Multicolour synthesis in lanthanide-doped nanocrystals through cation exchange in water

Meeting the high demand for lanthanide-doped luminescent nanocrystals across a broad range of fields hinges upon the development of a robust synthetic protocol that provides rapid, just-in-time nanocrystal preparation. However, to date, almost all lanthanide-doped luminescent nanomaterials have relied on direct synthesis requiring stringent controls over crystal nucleation and growth at elevated temperatures. Here we demonstrate the use of a cation exchange strategy for expeditiously accessing large classes of such nanocrystals. By combining the process of cation exchange with energy migration, the luminescence properties of the nanocrystals can be easily tuned while preserving the size, morphology and crystal phase of the initial nanocrystal template. This post-synthesis strategy enables us to achieve upconversion luminescence in Ce3+ and Mn2+-activated hexagonal-phased nanocrystals, opening a gateway towards applications ranging from chemical sensing to anti-counterfeiting.


S29 | Page
Positive and negative charges are shown in yellow and blue, respectively. It is obvious that the charge transfer mainly occurs between Gd/dopant and F atoms. In the Tb-doped nanocrystal, the charge transfer from Tb to F atoms is similar to that of the pure NaGdF 4 nanocrystal, which is consistent with the small formation energy calculated for the Tb-doped NaGdF 4 nanocrystal. However, the amount of the charge transferred between Mn and F atoms is much less than obtained between Gd and F (Supplementary Table 2

Supplementary Note 1: Structural studies of nanocrystals
The NaGdF 4 :Yb/Tm@NaGdF 4 nanocrystals were synthesized as tempate for cation exchange according to a coprecipitation method 1 . A representative TEM image reveals a spherical shape of the nanocrystals with an average size of 21 nm ( Supplementary Fig. 1a,b). The X-ray powder diffraction pattern of the as-synthesized sample can be well indexed to a hexagonal-phased NaGdF 4 structure ( Supplementary Fig.1c).
Oleate ligands have a strong absorption in the UV region, thus resulting in strong quenching of the migration energy stored in Gd sublattice. To overcome this problem, the oleate ligands on the particle's surface were removed by an acid-treatment approach according to the previous report 2 . The complete removal of oleate ligands was further confirmed by FTIR spectroscopic analysis ( Supplementary Fig. 2). As shown in Supplementary Fig. 3, the emission intensity of Gd 3+ at 311 nm was much stronger than that of oleate-capped nanocrystals. The sharp emission at 311 nm is attributed to the radiative decay from 6 P 7/2 state to the ground state of Gd 3+ ions. In addition, ligand-free NaGdF 4 :Yb/Tm@NaGdF 4 nanocrystals without the oleate ligand stabilization could be well dispersed in an aqueous solution and subsequently amenable for cation exchange.
In our study, the same size, phase, and morphology of the NaGdF 4 :Yb/Tm(49/1%)@NaGdF 4 nanocrystals are found to be largely retained after the cation-exchange treatment. A representative high-resolution STEM image shown in Supplementary Fig. 4 reveals the preservation of the highly crystalline hexagonal structure of nanocrystals after cation exchange with Eu 3+ ions. Elemental analysis and EELS line scan result carried out at single particle levels provide further evidence of the cation exchange process. Both elemental analyses show a higher Eu 3+ content at the particle's surface and a higher Gd 3+ content in the inner layer ( Supplementary Fig. 10).
Notably, the exchange process did not introduce any detectable surface defects.
We next examined the size and morphology evolution of the Gd 3+ -based nanocrystals in the presence of Tb 3+ or  Supplementary Figs 8 and 9).

Supplementary Note 2: Inductively coupled plasma mass spectroscopic (ICP-MS) analysis
To validate that our synthesis is governed by a cation exchange process, we carried out ICP-MS analysis to examine Gd 3+ content in the solution and Tb 3+ content in the nanoparticles after treatment of NaGdF 4 :Yb/Tm@NaGdF 4 nanoparticles with TbCl 3 . As shown in Supplementary Fig. 11, with increasing TbCl 3 concentration for cation exchange reaction, the Gd 3+ content in the solution and Tb 3+ content in the nanoparticles both significantly increased ( Supplementary Fig. 12). The similar trend of ICP-MS test for Gd 3+ content was observed in the EuCl 3 -treatment nanoparticle solution ( Supplementary Fig. 13). These data demonstrate that our synthesis is governed by the cation exchange process, rather than the deposition of the shell containing Tb 3+ .
To further confirm our hypothesis, we carried out ICP-MS analysis to examine Gd 3+ content in the solution after treatment of nanoparticles with Tb 3+ under different reaction temperatures. We found that the Gd 3+ content discharged from nanoparticles gradually increased with increasing the reaction temperature (Supplementary Fig. 14), further substantiating the occurrence of cation exchange.

Supplementary Note 3: Optimization of cation exchange-induced color tuning
For luminescence tuning via cation exchange, energy migration through Gd 3+ sublattices is of significance because it can bridge the energy transfer from the inner layers of the nanoparticle to its surface. To validate this hypothesis, we prepared two sets of NaYF 4 :Yb/Tm@NaYF 4 and NaGdF 4 :Yb/Tm@NaGdF 4 nanoparticles and treated them with Tb 3+ separately. As a result, no emission of Tb 3+ could be detected from Y 3+ -based nanocrystals, while a strong Tb 3+ emission was observed from Gd 3+ -based nanocrystals ( Supplementary Fig. 15).
In a typical energy migration upconversion (EMU) process involving 980 nm excitation, Yb 3+ ions harvest the excitation energy and subsequently populate the excited states of Tm 3+ via multi-step energy transfer upconversion process. The excitation energy stored at the 1 I 6 state of Tm 3+ ions is further transferred to the 6 P 7/2 state of Gd 3+ , followed by energy migration over the host lattice to the particle surface. Ultimately, the excitation energy is trapped by the surface-exchanged activators, allowing the emission color modulation to be achieved. On the basis S41 | Page of this principle, the emission color could be readily modulated by varying the composition and concentrations of the activator precursors ( Supplementary Fig. 16).
The color tuning strategy by cation exchange is strongly dependent on three parameters, namely, the reaction time, the concentration of the exchange ion, and the reaction temperature. To investigate the influence of reaction time on the optical property of the resulting nanocrystals upon cation exchange, we prepared an aqueous solution of ligand-free NaGdF 4 :Yb/Tm@NaGdF 4 nanocrystals (0.9 mL; 26.2 mg). After addition of Tb 3+ ions (0.2 M; 100 L) at room temperature, the upconversion luminescence of the colloidal solution was continuously monitored at a time interval of 1 min. By benchmarking the luminescence profile of Tb 3+ with the Tm 3+ emission at 646 nm, we noticed that the emission of Tb 3+ was observed right at the onset of cation addition. Then the emission intensity of Tb 3+ was gradually boosted, reaching a plateau after 8 min ( Supplementary Fig. 19).
To optimize the concentration of the exchanged activators for intense emissions, we performed a series of cation exchange experiments by mixing a fixed quantity of NaGdF 4 :Yb/Tm@NaGdF 4 nanoparticles with Tb 3+ ions of different concentrations at room temperature for 1 hour. As shown in Supplementary Fig. 17, with increasing Tb 3+ concentration, the emission intensity of Tb 3+ initially increased and then gradually decreased owing to the concentration quenching effect, while the intensity of Gd 3+ emission at 311 nm ( 6 P 7/2  8 S 7/2 transition) decreased gradually ( Supplementary Fig. 17c). The decay curves of Gd 3+ emissions from different batches of nanoparticles obtained with increasing Tb 3+ concentration reveal a consistent, steady decrease in emission lifetime as shown in Supplementary Fig. 17d, providing further evidence of success in the cation exchange process. In reference to Tm 3+ emission at 450 nm, we found that the optimal Tb 3+ concentration for maximum emission intensity was estimated to be 15 mM by evaluating the emission profiles of Tb 3+ at 546 nm ( Supplementary Fig. 17b).
We next prepared a set of at 311 nm reveals the partial substitution of Gd 3+ with Ce 3+ in the host lattice ( Supplementary Fig. 23b). To the best of our knowledge, this is the first demonstration of upconverted Ce 3+ emission from hexagonal-phased nanocrystals. The broadband emission of Ce 3+ in the UV range should be useful particularly for the promotion of photochemical processes.
To optimize the emission of Ce 3+ , we further prepared a collection of NaGdF 4 :Yb/Tm@NaGdF 4 nanocrystals treated with Ce 3+ ions at different concentrations at room temperature. As shown in Supplementary Fig. 24, with increasing Ce 3+ concentration, the emission intensity of Ce 3+ increased gradually and then reached a plateau in reference to Tm 3+ emission at 450 nm.
As shown in Supplementary Fig. 25a, a record-long lifetime of 1.6 ms could be detected for Ce 3+ emission at 380 nm from the resulting Ce 3+ -exchanged NaGdF 4 :Yb/Tm@NaGdF 4 nanocrystals. Notably, this lifetime is much longer than that of Tm 3+ emissions at 346 (221 s) and 364 nm (290 s), which allows the emission profile of Ce 3+ to be separated from the Tm 3+ emission by time-resolved emission spectroscopy ( Supplementary Fig. 25b, c).  Table 1).

Supplementary Note 5: Mechanistic study of cation exchange-induced color tuning
The cation exchange reaction is strongly influenced by the concentration of the exchange ions or the reaction temperature. To validate this effect, we carried out inductively coupled plasma mass spectroscopy (ICP-MS) analysis of the supernatant of the cation exchanged colloidal solution after dialysis. As shown in Supplementary   Fig. 28a, with increasing amounts of Tb 3+ at room temperature (25 o C), the Gd content discharged from the S44 | Page nanocrystals gradually increases after the treatment of NaGdF 4 :Yb/Tm@NaGdF 4 nanocrystals with Tb 3+ . That suggests that high concentrations of exchanged ions could promote the process of cation exchange. Next, we performed a control reaction between NaGdF 4 :Yb/Tm@NaGdF 4 nanocrystals and Tb 3+ ions at 90 o C. After dialysis, the ICP-MS result revealed that the Gd 3+ content discharged from the nanocrystals is higher as compared with that attainable at 25 o C. This clearly demonstrates that the cation exchange reaction could occur more easily at an elevated temperature.
Supplementary Fig. 28b shows that the cation exchange reaction through the use of Mn 2+ is less favorable than that driven by lanthanide ions, as evidenced by the relatively small amount of Gd 3+ ions obtained from the supernatant of the colloidal solution after the reaction. This can be explained by the fact that the calculated formation energy in Table 2 and Supplementary Fig. 29).

Mn-doped nanocrystals is higher and the transferred charge between Mn and Gd is lower (Supplementary
On the basis of optical investigations, we found that the nanoparticles obtained by cation exchange reactions under higher temperature tend to afford intense activator emission (Supplementary Fig. 20). We attributed this effect to the increased cation exchange rate and subsequently accelerated ion diffusion from the particle's outmost surface  Fig. 30). The enhanced emission is attributed to the accelerated rate of ion diffusion at elevated temperatures, as revealed by our calculated moderate diffusion barriers (Supplementary Table 2 and Supplementary Fig. 31).  Fig. 32). Taken together, these data further confirm that the enhanced emission intensity is attributed to the improved processes of cation exchange and ion diffusion at high temperatures.

Supplementary Note 6: Tb 3+ -mediated energy migration upconversion
To demonstrate the generality of our approach for color tuning in other systems, we carried out cation exchange reactions in Tb 3+ -activated nanoparticles. As a proof-of-concept experiment, we prepared the ligand-free NaYbF 4 :Tb(30%)@NaTbF 4 core-shell nanocrystals and then performed the cation exchange reaction with Eu 3+ ions in water at 90 o C. Following the 980 nm excitation, the excitation energy is firstly harvested by Yb 3+ and transferred to neighboring Tb 3+ ions via a cooperative sensitization upconversion process. After that, Tb 3+ is expected to transfer its energy to exchanged Eu 3+ ions, thereby allowing the particle's emission to be modulated ( Supplementary Fig. 33).
Prior to luminescence investigations, we examined the size, phase and morphology of the as-prepared nanocrystals before and after the treatment of Eu 3+ ions at 90 o C. As displayed in Supplementary Fig. 34, it is evident that the cation exchange process did not induce noticeable changes in particle size, phase, and morphology. To validate the emission tunability enabled by our strategy, we performed cation exchange reactions by mixing NaYbF 4 :Tb@NaTbF 4 nanocrystals (26.2 mg) with Eu 3+ at varying concentrations. As shown in Supplementary Fig.   35, the Eu 3+ emission increases gradually with increasing Eu 3+ concentration.

Supplementary Note 7: Down-conversion luminescence
Our cation exchange approach is also applicable to fine-tuning of down-conversion luminescence. To validate this hypothesis, we reacted NaGdF 4 :Ce(15%)@NaGdF 4 core-shell nanoparticles with different types of activator (Tb 3+ , Eu 3+ , Dy 3+ , and Mn 2+ ) at a temperature of 90 o C. As shown in Supplementary Fig. 36 hydroxide (NaOH; >98%), ammonium fluoride (NH 4 F; >98%), 1-octadecene (90%), and oleic acid (90%) were all purchased from Sigma-Aldrich. Unless otherwise noted, all the chemicals were used without further purification.  Computational details: In order to unravel the cation exchange barrier between the host Gd 3+ and guest ions (Ce 3+ , Eu 3+ , Tb 3+ , Dy 3+ or Mn 2+ ), we calculated the formation energies of Ce-, Eu-, Tb-, Dy-and Mn-doped β-NaGdF 4 nanocrystals and the corresponding diffusion barriers of these guest atoms in the host lattices using first principles calculations based on the density functional theory (DFT). All calculations were carried out using the Vienna ab initio simulation package (VASP) with generalized gradient approximation (GGA) and the projector-augmented wave (PAW) method [11][12][13][14] . A kinetic energy cut-off of 500 eV was used for the plane wave. The formation energy of the doped system is given by

S48 | Page
where N is the number of Gd atoms in the system, m is the number of the dopant in the system, and R indicates Ce, Eu, Tb, Dy and Mn atoms. Note that E denotes to the energy of different systems shown in the parentheses. To determine the lowest energy trajectory for dopant motion, we moved the dopant atom toward the Gd vacancy along the (0 0 1) surface normal in 0.9 Å steps. The lateral position of all atoms is allowed to relax until the forces acting on each atom are smaller than 0.02 eV/Å.

S49 | Page
Characterization. Elemental analysis was performed on an Elementar Vario MICRO elemental analyzer. The lifetime and time-resolved emission spectra were measured using an Edinburgh FLSP920 fluorescence spectrophotometer equipped with a xenon arc lamp (Xe900), a microsecond flash-lamp (F900) and a 980 nm diode laser. The lifetimes (τ) of the luminescence were obtained by fitting the decay curve with a multiexponential decay function of ( ) where A i and τ i represent the amplitude and lifetime of individual components for multi-exponential decay profiles, respectively. The photos and supporting videos were recorded by a Nikon D90 camera and iPhone 6, respectively.
Scanning transmission electron microscopy measurement: Atomic resolution scanning transmission electron microscopy (STEM), and electron energy loss spectroscopy (EELS) experiments were carried out on an FEI aberration-corrected Titan Cubed S-Twin transmission electron microscope operated at 200 kV. The STEM measurement with a spatial resolution of ~1.0 Å was performed at a 2-s/pixel scanning rate with 70 m C2 aperture, spot size 9, a high-angle annular dark-filed (HAADF) detector, and 146 mm camera length. For EELS point analysis, 150 m C2 aperture, spot size 6, 29.6 mm camera length, 5 mm entrance aperture (collection angle = 54 mrad) and 1s collection time were used. The EELS line scan was conducted using 70 m C2 aperture, spot size 9, 29.6 mm camera length, 5 mm entrance aperture (collection angle  = 54 mrad) and 0.1 s/pixel collection time.