NIR II-responsive photon upconversion through energy migration in an ytterbium sublattice

Abstract

Smart control of photon upconversion is a key strategy for lanthanide-based materials used in biological and photonic applications. However, this has remained a challenge for the upconversion luminescence of lanthanides under excitation in the second near-infrared (NIR II) biowindow instead of at the conventional 980 and 808 nm wavelengths. Here, we report a conceptual design for an energy-migratory ytterbium sublattice in an erbium-sensitized multilayer core–shell nanostructure that is able to achieve photon upconversion from a broad range of lanthanide ions (Yb³⁺, Tm³⁺, Ho³⁺, Gd³⁺, Eu³⁺ and Tb³⁺) under 1,530 nm irradiation. The quasi-single-band upconversion in the first near-infrared (NIR I) biowindow is also realized through fine manipulation of the introduced cross-relaxations. By establishing an interfacial energy-transfer-mediated nanostructure, we also gain a deep insight into the mechanistic features of the energy migration. These results open new opportunities in a variety of frontier applications, such as information security.

Main

Photon upconversion is a nonlinear anti-Stokes process featuring high-energy (short-wavelength) photon emission under low-energy (long-wavelength) photon irradiation¹. This phenomenon has been observed in a broad range of materials, from bulk to films, phosphors and nanomaterials. Lanthanide ions have been shown to be ideal candidates for photon upconversion due to their unique electronic configurations with abundant discrete energy levels^1,2,3,4. The emergence of lanthanide-doped nanoparticles has provided more opportunities related to their many biological applications^{5,6,7,8,9,10,11,12,13,14}, as well as other frontier fields such as three-dimensional (3D) volumetric displays¹⁵, super-resolution nanoscopy¹⁶, information storage and security^17,18 and upconversion lasers^19,20. Many mechanistic efforts have been made to manipulate photon upconversion, including a core–shell structure design for upconversion of lanthanides without intermediate states²¹, Nd³⁺-sensitized upconversion^6,22,23, colour-switchable upconversion^15,24,25 and a photon-avalanche-like process for amplified stimulated emission¹⁶. Recently, it has been shown that interfacial energy transfer (IET) is also an efficient way to tune upconversion and examine the interactions between lanthanides on a nanometre length scale^26,27.

Despite the great progress made in research on upconversion, previous works have been heavily dependent on excitation wavelengths in the short near-infrared region (for example, 980 and 808 nm). The longer infrared second near-infrared (NIR II) spectral region, particularly at around 1,530 nm, has attracted increasing interest because of its great promise in deep-tissue and high-resolution imaging^28,29. It is also associated with optical communications, where versatile lasers and optical components are commercially available^30,31. Research on NIR II-responsive photon upconversion would contribute markedly to fundamental research into luminescent materials as well as their biological and photonic applications. However, most lanthanide ions are non-responsive to the 1,530 nm wavelength due to the mismatch of metastable energy levels and a lack of feasible pump configurations¹. Although Er³⁺ has a good spectral response around 1,490–1,550 nm (refs. ^32,33,34), it cannot function well as the NIR II sensitizer in the commonly used sensitizer–activator co-doping scheme because of deleterious quenching interactions³⁵. It has thus remained a challenge to achieve smart control of photon upconversion from a series of lanthanide ions upon NIR II excitation.

Recently, Yb³⁺ has demonstrated good energy-migratory performance to accompany its usual role as a sensitizer in conventional upconversion materials (for example, in an Yb–Er coupled system)^6,22. The use of such energy migration can effectively enhance Nd³⁺-sensitized upconversion and mediate upconversion and downshifting luminescence dynamics with tuning of the rise and decay times^7,36,37. Therefore, a combination of sensitizing Er³⁺ and a migratory Yb sublattice in a well-designed nanostructure could offer new opportunities for the development of efficient and versatile upconversion materials.

Here, we describe a novel mechanistic strategy to realize NIR II-responsive upconversion from a series of lanthanide ions. As illustrated in Fig. 1a, the upconversion scheme was established by constructing an energy-migratory Yb sublattice in an erbium-sensitized nanostructure in which the sensitizer and emitter are spatially separated. We show that the excitation energy can be efficiently transported from sensitizer to emitter through the Yb³⁺-mediated energy-migration channel. In this design, Er³⁺ can work well as a sensitizer to harvest the 1,530 nm irradiation energy via its ⁴I_13/2 ← ⁴I_15/2 absorption transition. Notably, the good spectral overlap between upconverted Er³⁺ emission (⁴I_11/2 → ⁴I_15/2 transition) and Yb³⁺ absorption (²F_5/2 ← ²F_7/2 transition) at around 977 nm greatly boosts resonant energy transfer from Er³⁺ to Yb³⁺ (Fig. 1b,c). Another key feature of the migratory Yb sublattice is that it can effectively isolate any unwanted deleterious interactions between sensitizer and emitter that could quench the upconversion luminescence. As a result, this design makes it possible to achieve photon upconversion from a broad range of lanthanide ions under 1,530 nm excitation (Fig. 1d). Moreover, it inspires us to revisit and break stereotypes about the conventional role of each lanthanide ion in upconversion systems towards versatile control of luminescence dynamics. Because of the NIR II excitation wavelength, this conceptual design is also able to reach a much larger anti-Stokes shift compared to upconversion systems at short near-infrared excitations.

Fig. 1: NIR II-responsive photon upconversion through energy migration in an Yb sublattice.

a, Schematic of the conceptual upconversion model, achieved by constructing a migratory Yb sublattice in an erbium-sensitized nanostructure, and the energy transport process involved. ET, energy transfer; EM, energy migration; Ex., excitation; Em., emission. b, Spectral overlap between Er³⁺ emission and Yb³⁺ absorption allows an efficient energy transfer from Er³⁺ at its ⁴I_11/2 level to Yb³⁺ at its ²F_5/2 level. c, A comparison of near-infrared upconversion emission profiles from NaErF₄:Yb(10 mol%)@NaYF₄ and NaErF₄@NaYF₄ core–shell nanoparticles under 1,530 nm excitation. d, A summary of the upconversion transitions of lanthanide ions that are available under 1,530 nm excitation.

As a proof of concept, we designed a set of multilayer core–shell nanostructures, separately incorporating each lanthanide dopant (Fig. 2a). The possibility of upconversion from Tm³⁺ at 1,530 nm excitation was first examined by preparing NaErF₄:Yb/Y@NaYbF₄@NaYF₄:Yb/Tm@NaYF₄ multilayer core–shell nanoparticles (Type-I design in Fig. 2a) using a modified co-precipitation method (Supplementary Methods)²⁶. These nanoparticles exhibit a monodisperse characteristic with an average size of 57.3 nm (top panel of Fig. 2b, Supplementary Fig. 1 and Supplementary Table 1) and are in the hexagonal phase according to their Fourier-transform diffraction pattern (right bottom panel of Fig. 2b) and X-ray diffraction results (Fig. 2c and Supplementary Fig. 2a). The observation of clear lattice fringes with a d-spacing of 0.52 nm in Fig. 2b (left bottom panel) reveals its good crystalline property. Besides an increment in size (Supplementary Fig. 2b), the multilayer core–shell nanostructure is also evidenced by the distributions of each lanthanide element (Fig. 2d).

Fig. 2: Sample designs, characterization and upconversion emissions.

a, Schematic of three structural designs of the samples (Types I–III) and the energy transport processes involved. b, Transmission electron microscopy (TEM) image (top), high-resolution TEM image (bottom left) and Fourier-transform diffraction pattern (bottom right) of the as-synthesized NaErF₄:Yb/Y(10/40 mol%)@NaYbF₄@NaYF₄:Yb/Tm(50/1 mol%)@NaYF₄ multilayer core–shell nanoparticles. c, X-ray diffraction pattern of the sample from b. Data from hexagonal NaYF₄ is also plotted (PDF#16-0334). d, Element mappings of Er, Yb, Tm, Y and some overlaps of them for the sample in b. e, Upconversion emission spectra obtained from the NaErF₄:Yb/Y(10/40 mol%)@NaYbF₄@NaYF₄:Yb/Tm(50/2 mol%)@NaYF₄, NaErF₄:Yb/Y(10/40 mol%)@NaYbF₄@NaGdF₄:Yb/Tm(50/1 mol%)@NaYF₄ and NaErF₄:Yb/Y(10/40 mol%)@NaYbF₄@NaGdF₄:Yb/Tm(50/1 mol%)@NaYF₄:A (A = Eu, Tb; 5 mol%) multilayer core–shell nanoparticles together with a control sample of NaErF₄:Yb/Y(10/40 mol%)@NaYF₄ core–shell nanoparticles under 1,530 nm excitation.

Under 1,530 nm excitation, typical upconverted emissions of Tm³⁺ ranging from the near-infrared (for example, 802 nm from the ³H₄ → ³H₆ transition) to the visible (for example, 476 nm from the ¹G₄ → ³H₆ transition) and ultraviolet (for example, 360 nm from the ¹D₂ → ³H₆ transition) were observed, separate from the green and red emissions of the Er³⁺ sensitizer (Fig. 2e and Supplementary Figs. 3 and 4). The outermost NaYF₄ layer in these samples was used to eliminate possible surface quenching effects (Supplementary Fig. 5). Acting as a control, the absence of a migratory NaYbF₄ interlayer in the sample only resulted in much weaker Tm³⁺ emission, in particular in the short-wavelength range (Supplementary Fig. 6), and no Tm³⁺ emission was recorded for the sample without doping of Yb³⁺ in the interlayer (Supplementary Fig. 7). This implies the existence of strong interactions between Er³⁺ and Tm³⁺, which seriously quench the emission of Tm³⁺. This observation is in agreement with structural designs for multicolour switchable upconversion, in which an inert interlayer is essential for Tm³⁺ emissions^24,25,38. Further control experiments indicated that a 2.9-nm-thick interlayer can effectively isolate interfacial interactions (Supplementary Fig. 8), a thinner migratory layer with high Yb³⁺ concentration is much better than a thicker migratory layer with low Yb³⁺ concentration, and the NaYbF₄ layer has an optimal thickness (Supplementary Figs. 9 and 10). The presence of an appropriate amount of Yb³⁺ in the sensitizing core and luminescent layer can further improve the total energy-transport process, whereas it is almost useless for the simple co-doping of all lanthanides in the core region, as shown for NaErF₄:Yb/Tm@NaYF₄ nanoparticles (Supplementary Fig. 11). It should be noted that direct activation of Tm³⁺ by Er³⁺ is only available for a very low Er³⁺ concentration³⁹, such as NaYF₄:Er/Tm(2/0.5 mol%)@NaYF₄, resulting in very weak upconversion (Supplementary Fig. 12). These results clearly confirm the critical role of the energy-migratory Yb sublattice in achieving upconversion emission of Tm³⁺ upon excitation at 1,530 nm. Details of the energy-transport process are illustrated schematically in Supplementary Fig. 13.

Strikingly, the observation of ultraviolet Tm³⁺ emissions under 1,530 nm irradiation would further enable photon upconversion of more lanthanide ions towards emissions at specific wavelengths. Indeed, the presence of Gd³⁺ in the luminescent layer (Type-II design in Fig. 2a) led to its 311 nm emission (⁶P_7/2 → ⁸S_7/2 transition; Fig. 2e and Supplementary Fig. 14). It is worth noting that the anti-Stokes shift of this emission is as large as 1,219 nm, a value that is not achievable for conventional 980 or 808 nm upconversion systems. By further introducing the lanthanides of Eu³⁺ and Tb³⁺ into the outermost shell layer (Type-III design in Fig. 2a), their typical upconversion emissions, such as the red emission of Eu³⁺ at 615 nm (⁵D₀ → ⁷F₂ transition), were also easily accessible (Fig. 2e and Supplementary Fig. 15). In the above two sample designs, Tm³⁺ was involved in the upconversion process, as its ladder-like energy levels can further upconvert the energy that Yb³⁺ transfers to the shell (Supplementary Fig. 16). The Yb-mediated energy migration also applies to the upconversion of Ho³⁺ under 1,530 nm excitation (Supplementary Figs. 13b and 17). Notably, energy migration among Er³⁺ ions in the core region cannot be taken into consideration (Supplementary Fig. 18), although it is a primary origin of the energy loss for an erbium-based host such as NaErF₄ (ref. ⁴⁰). These observations have extensively demonstrated that the migratory Yb sublattice-mediated nanostructure is a general approach to the 1,530-nm-responsive upconversion of lanthanide ions.

Although highly effective for upconversion, the sample designs in Fig. 2a cannot be used to examine the mechanism of energy migration because of the presence of different Yb³⁺ compositions in the migratory interlayer and other regions (Supplementary Fig. 19). To gain a deeper insight into this issue, we designed a NaYF₄:Nd(40 mol%)@NaYF₄:Yb(0–100 mol%)@NaYF₄:Ho(2 mol%) core–shell–shell nanostructure to confine Yb³⁺ in the interlayer, enabling the detection of Yb-mediated energy migration (Fig. 3a,b and Supplementary Fig. 20a). In detail, Nd³⁺ was used in the core to sensitize Yb³⁺ in the migratory interlayer through an IET process, from Nd³⁺ (at its ⁴F_3/2 state) to Yb³⁺ (at its ²F_5/2 state), at 808 nm excitation (Fig. 3c). Ho³⁺ was adopted in the outermost shell to detect energy migration in the interlayer because it can be activated by the Yb-to-Ho IET process, with the resultant observation of Ho³⁺ emission (Fig. 3d). It should also be noted that both Yb³⁺ and Ho³⁺ are non-responsive to the incident 808 nm excitation. Therefore, this design allows an in-depth study of energy migration over the Yb sublattice by monitoring changes in the Ho³⁺ emission profile.

Fig. 3: Mechanistic investigation of energy migration in the Yb sublattice.

a, Schematic of the proposed NaYF₄:Nd@NaYF₄:Yb(0–100 mol%)@NaYF₄:Ho core–shell–shell nanostructure for investigating energy migration in the Yb sublattice by recording the change in Ho³⁺ emission from the outermost detecting shell layer with irradiation at 808 nm. b, The energy transport process involved in the design of a. c, Near-infrared emission spectra of NaYF₄:Nd(40 mol%)@NaYF₄:Yb(10 mol%) and NaYF₄:Nd(40 mol%)@NaYF₄ core–shell nanoparticles under 808 nm excitation. d, Upconversion emission spectra of NaYF₄:Nd(40 mol%)@NaYF₄:Yb(50 mol%)@NaYF₄:Ho(2 mol%) and NaYF₄:Nd(40 mol%)@NaYF₄@NaYF₄:Ho(2 mol%) nanoparticles under 808 nm excitation. Inset: Ho³⁺ emission photograph of the former sample. e, Dependence of Ho³⁺ emission intensity and lifetime of Yb³⁺ at its ²F_5/2 state on the concentration of Yb³⁺ in the NaYF₄:Nd(40 mol%)@NaYF₄:Yb(0–100 mol%)@NaYF₄:Ho(2 mol%) core–shell–shell nanoparticles with excitation at 808 nm. f, Emission and absorption spectral overlap for the ²F_5/2 ↔ ²F_7/2 transitions of Yb³⁺.

As shown in Fig. 3e and Supplementary Fig. 20b, under 808 nm excitation, the emission intensity of Ho³⁺ depends closely on the Yb³⁺ concentration in the migratory interlayer. The initial increase in Ho³⁺ emission with increasing Yb³⁺ concentration in the interlayer can be understood by a reduction of interionic Yb³⁺ separation that can, in principle, facilitate more efficient energy migration, according to Dexter’s theory^1,41. The decline in emission observed at much higher Yb³⁺ content might result from the concentration effect, which may cause additional energy loss. This is in accordance with the decrease in the lifetime of Yb³⁺ in the ²F_5/2 state during this process (Fig. 3e and Supplementary Fig. 21a). Intriguingly, the interlayer with low Yb³⁺ concentration (for example, 10 mol%) begins to show migratory performance, possibly due to its simple electronic energy levels with a broad spectral overlap between the associated emission and absorption transitions (Fig. 3f). An Yb³⁺ content of 50 mol% produced an optimum energy-migration property with an Yb³⁺–Yb³⁺ separation of 5.25 Å (Supplementary Fig. 22)⁴². These observations suggest that Yb³⁺ is a good candidate to enable upconversion of lanthanides via its energy-migration property. On the other hand, intense Yb³⁺ intrinsic emission at around 977 nm was also recorded (Supplementary Fig. 21b), implying that there is competition between energy migration and spontaneous radiation at the migratory ²F_5/2 level. However, this might leave room to further improve the migration performance over the Yb sublattice by suppressing its spontaneous radiative transition (Supplementary Fig. 23).

We next investigated how to realize NIR I single-band upconversion of Tm³⁺ at around 802 nm, a commonly used wavelength region for bioapplications^6,23. To minimize unwanted visible emission from the Er³⁺ sensitizer, Ce³⁺ was introduced into the core matrix to depopulate its visible-emitting levels through cross-relaxation. Figure 4a shows four typical cross-relaxation processes (CR1–4) that can occur at the ⁴S_3/2, ⁴F_9/2, ⁴I_9/2 and ⁴I_11/2 levels of Er³⁺, respectively. Note that CR1–3 can help populate Er³⁺ at ⁴I_11/2, while CR4 is a depopulating process for this energy level because it reduces the energy transfer from Er³⁺ (⁴I_11/2) to Yb³⁺ (²F_5/2). Consequently, optimization of the Ce³⁺ concentration is needed. As shown in Fig. 4b and Supplementary Fig. 24, the visible upconversion of Er³⁺ is effectively depressed in the presence of increased Ce³⁺ in the core region. When Ce³⁺ reaches 0.4 mol%, the 802 nm emission intensity accounts for over 90% of the total upconversion emission spectrum (Fig. 4c). This percentage is well maintained over a broad range of pump power densities (up to 1.86 × 10² W cm⁻²; Fig. 4c and Supplementary Fig. 25). In these samples, the presence of a small amount of Ce³⁺ did not lead to its emission, and had almost no impact on the size and morphology of the nanoparticles (Supplementary Figs. 26 and 27). Thus, we can experimentally consider it to be an NIR II-to-NIR I quasi-single-band upconversion for samples with Ce³⁺ content higher than 0.4 mol%. This is particularly attractive for biological applications. A detailed investigation showed that emission at 802 nm has much deeper penetration depth in pork tissue slides than that at 662 nm under 1,530 nm excitation (Supplementary Fig. 28). As a control, visible upconversion light from a typical 808 nm excitation scheme (for example, NaYF₄:Yb/Er@NaYF₄:Nd core–shell nanoparticles) was barely observed, even with a 0.9-mm-thin pork tissue slide (Supplementary Fig. 29).

Fig. 4: Tuning photon upconversion through Ce³⁺-mediated cross-relaxation.

a, Schematic of Ce³⁺-mediated cross-relaxations (CR1–4) towards quasi-single-band upconversion of Tm³⁺ in the NIR I biowindow. CR1: [Er³⁺ (⁴S_3/2), Ce³⁺ (²F_5/2)] → [Er³⁺ (⁴F_9/2), Ce³⁺ (²F_7/2)]; CR2: [Er³⁺ (⁴F_9/2), Ce³⁺ (²F_5/2)] → [Er³⁺ (⁴I_9/2), Ce³⁺ (²F_7/2)]; CR3: [Er³⁺ (⁴I_9/2), Ce³⁺ (²F_5/2)] → [Er³⁺ (⁴I_11/2), Ce³⁺ (²F_7/2)]; CR4: [Er³⁺ (⁴I_11/2), Ce³⁺ (²F_5/2)] → [Er³⁺ (⁴I_13/2), Ce³⁺ (²F_7/2)]. b, Upconversion emission spectra of NaErF₄:Yb/Y/Ce(10/40/x; x = 0–1.0 mol%)@NaYbF₄@NaYF₄:Yb/Tm(49/1 mol%)@NaYF₄ multilayer core–shell nanoparticles under 1,530 nm excitation. c, The integrated intensity percentage (γ) of 802 nm emission as a function of Ce³⁺ concentration for the samples in b. Inset: the percentage of the sample doping with 0.4 mol% Ce³⁺ under different pump power densities. d, Dependence of the upconversion emissions of Er³⁺ on Ce³⁺ concentration for NaErF₄:Y/Ce(50/x; x = 0–1.0 mol%)@NaYF₄ core–shell nanoparticles. e, Upconversion emission spectra of NaErF₄:Yb/Y/Ce(10/40/x; x = 0, 0.4 mol%)@NaYbF₄@NaGdF₄:Yb/Tm(50/1 mol%)@NaYF₄:Eu(5 mol%) nanoparticles under 1,530 nm excitation.

To shed more light on the cross-relaxation, the impact of Ce³⁺ on the ⁴I_9/2 and ⁴I_11/2 energy levels of Er³⁺ was investigated further. The emissions from these two levels were almost completely covered by those of Tm³⁺ (at 802 nm) or Yb³⁺ (at 977 nm) for the aforementioned sample (Fig. 4a and Supplementary Fig. 30), making it difficult to distinguish them precisely from the measured spectrum. Here, the control sample of NaErF₄:Y/Ce@NaYF₄ was prepared by substituting inert Y³⁺ for Yb³⁺ in the core region as well as removing the other layers to get rid of the interfering emission bands. As displayed in Fig. 4d and Supplementary Fig. 31, the near-infrared and visible emissions of Er³⁺ were recorded and show a rapid decline at a higher Ce³⁺ dopant concentration, suggesting the occurrence of cross-relaxation at these energy levels. Strikingly, an initial increase in 977 nm emission was observed for the sample doped with a small amount of 0.2 mol% Ce³⁺. This implies that, at lower Ce³⁺ concentration, cross-relaxation at the ⁴I_11/2 level might be less efficient than that at higher energy levels, resulting in a positive population of Er³⁺ at such an energy level. This is in agreement with estimates of cross-relaxation rates that are smaller for CR4 than the other three processes (Supplementary Figs. 32 and 33, Supplementary Scheme 1 and Supplementary Table 2)^1,43,44,45. Introducing Ce³⁺ into the Type-II and Type-III samples can also improve the upconversion emissions of emitters, although they are still relatively weak (Fig. 4e and Supplementary Fig. 34). On the other hand, there was almost no quenching effect observed for Er³⁺ at its ⁴I_13/2 state due to the presence of Ce³⁺ (Supplementary Fig. 35), which helps in maintaining an efficient excitation scheme at a wavelength of 1,530 nm.

As an added benefit, the use of an energy-migratory Yb sublattice provides an approach to manipulating the upconversion dynamics to achieve switchable control of emission colours. The NaErF₄@NaYbF₄ core–shell nanostructure design resulted in a red-to-green colour change on switching the excitation from 980 to 1,530 nm (Fig. 5a, Supplementary Fig. 36a and Supplementary Table 3). This might be attributable to the NaYbF₄ shell, which provides an energy-transport channel to suppress red upconversion emission by depopulating the intermediate state (⁴I_11/2) upon 1,530 nm excitation (Supplementary Fig. 36b)^35,46. Control samples without the NaYbF₄ shell or with only Yb³⁺ in the core did not give rise to the colour-switchable performance (Supplementary Figs. 37 and 38). Such smart control of output emission colours offers an opportunity for information security. A green ‘dragonfly-on-lotus’ pattern made of NaErF₄@NaYbF₄ nanoparticles was clearly distinguished from dazzle light (at 980 nm irradiation) by the use of 1,530 nm excitation (Fig. 5b, Supplementary Fig. 39 and Supplementary Table 4). Further preparing a quick-response (QR) code using these nanoparticles demonstrated their potential to be applied in anti-counterfeiting (Fig. 5c).

Fig. 5: Colour-switchable output and its application.

a, Upconversion emission spectra of NaErF₄@NaYbF₄ under 980 nm (top) and 1,530 nm (bottom) excitations. Insets: corresponding emission photographs. b, The use of 1,530 nm irradiation to decode a ‘dragonfly-on-lotus’ pattern (right), which presents as dazzle light under 980 nm irradiation (left). c, Schematic of decoding the information patterned in a quick-response (QR) code by scanning with a mobile phone. The patterns in b and c were prepared using the sample in a and control nanoparticles of NaYF₄:Yb/Ho/Ce(20/2/10 mol%)@NaYF₄, NaYF₄:Yb/Er(20/2 mol%)@NaYF₄ and NaYF₄:Yb/Tm(30/1 mol%)@NaYF₄.

In conclusion, we have mechanistically demonstrated that the energy-migration-mediated nanostructure is an efficient and general approach to realizing NIR II-responsive upconversion from a broad range of lanthanide ions. Our results provide a thorough understanding of energy migration in an Yb sublattice and its intrinsic role during the dynamic control of upconversion. Moreover, they should help stimulate new concepts for photon upconversion and the development of a new class of luminescent materials. On a separate note, the Yb sublattice still yields radiation from its migratory energy level, which will result in competition with energy migration and limit its performance. We anticipate that future investigations will search for new strategies to minimize the intrinsic spontaneous radiation of the migratory sublattice, develop facile synthetic methods for multilayer core–shell nanostructures and even simplify the sample structural designs. Given its versatility and general tunability, the striking conceptual model described here shows great potential for frontier applications such as biophotonics, information security and anti-counterfeiting.

Methods

All experimental details including samples synthesis, characterization and optical measurements are provided in the Supplementary Information.