Full-color tuning in multi-layer core-shell nanoparticles from single-wavelength excitation

Introduction

Lanthanide-based materials are a class of ideal candidates for efficient nonlinear anti-Stokes visible photon upconversion under infrared excitations^1,2,3. Lanthanide ions have abundant discrete energy levels that allow for multiple emission colors in the visible spectral region^4,5,6,7, showing great promise in diversities of frontier applications including laser^8,9, volumetric display^10,11, super-resolution nanoscopy^12,13,14, and biophotonics^15,16,17. However, conventional upconversion and luminescent materials usually exhibit a specific emission color when the dopants and compositions are determined. Recently, an interesting orthogonal excitation–emission phenomenon was observed in multi-layer core-shell (MLCS) nanoparticles, featuring a change in the emission colors by simply switching excitation wavelengths¹⁸. For instance, a spatial separation of Er³⁺ and Tm³⁺ into suitable layers in an MLCS nanostructure results in green light emission (from Er³⁺) under 980 nm irradiation and blue light emission (from Tm³⁺) under 808 nm irradiation^19,20,21,22. As a result, the red-green-blue (RGB) colors become obtainable by integrating more luminescent layers in an MLCS nanostructure or by employing non-steady state upconversion^10,23,24,25. Such extension in optical properties of luminescent materials further promotes their frontier applications in information security^26,27, optical memory²⁸, super-resolution nanoscopy²⁹, photodynamic therapy³⁰, and optogenetics^31,32.

However, it should be noted that dual or more excitation wavelengths are necessary for the color tuning of MLCS nanoparticles, resulting in a complex design of the pumping system. Also, the MLCS structure together with its compositions and dopant species have to be exquisitely designed to avoid any spectral cross-talk between different luminescent shell layers, which raises stringent demands for the sample structure and experimental synthesis. These intrinsic constraints seriously limit the design and application of luminescent materials with a full-color switchable output by using only a single-wavelength excitation.

Recent work suggests that temporal control of upconversion dynamics provides a new possibility to tune and edit emission colors³. As a typical nonlinear anti-Stokes process^3,33, it requires time to populate the emitting energy level of a lanthanide emitter from its lower-lying intermediate state, giving rise to a characteristic rise time for the resultant emission (Fig. 1a). This provides a new parameter, apart from the decay time, to describe the temporal evolution process of a photon upconversion by comparison to conventional Stokes luminescence. This feature means that a color change would be available by modulating the upconversion rise times for specific lanthanide emitters that have multiple emission peaks (Fig. 1b). On the other hand, for a given emitting state, it has an intrinsic decay time under infrared excitation. When multiple emission profiles are designed in one upconversion system, the emissions with longer decay times can be singled out by filtering out those shorter decay emissions through the time-gating technique²². Therefore, a rational combination of modulating approaches on both the initial rise and final decay processes of upconversion luminescence could be greatly capable of producing full-color switchable output from a single nanoparticle under a single-wavelength excitation.

**Fig. 1: Proposed concept for full-color tuning with a single-wavelength excitation.**

Here, we describe a conceptual model to realize the full-color tuning in a single MLCS nanoparticle upon a commercial 980 nm laser (Fig. 1c). The conventional emitters of Er³⁺ (green and red emissions) and Tm³⁺ (blue emission) were selected and spatially doped into different nanoscale regions toward its specific emission colors. We find that Er³⁺ in shell can produce red-to-green color-switchable emissions by tuning its upconversion dynamics through the cooperative modulation effect of the Tm/Yb couple. The blue light of Tm³⁺ in core can be singled out by applying the time-gating technique which removes the short-decayed green and red emissions. More importantly, it is capable of simplifying the pump systems greatly by only using one single-wavelength laser excitation instead of two or more, as shown in the previous reports. We further demonstrate the new role of Tm³⁺ in manipulating upconversion dynamics in addition to the typical role of a blue-emitter, which plays a key role in enabling the temporal control of emission colors of Er³⁺. These findings help deepen the understanding of the photophysics of lanthanide ions, showing great promise in the frontier photonic applications.

Results

To validate our hypothesis, we first explore how to realize the temporal control of red/green color-switchable upconversion of Er³⁺. In general, its emission profile consists of two adjacent green emission bands (²H_11/2,⁴S_3/2 → ⁴I_15/2 transitions) and a red emission band (⁴F_9/2 → ⁴I_15/2 transition) under the commonly used 980 nm excitation^34,35,36. Note that the conventional Er³⁺/Yb³⁺ codoped materials usually show a fixed emission color, such as green for the regular NaYF₄:Yb/Er@NaYF₄ core-shell nanoparticles under both continuous-wave (c.w.) and pulse 980 nm laser excitations (Supplementary Fig. 1a). However, it is interesting to observe a faster rise time for the green emission than that of the red emission (Supplementary Fig. 1b). This suggests that temporal control of emission colors would be highly desirable by pre-editing the steady-state upconversion emission profiles of Er³⁺. Here we introduced a small amount of Tm³⁺ in the Er/Yb couple to mediate the upconversion dynamics of Er³⁺. It can promote the population of Er³⁺ from its ⁴I_11/2 state to the lower-lying intermediate ⁴I_13/2 state through an energy-transfer looping (ETL) process, namely Er³⁺ (⁴I_11/2) → Tm³⁺ (³H₅) → Er³⁺ (⁴I_13/2), consequently leading to a red color output under c.w. 980 nm excitation (Fig. 2a; Supplementary Fig. 2).

**Fig. 2: Red/green color-switchable upconversion in Er/Tm/Yb tri-doped nanoparticles.**

A series of NaYF₄:Er/Tm/Yb@NaYF₄ core-shell nanoparticles were synthesized by a modified co-precipitation method (Supplementary Note 1)²⁶, showing sphere-like morphology and hexagonal phase (Supplementary Fig. 3). The inert NaYF₄ shell layer was used to prevent the lanthanide emitters in core from being quenched by surface quenchers. The upconversion emission colors are closely dependent on the concentrations of each lanthanide dopant (Fig. 2a–c), and the doping of 15Er:0.5Tm:40Yb (mol%) results in an emission color change by only modulating the pulse width of 980 nm excitation laser (Fig. 2d, e; Supplementary Figs. 4–10 and Supplementary Tables 1–4). Namely, the sample shows red color under c.w. or long pulse (e.g., >8 ms for pulse width) 980 nm excitation; while it starts to present a gradual emission color change from red to orange, yellow and green with reducing the excitation pulse width down to 0.05 ms. As a control, other excitation wavelengths (e.g., 808 or 1530 nm) do not lead to an emission color change (Supplementary Fig. 11a, b). The recently reported NaErF₄@NaYF₄ core-shell nanoparticles only show a red color under both c.w. and pulse 980 nm excitation (Supplementary Fig. 11c, d)^37,38,39. In addition, the steady-state red upconversion of the Er/Tm/Yb tri-doped sample is much greater than that of the Yb/Ho/Ce tri-doped system (Supplementary Fig. 12). These observations confirm the feasibility of achieving emission color tuning of Er³⁺ through temporal control of its upconversion dynamics under conventional 980 nm excitation, laying a solid foundation for the following full-color tuning with a single wavelength excitation.

To shed more light on the upconversion mechanism of Er³⁺ in the Er/Tm/Yb tri-doped system, we measured the time-dependent upconversion emission profiles of the samples. As shown in Fig. 2f, it is interesting to observe a faster rise time for the green emission than the red emission during the excitation pulse width. This feature ensures the observation of green emission color by simply reducing the excitation laser pulse width. The faster rise time of green emission than that of red emission can originate from the difference in their upconversion processes. The green upconversion is based on the resonant energy transfer from Yb³⁺ (at its ²F_5/2 state) to Er³⁺ (at its intermediate ⁴I_11/2 state) followed by an up-transition process³. In contrast, for the red upconversion its intermediate ⁴I_13/2 state has to be pre-populated through a Tm³⁺-mediated ETL process and/or multi-phonon relaxation (MPR) from ⁴I_11/2 to ⁴I_13/2 (Supplementary Fig. 2b). Such additional processes result in a temporal delay in the up-transition of Er³⁺ to the red emitting ⁴F_9/2 state. Although the presence of Tm³⁺ accelerates the rise time of red emission, it is still slower than that of the green emission (Fig. 2f; Supplementary Fig. 13).

Another interesting observation in Fig. 2f is the abrupt drop in the time-dependent emission profile of Er³⁺ when the pulse duration is over ~0.5 ms. This suggests that some relatively slow depleting processes begin to occur at the green emitting levels. To shed more light on this issue, we prepared a set of core–shell control samples with single doping of Er (15 mol%), and codoping of Er/Tm (15/0.5 mol%) and Er/Yb (15/40 mol%) in the core (Supplementary Fig. 14). Their time-dependent emission profiles were compared in Fig. 2g (top panel), showing no abrupt drop for all of them. Thus the abrupt decline of Er³⁺ should result from the cooperative modulation effect due to the presence of both Tm³⁺ and Yb³⁺, which work together to make a dynamic manipulation of Er³⁺ upconversion. Importantly, it was found that the lifetime of green emission exhibits a decrease for the Er-doped control samples after the presence of Tm³⁺ or Tm/Yb couple (Fig. 2h). The abrupt decline is presumably attributed to the cross-relaxation (CR) between Er³⁺ at its green emitting ²H_11/2,⁴S_3/2 states and Tm³⁺ at its intermediate ³F₄ state instead of the ground state ³H₆. Taking into account a small energy mismatch (~1600 cm⁻¹) ³, the energy transfer from Yb³⁺ (at its ²F_5/2 state) to Tm³⁺ (at its ³H₅ state) needs more time than that to Er³⁺ (at its ⁴I_11/2 state), which is a resonant process. As a result, the CR occurring at the ³F₄ state of Tm³⁺, populated by a rapid MPR process from the upper-lying ³H₅ state, would have a slower temporal feature. In this case, the upconversion emission based on the intermediate ³F₄ state of Tm³⁺ should also have a slower rise time than that of Er³⁺ (Supplementary Fig. 15). This was confirmed by the time-dependent upconversion emission profiles of Tm³⁺ (e.g., 695 and 800 nm emissions) upon 980 nm excitation (Fig. 2g bottom panel). For the Er/Tm (15/0.5 mol%) coupled control sample, much less Tm³⁺ ions populate at its ³F₄ state, which cannot support efficient CR processes, and there is no change in the emission color with tuning the pulse width of excitation laser (Supplementary Fig. 7a). While after the presence of Yb³⁺, much more Tm³⁺ ions were pumped in its ³F₄ state followed by the energy transfer from Yb³⁺ (²F_5/2) to Tm³⁺ (³H₅) and subsequent relaxation to its ³F₄ state, as evidenced by the remarkable enhancement of infrared emission of Tm³⁺ from its ³F₄ → ³H₆ transition (Fig. 2i). In contrast, no obvious abrupt decline was observed in the Er/Yb codoped control sample, implying that the CR between Er³⁺ and Yb³⁺ has no notable contribution to the color change despite a decline in the lifetime of green emission (Supplementary Fig. 16)^40,41,42. The observation of temporal abruptness was also in agreement with the simulation through rate equations via an upconversion model (Supplementary Fig. 17; Supplementary Note 2). Thus, we have demonstrated the key role of the Tm^/Yb couple in manipulating both the rise and decay dynamics of Er³⁺, and the involved processes were schematically illustrated in Fig. 2j.

We next investigate how to realize the long-decay blue emission light together with a slower rise time than that of green and red emissions of Er³⁺. This is a prerequisite to single out the blue emission from the total upconversion emission profiles through the time-gating technique. Among a series of lanthanide ions, Tm³⁺ is a possible candidate for blue emission light by taking advantage of its ¹G₄ → ³H₆ transition (peaking at around 475 nm), which is widely observable in Yb³⁺/Tm³⁺ codoped systems upon 980 nm excitation (Fig. 3a) ⁴. However, a crucial issue for Tm³⁺ lies in the serious concentration quenching effect when it is doped into host with higher concentrations due to the multiple CR processes^43,44,45, such as [¹G₄; ³H₆] → [³H₄; ³H₅], which would result in a rapid decline in the lifetime of Tm³⁺ at its ¹G₄ state in addition to decreased blue emission intensity (Fig. 3b, c). Other emission levels of Tm³⁺ show a similar result (Supplementary Fig. 18a). To get rid of this limit, we attempt to enlarge the ionic separation by employing a smaller Tm³⁺ dopant content, and the CR processes occurring in the blue emitting level were well suppressed, resulting in markedly prolonged lifetime (Fig. 3c). Also, it can be observed that a decrease of Tm³⁺ content helps to slow the rise time of blue emission (Fig. 3d), might be resulted from a greater Tm³⁺-Yb³⁺ ionic separation. This is a key point to remove the interference of blue emission on other emissions under short pulse 980 nm excitation. Relatively low doping of Yb³⁺ content is also helpful for intense blue emission with slower rise time and longer decay time (Fig. 3e; Supplementary Fig. 18b). Thus, the Yb/Tm couple with the concentration ratio of 20/0.1 mol% was selected for the on-demand blue emission light.

**Fig. 3: Tuning rise and decay times of blue upconversion of Tm³⁺.**

To achieve the single-wavelength responsive full-color output in a single nanoparticle, we construct an MLCS nanostructure to integrate each functional unit. In detail, the Yb/Tm couple was codoped into the core and Er/Tm/Yb into the outer luminescent shell layer, both of which were spatially separated by growing an optically inert NaYF₄ epitaxial interlayer aiming to avoid any spectral cross-talk between them²⁶, and finally another NaYF₄ shell was coated outside to prevent surface quenching effect⁴⁶. In such a design, a newly emerging issue encountered is that the blue emission from the core was concurrently activated along with the red emission upon c.w. 980 nm excitation (Supplementary Fig. 19; Supplementary Table 5). This means that the content of 40 mol% Yb³⁺ in the luminescent shell layer is not sufficient enough to block the 980 excitation photons reaching the blue light emitting core^21,47. In this case, we attempted to use the Yb sublattice to raise Yb³⁺ content in the luminescent shell layer, some of which could act as the energy harvester to facilitate the trapping of incident 980 nm excitation photons and properly minimize the excitation energy to activate Tm³⁺ in the core (Fig. 4a, b; Supplementary Fig. 20), resulting in observation of red color output under c.w. 980 nm excitation. Also, it is still capable of enabling green emission light under short pulse 980 nm excitation due to the faster rise time of the green emission. More importantly, the lifetime of Tm³⁺ emission is much longer than that of Er³⁺ (Fig. 4c), which allows for the blue emission light output through the time-gating technique to filter out those short-decay emissions (Supplementary Note 3), resulting in RGB color-tunable emissions under suitable excitation modes (Fig. 4d). The RGB colors can be kept in a large range of pump power densities under 980 nm excitation (Supplementary Fig. 21; Supplementary Tables 6–8). We further examined the role of the inert NaYF₄ interlayer between the core and luminescent shell (Supplementary Fig. 22) and found that it is indispensable for the observation of the blue emission by suppressing the interfacial quenching effect (Supplementary Figs. 23–26). In comparison, the thickness of the outermost NaYF₄ shell has a slight effect on the color tuning besides a change in the luminescence intensity (Supplementary Figs. 27–31). So far, we have successfully realized the RGB colors output in an MLCS nanoparticle by simply tuning the excitation modes of a single 980 nm laser (Fig. 4e) for the first time to the best of our knowledge (Supplementary Table 9).

**Fig. 4: Single-wavelength responsive full-color tuning of upconversion and its potential application.**

The realization of primary RGB color-switchable emissions permits new chances for frontier photonic applications. As a proof of concept, we prepared a painting of a “butterfly-over-rose” pattern using such nanoparticles through the screen printing method (Fig. 4f top). Interestingly, it can precisely show a red rose flower under c.w. 980 nm irradiation, green leaves, and bud under short pulse 980 nm irradiation, and then blue butterfly under the time-gating mode (Fig. 4f; bottom right panel). In contrast, under regular c.w. 980 irradiation the entire painting only shows a single red color (Fig. 4f; bottom left panel). More colors can be easily achievable by fine-tuning the pulse widths of the incident laser (Fig. 4g). This may contribute to the full-color volumetric display under a single excitation wavelength and reduce the complexity of pump systems¹⁸. As an added benefit, such smart control of emission light colors also holds great promise in multi-level anti-counterfeiting²⁶. The concealed information can be clearly distinguished by applying a time-gating technique and 1530 nm irradiation (Fig. 4h; Supplementary Fig. 32).

Discussion

We have demonstrated a new conceptual model that is convenient to realize the full-color tuning from a single nanoparticle upon a single 980 nm wavelength excitation. The temporal control of up-transition and decay dynamics is confirmed to be a facile but effective strategy to achieve color-switchable output from conventional lanthanide emitters of Er³⁺ and Tm³⁺. Intriguingly, the new roles of lanthanide ions were discovered during the manipulation of luminescence dynamics and energy transport over the lanthanide sublattice, which helps break the stereotypes of each lanthanide ion. The results described in this work contribute to the design of full-color emissive nanoparticles for volumetric display, which would be further promoted by the new strategies of full-color tuning and device miniaturization in the future. Moreover, in light of versatile color-tunable properties, they would also help develop a new class of smart materials for frontier applications such as information security and storage, functional flexible devices, and wavelength-tunable lasers.

Methods

Materials

The materials including erbium (III) acetate hydrate (99.9%), ytterbium (III) acetate hydrate (99.99%), thulium (III) acetate hydrate (99.9%), yttrium (III) acetate hydrate (99.9%), holmium (III) acetate hydrate (99.9%), cerium (III) acetate hydrate (99.99%), oleic acid (90%), 1-octadecene (90%), sodium hydroxide (NaOH; >98%), and ammonium fluoride (NH₄F; >98%) were all purchased from Sigma-Aldrich and used as received unless otherwise noted.

Synthesis of nanoparticles

The core and core-shell nanoparticles were synthesized using a modified co-precipitation method. The MLCS nanoparticles were prepared by combining coprecipitation and thermal decomposition methods. The synthetic procedures for other control nanoparticles were identical to the above methods except for the use of different core seed nanoparticles and corresponding lanthanide shell precursors. The experimental details are provided in the Supplementary Information.

Characterization

The low- and high-resolution transmission electron microscopy measurements together with energy-dispersive X-ray spectroscopy were carried out on a JEOL JEM-2100F transmission electron microscope (200 kV). Powder X-ray diffraction was performed by a Philips Model PW1830 X-ray powder diffractometer with Cu Kα radiation (λ = 1.5406 Å). The upconversion emission and infrared emission spectra were recorded by a Zolix spectrofluorometer equipped with external power-controllable laser diodes of 980, 808, and 1530 nm. The decay curves were measured using the same spectrofluorometer through using the pulse laser as an excitation sources. The lifetime values were determined by fitting the decay curves by the equation: I = I₀exp(-t/τ), where I₀ is the initial emission intensity at t = 0 and τ is the lifetime. The upconverting emission photographs were taken by a digital camera with suitable optical filters. All the measurements were conducted at room temperature. The time-gating observation is obtained based on the setup in which a chopper and a pulse synchronizer were used to set the delay time, and the details are provided in the Supplementary Information.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.