 K to 298 K. With increasing temperature, the fluorescence emission peak continuously redshifted, and the linewidth broadened. QDs with higher levels of Zn doping showed larger shifts in emission wavelength.

 

Fig. 5. The temperature dependence of steady-state PL. (a-d) The PL evolution at different temperatures. The inset is the normalized PL plot.

Download Full Size | PDF

 

Fig. 6. The temperature dependence of steady-state PL. (a-d) The fitting curves of the emission peak and its FWHM at different temperatures.

Download Full Size | PDF

With the temperature increased from 80 K to 298 K, the PL intensity decreases continuously, the peak wavelength of S_MH redshifts from 618 nm to 630 nm, and the emission FWHM also widens from 114 meV to 131 meV. The contribution of each part of linewidth is studied by fitting the following Eq. (1) [37–40]:

Γ_inh is the inhomogeneous broadening term (temperature- independent); Γ_ac and Γ_LO are the homogeneous broadening terms, which are caused by acoustic and LO phonon interaction. γ_ac and γ_LO are the exciton-phonon coupling strengths of the acoustic and LO phonon. k_B is the Boltzmann constant and E_LO is the LO phonon energy.

To facilitate fitting, the LO phonon energy of InP/ZnSe/ZnS QDs was calculated to 43.8 meV through Raman curves (Fig. S4). The fitting results obtained using Eq. (1) are shown in Table 1. The inhomogeneous broadening terms of the QDs synthesized at different temperatures were relatively close. The value of S_MH was the lowest, indicating that this multistage heating shell growth strategy effectively reduced the differences between the QDs. Among them, the inhomogeneous broadening of S₂₄₀ was ∼11% larger than the other samples, indicating that the reaction at only 240°C was insufficient to grow high-quality shells. In addition, the coupling between excitons and acoustic phonons in these samples exhibited a noticeable pattern. However, the coupling between excitons and LO phonons showed a clear correlation with the reaction temperature. S₂₄₀ had the lowest value of γ_LO, indicating a lower exciton-LO phonon coupling strength, and as the shell growth temperature increased, γ_LO also increased. This suggests that the entry of Zn ions from ZnSe into the InP core during the growth process significantly enhances the exciton-LO phonon coupling, thus increasing the homogeneous broadening of the emission linewidth. When Zn ions enter the InP cores as divalent cations, they become P-type impurities and produce defects in InP, which disturbs the electronic state and enhances the interaction between excitons and phonons. The γ_LO of S₂₈₀ increased by 59% compared to S₂₄₀ and 49.6% compared to S_MH. Although the γ_LO of S_MH was larger than S₂₄₀, the difference between them was small, indicating that the multistage heating shell growth method effectively suppresses the broadening of the InP linewidth caused by Zn²⁺ exchange.

Table 1. Results of Phonon Broadening Model Fits by Eq. 1

View Table

4. Conclusion

In summary, we employed a multistage heating strategy to synthesize high-quality amino-phosphine-based InP/ZnSe/ZnS QDs, facilitating size-focused growth and effectively suppressing exciton-phonon interactions. By controlling the temperature based on continuous nucleation characteristics, we truncated the nucleation window and directed monomers towards the growth of small-sized QDs rather than continuous generation of small-sized cores, thereby reducing inhomogeneous broadening (decreased by 13.2%). Subsequent low-temperature ZnSe layer growth effectively minimized diffusion of Zn ions into InP cores, followed by high-temperature growth to ensure a high-quality shell formation. We observed a positive correlation between exciton-LO phonon coupling and Zn²⁺ content in InP cores. Compared to control samples synthesized at high temperature, the LO phonon broadening of our optimized InP/ZnSe/ZnS QDs is reduced by 49.6%, with the total homogeneous broadening of 22.6 meV (decreased by 30.9%) at room temperature. Furthermore, these optimized QDs achieved narrow red (630 nm) emission linewidths of 41.5 nm (129.6 meV). This method also provides valuable insights for synthesizing other nanocrystals with similar reaction-limited growth systems.