Core-shell structured NaMnF3: Yb, Er nanoparticles for bioimaging applications

NaMnF3: Yb,Er upconversion nanoparticles (UCNPs) have received considerable attention due to their single-band emission. In this paper, a modified thermal decomposition method to synthesize single or core/shell structured lanthanide-doped NaMnF3 nanoparticles is proposed. The NaMnF3: 20% Yb , 2% Er nanoparticles displayed pure red emission under 980 nm laser excitation, and the emission intensity could be significantly enhanced by coating a NaMnF3 or NaMnF3: 20% Yb 3+ shell layer on their surfaces. Moreover, NaMnF3: 20% Yb , 2% Er nanoparticles shelled with NaMnF3: 20% Nd 3+ or NaMnF3: 20% Yb, 20% Nd were synthesized for the first time, and displayed pure red emission when excited by an 808 nm laser. The core/shell structured NaMnF3: Yb/Er@NaMnF3: Yb or NaMnF3: Yb/Er@NaMnF3: Yb/Nd nanoparticles were applied to bioimaging. The NaMnF3 based UCNPs exhibited good photostability and biocompatibility in HeLa cells. The obtained images indicated that the UCNPs could mainly exist in the cytoplasmic regions in the cells. Besides, images from the same region exposed to laser irradiations at 808 nm and 980 nm were found to be comparable. This result indicated that, for NaMnF3: 20% Yb , 2% Er@NaMnF3: 20% Yb , 20% Nd nanoparticles, laser excitation at 808 nm was as efficient as that at 980 nm for bioimaging.


Introduction
Lanthanide-doped upconversion nanoparticles (UCNPs) have received considerable attention due to their intriguing features including non-autouorescence, low photobleaching, strong penetration abilities, and low toxicity. [1][2][3][4][5][6][7][8][9][10] UCNPs have been widely used in solar cells, biolabeling, bioimaging, optical data storage, drug delivery, etc. [11][12][13][14][15] UCNPs with red single-band emission are preferred for biological imaging applications, as red light can penetrate deeper than other visible wavelengths through most tissues. 16,17 Besides, UCNPs with only red singleband emission are ideal candidates for photodynamic therapy applications, as the absorption peak of most commercial photosensitizers is located in the red region at $650-670 nm. 18,19 Therefore, to satisfy the requirements of several applications, the synthesis of UCNPs with pure red emission is crucial. 20 Until now, most research studies on the synthesis of UCNPs with strong red or near-infrared (NIR) emission have focused on Yb/Er or Yb/Tm doped NaYF 4 nanoparticles, as strong green ($540 nm) and weak red ($654 nm) emissions are produced by Yb/Er doped NaYF 4 nanoparticles, whereas blue ($480 nm) and strong NIR ($800 nm) emissions are produced by Yb/Tm doped NaYF 4 nanoparticles. [21][22][23][24] Although each trivalent lanthanide activator ion has a unique energy level structure to produce emission peaks at specic wavelengths, the excitation and relaxation dynamics of energy levels involved in UC processes can be manipulated to vary the relative luminescence intensity between different UC bands. [25][26][27][28][29][30][31][32] The UC emission color may be tuned by incorporating other ions such as Mn 2+ , Zr 4+ , Mg 2+ or Ce 3+ in the host lattice. [33][34][35][36][37][38] Mn 2+ ion doped NaYF 4 nanoparticles show very promising performances, because the 4 T 1 energy states of Mn 2+ ions can facilitate non-radiative energy transfer from the 2 H 9/2 and 4 S 3/2 levels to the 4 F 9/2 level of Er 3+ ions (from 1 D 2 and 1 G 4 levels to 3 F 4 levels of Tm 3+ ions), which can tune the color emission from green to red or from blue to NIR for the Yb/ Er and Yb/Tm doped UCNPs, respectively. Thus, UCNPs with single red or NIR band emissions can be obtained.
Besides, UCNPs with Yb 3+ as sensitizer are generally excited by a 980 nm laser. However, the use of 980 nm NIR photons has an intrinsic disadvantage due to the strong absorption of water molecules in biological tissues at that wavelength, resulting in the risk of local temperature rise or even tissue overheating under continual irradiation. [39][40][41] To avoid water absorption in tissues, the wavelength of the exciting light source should be shorter than 800 nm. A promising strategy was proposed to incorporate Nd 3+ ions into UCNPs as sensitizers to absorb the energy of the 800 nm photons, followed by transfer of non-radiative energy to Yb 3+ ions. 42 Owing to the quenching effect between them, Nd 3+ and activator ions (Er 3+ , Tm 3+ ) needed to be separated by a core/shell structure to obtain effective Nd 3+ sensitized UCNPs. 43,44 To maintain the penetration depth and simultaneously avoid overheating during in vivo bioimaging, it was necessary to develop UCNPs which can upconvert from $800 nm NIR laser to red light. Obviously, Nd 3+ sensitized NaMnF 3 : Yb/Er nanoparticles were the most promising candidate, although Nd 3+ sensitized Ho 3+ single band UCNPs have been reported. 45 Until now, solvothermal routes have been typically used to synthesize NaMnF 3 : Yb/Er nanoparticles; 46,47 the syntheses have been conducted in teon-lined autoclaves under high pressure, and core/shell UCNPs with perfect morphologies have been hardly obtained. In this work, a novel method to synthesize single or core/shell structured NaMnF 3 : Yb, Er nanoparticles with pure red emission under 808 or 980 nm laser excitation for bioimaging was developed. Then, bioimaging of core/shell structured NaMnF 3 : Yb/Er nanoparticles in HeLa cells was investigated.

Materials
For the proposed procedure, appropriate selection of the raw materials is critical. MnCl 2 hardly reacts with oleic acid to produce the necessary metal-oleic complex. The frequently used rare earth chloride must be abandoned, because chloride ions could hinder the complex reaction of Mn 2+ ions. Rather, rare earth acetate hexahydrate and manganese(II) pentanedionate were used. Chemicals were purchased from Alfa Aesar and used without further purication. The acetate hexahydrate had a trace metal basis of 99.9%.

Synthesis of single nanoparticles
In this work, NaMnF 3 : 20% Yb 3+ , A (A ¼ Er 3+ , Ho 3+ , or Tm 3+ ) nanoparticles were prepared by a modied one-step thermal decomposition. A mixture containing 0.2 mmol of Yb(CH 3 -COO) 3 $4H 2 O, (0.8 À x) mmol of manganese(II) pentanedionate, and x mmol of A(CH 3 COO) 3 $4H 2 O was placed in a 100 ml ask. Then, 7 ml of oleic acid and 15 ml of 1-octadecene were added, and the solution was heated to 160 C for 60 min under Ar atmosphere and vigorous stirring. Aer cooling to room temperature, a solution comprising 4 mmol of NH 4 F and 2.5 mmol of NaOH in 10 ml of methanol was added, and the mixture was heated to 100 C for 30 min to evaporate methanol. Finally, the solution was heated to 300 C under Ar atmosphere for 60 min, and then cooled to room temperature. The nanoparticles were precipitated with 50 ml of ethanol, collected aer centrifugation (7000 rpm, 5 min), and redispersed in 10 ml of hexane for later use.

Synthesis of core/shell structured nanoparticles
To prepare core/shell structured NaMnF 3 : 20% Yb 3+ , A nanoparticles, 0.5 mmol of manganese(II) pentanedionate was added to a 100 ml ask, followed by 7 ml of oleic acid and 15 ml of 1octadecene. The solution was heated to 160 C for 60 min under Ar atmosphere and vigorous stirring. Aer cooling to room temperature, 5 ml of colloidal NaMnF 3 : 20% Yb 3+ /A, synthesized as in Section 2.2, was added, followed by a solution consisting of 2 mmol of NH 4 F and 1.25 mmol of NaOH in 5 ml of methanol; the mixture was then heated to 100 C for 30 min to evaporate methanol and hexane. Finally, the solution was heated to 290 C under Ar atmosphere for 60 min, and subsequently cooled to room temperature. The nanoparticles were precipitated with 50 ml of ethanol, collected aer centrifugation (7000 rpm, 5 min), and redispersed in 5 ml of hexane for characterization.

Surface modication of oleic acid (OA) UCNPs by polymethacrylamide (PAAM)
To obtain hydrophilic UCNPs for biological application, PAAM was used to modify OA-capped UCNPs. Firstly, 0.5 ml of a solution of hexane-dispersed OA-UCNPs was added to 5 ml of anhydrous ethanol containing 0.5 ml of PAAM, and stirred vigorously at ambient temperature for 24 h. The UNCPs were precipitated by centrifugation (10 000 rpm; 15 min) and redispersed in phosphate buffer saline solution by sonication for further research.

Characterization
The sample morphologies were observed by transmission electron microscopy (TEM; Joel JEM-2100; acceleration voltage: 200 kV). The UC luminescence spectra were recorded using an Ocean Optics QE65000 spectrouorometer with a slit width dening a spectral resolution of 1 nm. The colloidal UCNPs were excited at 980 or 808 nm by using a continuous-wave laser diode with tunable laser power in the scale of 0-0.8 W. The spectra of colloidal UCNPs were measured with 0.5 W laser irradiation if no special requirement was mentioned. The emission decay time was measured by using the proposed system illustrated in. 48 Fig. 1a shows the TEM morphologies of the synthesized NaMnF 3 : 20% Yb 3+ , 2% Er 3+ nanoparticles. Ultrasmall lanthanide-doped NaMnF 3 nanoparticles were synthesized following the proposed procedure; all the nanoparticles were uniformly monodispersed. The average size of the NaMnF 3 : 20% Yb 3+ , 2% Er 3+ nanoparticles was approximately 9.5 nm. The XRD pattern shown in Fig. S1 † was in accord with the standard NaMnF 3 host lattice of JCPDS 18-1224, indicating the prepared UCNPs were NaMnF 3 phase. Fig. 1b shows the UC emission spectrum of the as-synthesized NaMnF 3 : 20% Yb 3+ , 2% Er 3+ nanoparticles. Only a single emission band centered at 654 nm, which corresponds to the energy transfer 4 F 9/2 / 4 F 15/2 of Er 3+ ions, is visible in the spectrum. The emission band corresponding to the energy transfer 4 S 3/2 / 4 F 15/2 of Er 3+ ionspresent in the NaYF 4 : 20% Yb 3+ , 2% Er 3+ nanoparticle spectrum-completely disappears in the NaMnF 3 : 20% Yb 3+ , 2% Er 3+ nanoparticle spectrum. This is ascribed to the nonradiative energy transfer from the 4 S 3/2 level of Er 3+ ions to the 4 T 1 level of Mn 2+ ions, followed by back energy transfer to the 4 F 9/2 level of Er 3+ ions. Hence, the color output of the NaMnF 3 :
Notably, the UC emission intensity can be signicantly enhanced by coating effective shell layers on the surface of the UCNPs. 4,49 Layers of different types (i.e., an inert NaMnF 3 layer or an active NaMnF 3 : 20% Yb 3+ layer) were coated on the surfaces of NaMnF 3 : 20% Yb 3+ , 2% Er 3+ nanoparticles to boost the UC emission intensity, as shown in Fig. 2. The red emission intensity was enhanced by a factor of 1.75 in the presence of the inert NaMnF 3 layer, whereas an enhancement factor of 3 was obtained by coating with the active NaMnF 3 : 20% Yb 3+ layer. The output color of all these coated nanoparticles was red, as shown in the inset images; however, a very weak peak centered at 540 nm appeared in the spectrum of the NaMnF 3 : 20% Yb 3+ , 2% Er 3+ @NaMnF 3 : 20% Yb 3+ nanoparticles, whose TEM image is shown in Fig. 2b. The average size was approximately 16 nm, almost twice that of the NaMnF 3 : 20% Yb 3+ , 2% Er 3+ nanoparticles.
To investigate the mechanism behind the intensity enhancement of the red emission in core/shell structured lanthanide-doped NaMnF 3 nanoparticles, the dependence of the red emission intensity on the pump laser power was measured for NaMnF 3 : 20% Yb 3+ , 2% Er 3+ nanoparticles with or without shell layer (Fig. 3). Generally, the number of photons required to populate the upper emitting state under unsaturated conditions is related to the intensity by: where I f is the photoluminescence intensity, P is the pump laser power, and n is the number of laser photons required. The  slopes of the red-emission bands of the NaMnF 3 : 20% Yb 3+ , 2% Er 3+ , NaMnF 3 : 20% Yb 3+ , 2% Er 3+ @NaMnF 3 , and NaMnF 3 : 20% Yb 3+ , 2% Er 3+ @NaMnF 3 : 20% Yb 3+ nanoparticles were 1.37, 1.51, and 1.79, respectively, indicating that the red emission involved a two-photon process in all these cases. Additionally, the enhanced slope of the shell-structured nanoparticles indicated that the shell layer was suppressing surface-related quenching effects.
The decay proles of the 4 S 3/2 / 4 F 15/2 transition of Er 3+ ions at 540 nm and 4 F 9/2 / 4 F 15/2 transition of Er 3+ ions at 654 nm were also measured, as shown in Fig. 4. The decay curves were non-exponential, owing to enhanced non-radiative energy transfer processes, as well as effects of the surrounding environment. The effective lifetime s m is where I(t) is the intensity at time t. The effective decay times for the 4 S 3/2 and 4 F 9/2 states of Er 3+ ions in NaMnF 3 : 20% Yb 3+ , 2% Er 3+ nanoparticles were 19.7 and 34.0 ms, respectively, which were signicantly prolonged in NaMnF 3 shelled nanoparticles. The prolonged lifetime was mainly attributed to the suppression of surface-related quenching effects. The NaMnF 3 : 20% Yb 3+ shell layer could not completely suppress the quenching effect, but it enhanced the absorption of the 980 nm excitation laser. Thus, the lifetime of the NaMnF 3 : 20% Yb 3+ , 2% Er 3+ @NaMnF 3 : 20% Yb 3+ nanoparticles was longer. The appearance of the green emission at 540 nm in the NaMnF 3 : 20% Yb 3+ , 2% Er 3+ @NaMnF 3 : 20% Yb 3+ nanoparticle spectrum (Fig. 2) was related to the segregation of Yb 3+ ions at the interface between core and shell, as the coating of the NaMnF 3 : 20% Yb 3+ shell layer increased the concentration of Yb 3+ activator ions at the interface.
The NaMnF 3 : 20% Yb 3+ , 2% Er 3+ @NaMnF 3 : 20% Yb 3+ , 20% Nd 3+ nanoparticles were also applied to bioimaging, as shown in Fig. 7. The HeLa cells incubated with PAAM-UCNPs were imaged in bright eld and uorescence eld by using a Leica confocal microscope equipped with 808 nm and 980 nm NIR lasers. The merged images ( Fig. 7d and e) indicate that the UCNP uorescence mainly occurred in the cytoplasmic regions. The images obtained for the same region by using 808 nm and 980 nm laser excitations ( Fig. 7a and b) appear comparable, which implied that, for NaMnF 3 : 20% Yb 3+ , 2% Er 3+ @NaMnF 3 : 20% Yb 3+ , 20% Nd 3+ nanoparticles, the 808 nm laser excitation was as efficient as the 980 nm laser excitation for bioimaging. Notably, it has been conrmed 39,43 that, compared with the popular excitation wavelength of 980 nm, the 808 nm excitation wavelength minimized the overheating effect for in vivo tissues.

Conflicts of interest
There are no conicts to declare.