Intense UV upconversion through highly sensitized NaRF₄:Tm (R:Y,Yb) crystals

Abstract

Photon upconverting luminescent hexagonal NaRF₄:Tm (0.5 mol%) (R:Y³⁺,Yb³⁺) crystals with Yb³⁺ concentrations between 20 and 99.5 mol% were synthesized by a modified thermal coprecipitation method. The effects of the Yb³⁺ sensitizer concentration on the shape, size, structure, upconversion luminescence intensities and dynamic luminescence lifetimes were studied in detail. The intensity of ultraviolet upconversion luminescence at 340–365 nm upon 980 nm excitation increased up to 20 times with increasing Yb³⁺ concentration. The mechanisms for the changes in morphology, size and UV upconversion emission are discussed. In addition, the effect of the Tm³⁺ concentration on the UV emission of NaYbF₄ crystals was investigated. The results demonstrate that the NaYbF₄ with Tm³⁺ concentration between 0.4 to 0.8 mol% produce the most intense UV upconversion luminescence. These crystals may find applications in e.g. UV-visible solid state lasers or as an internal UV radiation source for many photochemical reactions.

---

Introduction

Lanthanides and rare earths have gained wide research interest during the last decades due to their unique electronic, magnetic, optical and chemical characteristics arising from the 4f electrons.^1–3 Lanthanide-based photon upconversion (UC) has been known since the 1960's and photon upconverting nanoparticles (UCNPs) are already widely applied in many applications ranging from medical diagnostics to solar energy harvesting.^4–11 Photon upconversion is a nonlinear optical process that converts the energy of two or more NIR photons into a single emitted photon at shorter wavelength (higher energy). Among the photon upconverting nanomaterials, hexagonal NaYF₄ (β-NaYF₄) nanocrystals have been intensively investigated as the most efficient upconversion host material available to date, because of their low phonon energy and the high quantum efficiencies obtained.^12,13 It is known that also the nanocrystal size, structure, morphology and dopant ion concentration play crucial roles in the UC process.^14–16 In general, the nanocrystal structure and size highly depend on the synthesis temperature, while commonly used organic additives act as shape modifiers to control the morphology of UCNPs. Currently, the research challenges comprise e.g. the optimal dopant ion combinations and understanding the behaviour of the relative intensities of different radiative transitions in luminescence spectrum of UC materials.

In recent years, ultraviolet (UV) UC photoluminescence has become one hot topic of optical functional materials. High energy UV emission is extremely important in many photochemical reactions in the field of chemical biology and advanced imaging.^17–19 Intense UV emitting UCNPs are of special significance with potential technological applications in microelectronic devices and solid-state lasers.²⁰ In our previous report, it was successfully demonstrated that UV emitting UCNPs can be used as an internal light source to form a shell with different functional groups around the particles to further facilitate their use in applications like bioassays, biosensing and theranostics.²¹ The main limitation for practical applications of upconversion materials is the insufficient upconversion efficiency; especially 980 nm excited UV UC photoluminescence remains challenging because it requires at least four NIR photons to populate the UV emitting excited states of Tm³⁺. Although NIR-excited UC to visible/NIR from NaYF₄ with Er³⁺/Tm³⁺ and Yb³⁺ has been extensively studied, NIR-excited UC to UV is still less understood.^22,23

UV UC properties in β-NaYF₄ nanocrystals containing sensitizer Yb³⁺ and activator Tm³⁺ ions have been reported earlier.²⁴ The doping concentration of lanthanide ions however seems to be very important in getting efficient UV UC emission.²⁵ Recently, some researchers have focused their investigations on UV UC emission enhancement by doping with different ions or by a coreshell approach: e.g. Shen et al. have demonstrated the enhanced UV emission in CaF₂ coated cubic NaYF₄ (α-NaYF₄) nanocrystals.²⁶ Zeng et al. obtained higher UV emission from NaYbF₄ microtubes compared to cubic NaYF₄ nanocrystals²⁷ and Zang et al. reported the deep-ultraviolet UC emissions from Gd³⁺ doped NaYF₄:Yb,Tm nanorods.²⁸ It should be noted that although there are a few strategies to enhance the UV emission from UCNPs, systematic studies of sensitizer concentration dependency of UV emission are scarce in literature.

In this work, a simple method is presented to increase the NIR to UV UC photoluminescence by replacing Y³⁺ from NaYF₄:Yb,Tm crystals with Yb³⁺. The effect is investigated by stepwise increasing the sensitizer (Yb³⁺) concentration and optimizing the activator (Tm³⁺) concentration range for the NaYbF₄ (x_Yb = 100%) host. In addition, the influence of different Yb³⁺ and Tm³⁺ concentrations on crystal size, shape, structure and optical properties such as UC luminescence intensity and decay lifetime were also investigated.

Experimental section

Reagents

All of the chemicals were of analytical grade and used as received without further purification. Yttrium, ytterbium and thulium chloride hexahydrate (RCl₃·6H₂O, 99.99%), ammonium fluoride (NH₄F, ≥99.99%) and sodium hydroxide (NaOH, ≥98%), 1-octadecene (ODE, 90%) and oleic acid (OA, 90%) were purchased from Sigma-Aldrich. The purities of the rare earth chlorides are with respect to other rare earths. Absolute ethanol and methanol were used as received.

Material synthesis

Synthesis of NaY_1−xYb_xF₄:Tm (0.5%). The synthesis protocol for the UCNPs was adopted from our previously published procedure.²⁹ Briefly, the syntheses of 0.5% Tm doped NaY_1−xYb_xF₄ UCNPs were done by stepwise substitution of Y³⁺ by 20, 40, 60, 80 and 100% of the sensitizer ion Yb³⁺ (the percentage of Yb³⁺ is with respect to the total rare earth contents excluding the activator). In a typical experiment, 0.2 mol dm⁻³ RCl₃·6H₂O (R:Y and/or Yb, Tm) in methanol were added to a 250 cm³ two neck round-bottom flask containing 9 cm³ oleic acid (OA) and 21 cm³ 1-octadecene (ODE). The solution was magnetically stirred and heated to 160 °C under flowing argon for 40 min to form the rare earth oleate complex and then cooled down to room temperature. Thereafter, 15 cm³ of the methanol solutions of both NH₄F (4.8 mmol) and NaOH (3 mmol) were added and the solution was stirred for 30 min without a protective gas sphere to evaporate methanol and to ensure that all the fluorides were consumed completely. The temperature was then increased to 310 °C and maintained there under argon for 90 min. After the solution was slowly cooled down to room temperature, absolute ethanol was added to the reaction solution to precipitate the nanocrystals. The solution was then transferred to the centrifuge tube and centrifuged at 3766 g for 8 min, so that the nanoparticles separated out completely. This washing procedure was repeated 6 times to ensure washing away any NaF impurity that may have formed. Finally, the nanoparticles were re-dispersed in toluene for further experiments.

Synthesis of NaYbF₄:Tm (x%). Hexagonal NaYbF₄:Tm (x%) were synthesized by replacing all Y³⁺ ions with the sensitizer Yb³⁺ ions and varying the activator ion (Tm³⁺) concentration (x_Tm = 0.05, 0.2, 0.4, 0.6, 0.8, 1 and 5%: the percentage of Tm³⁺ is with respect to the total rare earth contents). The synthesis and washing procedure is the same as that described for the NaY_1−xYb_xF₄:Tm (0.5%) crystals. Finally, the nanoparticles were re-dispersed in toluene for further experiments.

Characterization

The size and morphology of the nanocrystals was characterized by transmission electron microscopy (Tecnai12 BioTwin TEM) with an acceleration voltage of 120 kV. The TEM specimens were prepared by directly drying a drop of a diluted toluene dispersion solution of the as prepared product on the surface of a carbon coated copper grid. The structural phase identification of the synthesized samples was carried out with X-ray powder diffraction (XPD) using a Huber G670 image plate Guinier camera (2θ range: 4–100°) with copper K_α1 radiation (λ: 1.5406 Å). Data collection time was 30 min with 10 scans of the image plate.

The crystallite size, shape and aspect ratio was calculated from the XPD data by using the Scherrer equation³⁰ with instrumental correction from the full width at half maximum (FWHM) determined for a Si reference material (NIST standard 640b):

where d (m) is the mean crystallite size, λ (m) the X-ray wavelength, β (rad) the FWHM of the chosen reflection and θ (°) half of the Bragg's angle (2θ). β_r, β_s and β are the reference, sample and corrected FWHM respectively. The (002) and (200) reflections were used for calculating the length of the hexagonal plates and the width of the hexagonal faces, respectively, thus giving the size and morphology of the crystallites.

The UC luminescence spectra were recorded on Varian Cary Eclipse fluorescence spectrophotometer (Varian Scientific Instruments, Mulgrave, Australia) equipped with a 980 nm laser diode module C2021-F1 (Roithner Lasertechnik, Vienna, Austria). The excitation power used in the experiment was 200 mW with a focusing area of approximately 2.4 mm². The elemental composition was analysed by LEO 1530 Gemini scanning electron microscope with a Thermo Scientific UltraDry SDD X-ray detector (SEM-EDX).

Results and discussion

Phase identification and morphology

All synthesized materials showed well-defined strong and sharp reflections in their XPD patterns (Fig. 1), confirming the good crystallinity and purity of the products. For the phosphors, high crystallinity is important as it means less energy traps and strong luminescence. It has been reported that the hexagonal NaRF₄ is a by far better host lattice for high UC emission than the cubic phase.³¹ All XPD patterns agreed well with the hexagonal NaYF₄ reference pattern indicating that increased Yb³⁺ and Tm³⁺ concentrations did not induce phase transitions in the NaY_1−xYb_xF₄:Tm (0.5%), and NaYbF₄:Tm (x%) crystals.³²

Fig. 1 XPD patterns of hexagonal NaY_1−xYb_xF₄:Tm (0.5%) with different concentrations of Yb³⁺ (top) and NaYbF₄:Tm (x%) crystals with different concentrations of Tm³⁺ (bottom), along with a reference pattern calculated with PowderCell³³ based on structural data reported.³² The additional reflections are the traces of * sodium chloride (NaCl) and ^§ cubic NaRF₄ impurities.

The shape and size of the synthesized crystals were further characterized with transmission electron microscopy (TEM). Fig. 2 shows the typical TEM images of as prepared NaY_1−xYb_xF₄:Tm (0.5%) with different concentrations of Yb³⁺: (a) 20, (b) 40, (c) 60, (d) 80, (e) 100% and NaYbF₄:Tm (x%) crystals with different concentrations of Tm³⁺: (f) 0.05, (g) 0.2, (h) 0.4, (i) 0.6, (j) 0.8, (k) 1 and (l) 5%. The increase or decrease in rare earth ion concentrations with respect to each other does not affect the crystal structure, while in case of crystal size and shape the Yb³⁺ and Y³⁺ concentrations play a vital role. UCNPs with Yb³⁺ concentration up to 40% have small spherical shape which changes to hexagonal plate when Yb³⁺ concentration is increased further. On the other hand increase or decrease in the Tm³⁺ ion concentration up to 0.6% keeps the particle size unaffected but further increase in the activator concentration gives also small UCNPs as a byproduct.

Fig. 2 TEM images of hexagonal NaY_1−xYb_xF₄:Tm (0.5%) with different concentrations of Yb³⁺: (a) 20, (b) 40, (c) 60, (d) 80, (e) 100% and NaYbF₄:Tm (x%) crystals with different concentrations of Tm³⁺: (f) 0.05, (g) 0.2, (h) 0.4, (i) 0.6, (j) 0.8, (k) 1 and (l) 5%.

As one can see from the TEM images, at low sensitizer ion concentration, β-NaYF₄: 20% Yb, 0.5% Tm nanoparticles appear nearly spherical in shape and monodisperse. The average diameter of the 20% Yb³⁺ doped nanoparticles is around 25 nm (Fig. 2a). When the Yb³⁺ concentration increased further to 40%, the particles grew up to 90 nm in all directions, keeping the spherical shape unchanged (Fig. 2b). At medium Yb³⁺ ion concentration, hexagonal plates with sharp edges can be prepared, which is shown in Fig. 2c and the average diameter calculated is about 150 nm. When the experiment is performed at 80% Yb³⁺ concentration, most of the Y³⁺ ions are already replaced by Yb³⁺ resulting in bigger hexagonal plates with a diameter of about 900–950 nm, which are basically polycrystalline plates made up from many small plates (Fig. 2d) as indicated by the Scherrer calculations. In the case of the NaYbF₄:Tm (0.5%) crystals, where all Y³⁺ was replaced by Yb³⁺, the biggest polycrystalline hexagonal plates of about 1000 nm are formed (Fig. 2e). According to the Scherrer calculations, these are also polycrystalline. The low resolution TEM images showing homogeneity and size distributions of the particles are presented in the ESI, Fig. S6.†

For the NaYbF₄ crystals with different Tm³⁺ ion concentrations, the shape and size of the particles remains the same until the 0.6% Tm³⁺ doping level, the average diameters of all hexagonal plates being between 1000 and 1200 nm (Fig. 2f–i). The Tm³⁺ ion concentration of 0.8% induces the formation of small spherical particles along with the bigger hexagonal plates. The diameter of the large plates is about 750–800 nm and that of the small particles is up to 10 nm (Fig. 2j). The SEM-EDX analysis of both small and large particles shows that the elemental composition is the same in both size fractions (results of 0.8% Tm³⁺ sample are presented in the ESI, Fig. S5†). For small particles, the Tm³⁺ is not detectable by the instrument as the concentration of Tm³⁺ is very small in the sample. A further increase in the Tm³⁺ concentration results in an increase in the population of small particles with about 25 nm in diameter along with few large hexagonal plates of about 900 nm (Fig. 2k and l). The effect of the Yb³⁺ and Tm³⁺ concentrations on the crystal size, shape and structure are listed in Tables 1 and 2.

Table 1 Summary of the morphologies, structure and dimensions obtained with TEM for the NaY_1−xYb_xF₄:Tm (0.5%) crystals

Yb^3+ (mol%)^a	Y^3+ (mol%)^a	Structure	Morphology	Size
a The percentages are with respect to the added quantity in the synthesis.
20	80	Hexagonal	Nanosphere	25
40	60	Hexagonal	Nanosphere	≈90
60	40	Hexagonal	Hexagonal plates	≈150
80	20	Hexagonal	Hexagonal plates	900–950
100	0	Hexagonal	Hexagonal plates	950–1000 and ≈20

---

Table 2 Summary of the morphologies, structure and dimensions obtained with TEM for the NaYbF₄:Tm (x%) crystals

Tm^3+ (mol%)^a	Structure	Morphology	Size
a The percentages are with respect to the added quantity in the synthesis.
0.05	Hexagonal	Hexagonal plates	≈950
0.2	Hexagonal	Hexagonal plates	≈1000–1200
0.4	Hexagonal	Hexagonal plates	900–950
0.6	Hexagonal	Hexagonal plates	900–950
0.8	Hexagonal	Hexagonal plates	800 and ≤10
1	Hexagonal	Hexagonal plates	900 and ≈25
5	Hexagonal	Hexagonal plates	900 and ≈25

---

Crystal growth mechanisms

The experimental results show that the substitution of Y³⁺ in the host lattice by Yb³⁺ has a clear effect on the particle morphology and size. To investigate the formation process the sensitizer concentrations of NaY_1−xYb_xF₄:Tm (0.5%) were increased stepwise by keeping all other synthesis parameters constant.

Regarding the growth mechanism, it is important to consider also the relation between the reaction temperature, initially formed crystals and the role of the chelating agent. It is known that the reaction temperature plays a crucial role in the α → β phase transition of NaYF₄ and also the concentration of Yb³⁺ embedded in NaYF₄ matrix affects the temperature of the phase transition in such a way that the temperature decreases with increasing the Yb³⁺ content.³⁴ In the present work, 310 °C used is quite high for NaRF₄ with high Yb³⁺ content and this results in large particles. Fig. 3 shows the possible formation mechanism for the β-NaRF₄:Tm crystals with different Y/Yb concentrations.

Fig. 3 Schematic representation of formation procedure of β-NaRF₄,Tm (0.5%) with different Yb³⁺ concentrations. Inset is the schematic diagram showing the anisotropy of the β-NaRF₄ crystals.

In the actual process, after dissolution and renucleation the unstable α-NaRF₄ phase is transformed to the stable β-NaRF₄. As the Yb³⁺ concentration in the reaction mixture increases the formation of the initial β-NaRF₄ crystals will be faster and there is more time for crystal growth. This is because the increased Yb³⁺ concentration lowers the phase transition temperature.³⁵ The nanocrystal formation mechanism for the UCNPs with 20 and 40% Yb³⁺ is a one step process and the phase transition from α → β was completed somewhat below 310 °C, which finally gives uniform 25 and 90 nm nanocrystals respectively (Fig. 3a and b). At these lower Yb³⁺ concentrations the growth rates in all directions are comparable which results in the spherical shape.

For the 60, 80 and 100% Yb³⁺ samples, the α → β phase transformation is completed well below 310 °C. As the initially formed β-NaRF₄ has the hexagonal shape with an anisotropic structure, the crystal growth is anisotropic, as well. The hexagonal structure has four axes; three assistant axes intersecting each other at 60 degrees and one principle axis perpendicular to the hexagonal (001) crystal surface (Fig. 3, inset).^36,37 The growth mechanism of the crystal depends on the growth rates along the different crystal axes which results into the different shapes of the final crystals. Once the hexagonal β-NaRF₄ is formed, the oleic acid (OA) starts to show its influence as shape modifier. OA strongly reacts with the top/bottom (001) surfaces of the growing crystallites and lowers their surface energy, which results into the fast growth along the assistant axes rather than the principle axis. For the 60% Yb³⁺ sample, this anisotropic growth results into 130–150 nm hexagonal nanoplates (Fig. 3c). For the 80 and 100% Yb³⁺ samples, the temperature during the second heating may assist the formed hexagonal plates to self-assemble and anneal into the larger hexagonal polycrystalline structures with about 1000 nm size (Fig. 3d and e). This hypothesis of self-assembly of the nanoplates into larger polycrystals was supported by the calculations of diffracting domain sizes from the XPD patterns (Table 3). The data shows that, for the 80 and 100% Yb³⁺ samples, the hexagonal particles formed at the end of reaction are polycrystalline made up by aggregation of 130–150 nm nanoplates.

Table 3 Summary of decay times from the emitting ³H₄, ¹G₄ and ¹D₂ level of NaY_1−xYb_xF₄:Tm (0.5%) and NaYbF₄:Tm (x%) with width to length (W/L) ratio and morphology calculated with the Scherrer method^a

NaY1−xYbxF4:Tm (0.5%) x mol%^#	W (nm)	L (nm)	W/L	Morphology	Decay times (μs)^§
					^3H4 → ^3H6	^1G4 → ^3H6	^1D2 → ^3H6
a Note: * the crystals are too large for Scherrer method, which commonly can treat sizes up to 150 nm, ^§ the decay time with the highest amplitude in case of multiexponential decay, ^# the percentages are with respect to the added quantity in the synthesis.
20	27	24	1.12	Spherical	630 ± 8	576 ± 2	168 ± 0.3
40	30	28	1.04	Spherical	422 ± 5	547 ± 2	142 ± 0.3
60	53	34	1.55	Hexagonal plate	378 ± 7	224 ± 2	138 ± 0.3
80	56	35	1.59	Hexagonal plate	440 ± 5	362 ± 2	157 ± 0.4
100	149	79	1.89	Hexagonal plate	375 ± 1	293 ± 4	137 ± 0.3

---

NaYbF4:Tm (x%) x mol%^#	W (nm)	L (nm)	W/L	Morphology	Decay times (μs)^§
					^3H4 → ^3H6	^1G4 → ^3H6	^1D2 → ^3H6
0.05	212*	104	2.04	Hexagonal plate	481 ± 4	510 ± 5	272 ± 1
0.2	117	68	1.72	Hexagonal plate	470 ± 5	427 ± 0.5	235 ± 2
0.4	92	70	1.31	Hexagonal plate	430 ± 7	358 ± 1	169 ± 0.5
0.6	125	89	1.40	Hexagonal plate	321 ± 5	256 ± 3	125 ± 76
0.8	146	115	1.27	Hexagonal plate	337 ± 2	169 ± 2	98 ± 0.2
1	1011*	95	—	Hexagonal plate	275 ± 2	148 ± 0.5	82 ± 0.3
5	325*	162	—	Hexagonal plate	70 ± 0.5	82 ± 3	80 ± 10

---

Upconversion luminescence properties

To investigate the optical properties, the UC photoluminescence of colloidal NaY_1−xYb_xF₄:Tm (0.5%) and NaYbF₄:Tm (x%) crystals (0.1 mg ml⁻¹ suspension in oleic acid) was measured under 980 nm diode laser excitation. As shown in the Fig. 4a and b, six UC photoluminescence lines were clearly resolved, centered at 346, 365, 450, 475, 649 and 800 nm, which correspond to the ¹I₆ → ³F₄, ¹D₂ → ³H₆, ¹D₂ → ³F₄, ¹G₄ → ³H₆, ¹G₄ → ³F₄, and ³H₄ → ³H₆ transitions of Tm³⁺, respectively (non-normalized intensity spectra are in the ESI, Fig. S1 and S4†). The weak emission at 511 nm arises from the transition of ¹D₂ → ³H₅ and the lines at 690 and 720 nm correspond to the second order of 345 and 365 nm.³⁸

Fig. 4 Upconversion luminescence spectra of colloidal (a) NaY_1−xYb_xF₄:Tm (0.5%) (x_Yb = 20, 40, 60, 80 and 100%) and (b) NaYbF₄:Tm (x%) (x_Tm = 0.05, 0.2, 0.4, 0.6, 0.8, 1 and 5%) crystals under 980 nm excitation in oleic acid suspension, normalized at the emission peak of 800 nm. The inset shows the intensity variation of all major emissions as a function of (a) Yb³⁺ and (b) Tm³⁺ concentrations.

The energy levels of Yb³⁺ and Tm³⁺ ions as well as the proposed UC pathways under excitation at 980 nm are shown in Fig. 5. After 980 nm excitation, the sensitizer ion Yb³⁺ transfers energy to the neighbouring Tm³⁺ ions. Three successive energy transfers from Yb³⁺ to Tm³⁺ populate the ³H₅, ³F₂ and ¹G₄ levels of Tm³⁺. These energy transfer processes along with some non-radiative relaxation are responsible for the emissions at 475, 650, 690 and 800 nm. In the samples with lower Tm³⁺ content (0.05%) the simple piling of NIR photons (black thick arrows shown in the Fig. 5) is the only way to populate the ¹D₂ level. It is however not a very efficient process, because of the large energy mismatch between the energy difference of the ¹D₂ and ¹G₄ levels and the NIR photon energy. In the samples with Tm³⁺ content up to 0.8%, the cross relaxation ³F₂ + ³H₄ → ³H₆ + ¹D₂ between the two Tm³⁺ ions plays an important role in populating the ¹D₂ level. The radiative relaxation from the ¹D₂ level gives the 365 and 450 nm emissions. Further increase in Tm³⁺ concentration, however, results in the depopulation of the higher emitting levels due to self-quenching. The Tm³⁺ ions in the ¹D₂ level can be excited to the ³P₂ level by the fourth energy transfer mechanism from Yb³⁺ ions which relax rapidly to the ¹I₆ level giving another UV band centered at 346 nm.

Fig. 5 Energy level diagrams of the Yb³⁺ and Tm³⁺ ions as well as the proposed UC mechanism for UV and visible emissions along with quenching mechanism by cross relaxation.

To increase the efficiency of the UV UC photoluminescence, we gradually increased the amount of the sensitizer ions (Yb³⁺), replacing Y³⁺ in the host material. As seen from the Fig. 4a, in hexagonal NaY_1−xYb_xF₄ crystals with fixed Tm³⁺ concentration at 0.5%, all the emission intensities were increased stepwise along with the increase of the concentration of Yb³⁺ from 20% to 100%. The maximum UV emission was observed with 100% Yb³⁺ and the UC intensities at 346 and 365 nm emissions were measured to be about 20 and 18 times higher than that of the 20% Yb³⁺ sample, respectively. Along with the increase in the UV emission, the visible emission centered at 450 nm was also increased 30 times. The enhancement in the UV and visible emissions partially come from the fact that, with higher Yb³⁺ concentration the particle size increases, which decreases the surface to volume ratio. This decreases the amount of surface quenching centers and thus facilitates the high UC emission. The increased particle size, however, does not explain the whole enhancement mechanism as the decrease in surface to volume ratio after ≈150 nm particle size has no such strong influence on the UC spectra as reported before.³⁹ Along with particle size there are thus other important factors which contribute to the intense UV and visible emissions. The increased Yb³⁺ concentration (keeping Tm³⁺ concentration fixed at 0.5%) increases the sensitization by absorbing more NIR photons and enhances the Yb-to-Yb energy migration as well as the Yb-to-Tm energy transfer probabilities, which leads to the increase of the population at ¹I₆ and ¹D₂ levels of the Tm³⁺ ions. The inset of Fig. 4a shows the variation of the different emission intensities as a function of Yb³⁺ concentration. Increasing the Yb³⁺ concentration beyond 60%, resulted in only small enhancement of the 475 and 800 nm emissions (4.5 and 2.7 times compared to 80% and 20% Yb samples) is observed. The increased sensitization by Yb³⁺ and the higher probability of energy migration due to shorter Yb-to-Yb distances can increase the probability of UV emission from Tm³⁺ which can ultimately result in the saturation of the emission from the lower emitting levels.²⁶

The effect of the Tm³⁺ concentration on UV photoluminescence for the NaYbF₄ crystals was also investigated using the UC emission spectra of NaYbF₄:Tm (x_Tm = 0.05, 0.2, 0.4, 0.6, 0.8, 1 and 5%) crystals (Fig. 4b). The UV emission intensity increases as the Tm³⁺ concentration increases from 0.05 to 0.4%; it remains steady in the range of 0.4 to 0.8% and decreases with further increase in Tm³⁺ concentration. From the inset in Fig. 4b it is clear that the optimal range of the Tm³⁺ concentration for highest UV and visible emission is between 0.4 and 0.8%. The reason for such a dramatic decrease in the UV and visible emission intensities with higher Tm³⁺ concentration can be the self-quenching of the Tm^{3+ 1}D₂ and ¹G₄ levels.⁴⁰ There are at least three different cross relaxation processes i.e. ¹D₂ + ³H₆ → ³F₃ + ³H₄, ¹D₂ + ³F₃ → ¹G₄ + ¹G₄ and ¹D₂ + ³H₆ → ³H₄ + ³F₂ which are responsible for depopulating the UV emitting ¹D₂ energy level in samples with higher Tm³⁺ content (thick black doted arrows shown in Fig. 5). Two out of these three cross relaxation processes feed the ³H₄ energy level which may explain the enhancement in the 800 nm emission with increase in the Tm concentrations up to 0.8%.

In Fig. 6, the ratio of the 365/800 nm emission intensities were plotted at different (a) Yb³⁺ and (b) Tm³⁺ concentrations. The 365/800 nm intensity ratio from the NaY_1−xYb_xF₄:Tm (0.5%) samples increases with Yb³⁺ concentration (Fig. 6a) and the highest ratio is obtained with 100% Yb³⁺ (the most excessive sensitization). From the NaYbF₄:Tm (x%) samples the highest ratio is obtained with 0.6% Tm³⁺ concentration (Fig. 6b).

Fig. 6 Dependence of the upconversion luminescence intensity ratio of 365/800 nm emissions on the (a) Yb³⁺ and (b) Tm³⁺ doping concentrations.

The emission decays of the ¹D₂ → ³H₆ (365 nm), ¹G₄ → ³H₆ (475 nm) and ³H₄ → ³H₆ (800 nm) transitions after wide-pulse excitation at 980 nm for the NaY_1−xYb_xF₄:Tm (0.5%) materials with 20–100% Yb and NaYbF₄ with 0.05–5% Tm are shown in the Fig. 7 (non-normalized intensity decay profiles are shown in the ESI Fig. S2 and S3†). The decay data were fitted with single, second or third order exponential decay functions. All the lifetimes for the different Yb³⁺ and Tm³⁺ concentrations are listed in Tables 3 and S1 in the ESI.† Usually the bi-exponential decays result from different pathways used to obtain the same emission. When the Tm³⁺ concentration was fixed at 0.5%, both the rise and decay time for all the emissions were found to decrease with increasing Yb³⁺ concentration (Fig. 7a–c). The same trend was observed in case of the NaYbF₄ samples with different Tm³⁺ concentrations (Fig. 7d–f). This indicates that high sensitization is responsible for the efficient energy transfer from Yb to Tm, which shortens the energy transfer time and decreases both rise and decay times. In addition to this, one possible reason for the short decay time in case of higher dopant concentration is the decrease in the distance between the luminescence centers and the surface related high frequency vibrational modes which causes the non-radiative vibrational processes to be more efficient.

Fig. 7 Upconversion luminescence rise and decay profiles for (a–c) NaY_1−xYb_xF₄:Tm (0.5%) and (d and e) NaYbF₄:Tm (x%) UCNPs at (a/d) 800, (b/e) 475 and (c/f) 365 nm emissions upon 980 nm excitation for 0–3000 μs.

The noticeable decrease in the lifetime of the emitting ³H₄ level with increase in Yb³⁺ and Tm³⁺ concentration can be partially explained by the Tm → Yb energy back transfer and self-quenching occurring among the Tm³⁺ ions.⁴¹

Previously, Shan et al. and Lim et al. have demonstrated the role of particle shape on the fluorescence decay times of UCNPs.^42–44 As per that discussion, crystals with different width to length (W/L) aspect ratios should have different phonon relaxation processes which strongly affect on the fluorescence decay. From our experiments it is observed that, in case of samples with higher Yb content where hexagonal plates have higher W/L ratios, the observed decays are lower, i.e. the W/L ratio and decay time are inversely proportional to each other. It is also interesting to notice that the NaYbF₄:Tm (5%) sample shows persistent upconversion luminescence for the 800 nm emission (Fig. 7d), i.e. a prolonged emission, similar to what has been reported for ZrO₂:Yb³⁺, Er³⁺ earlier.⁴⁵ Moreover, the same sample shows the sharp peak in the rise curves of the 365 and 475 nm emissions indicating optically stimulated luminescence. The explanation for these two phenomena is out of the scope of the present report, however, and thus they will be discussed in another report.

Conclusions

Improved NIR to UV upconversion is reported with high sensitizer concentration in NaYbF₄:Tm (0.5%). It was found that the increase in the Yb³⁺ concentration, with corresponding decrease in Y³⁺ concentration resulted in the enhanced UV emission and improved UV (340–365 nm) vs. NIR (800 nm) emission intensity ratio. The enhancement arises from the increased absorption by the Yb³⁺ ions and the increased efficiency of the Yb-to-Yb energy migration and the Yb–Tm energy transfers. The NaYbF₄ host with 0.4 to 0.8% Tm³⁺ is the optimal composition for UV UC photoluminescence. The effect of different Yb³⁺ concentrations on the particle growth mechanism was also studied and it was observed that higher Yb³⁺ content will result in self-aggregation and sintering of the hexagonal plates into larger polycrystalline plate at the end of the reaction. Upconversion emission decay time analyses of UCNPs with different Yb³⁺ and Tm³⁺ concentrations at all major emissions was also included. The decrease in the decay time of the highly sensitized UCNPs with high W/L ratio indicates an efficient energy transfer between Yb³⁺ and Tm³⁺. In prospect, this work provided deeper understanding of the effect of the activator and sensitizer concentrations on photon upconversion resulting in a simple way to obtain high intensity NIR to UV upconversion for use in optical, chemical and biological applications.

Acknowledgements

The authors gratefully acknowledge the financial support from the Research Executive Agency (REA) of the European Union under Grant Agreement number PITN-GA-2010-264772 (ITN CHEBANA).

Notes and references

1. S. Ghosh, T. F. Rosenbaum, G. Aeppli and S. N. Coppersmith, Nature, 2003, 425, 48.
2. G. Yi, H. Lu, S. Zhao, Y. Ge, W. Yang, D. Chen and L.-H. Guo, Nano Lett., 2004, 4, 2191.
3. H. Schäfer, P. Ptacek, K. Kömpe and M. Haase, Chem. Mater., 2007, 19, 1396.
4. S. F. Felix, E. A. Gouveia, M. T. de Araujo, A. S. B. Sombra and A. S. Gouveia-Neto, J. Lumin., 2000, 87 and 89, 1020.
5. F. van de Rijke, H. Zijlmans, S. Li, T. Vail, A. K. Raap, R. S. Niedbala and H. J. Tanke, Nat. Biotechnol., 2001, 19, 273.
6. J. Ryu, H.-Y. Park, K. Kim, H. Kim, J. H. Yoo, M. Kang, K. Im, R. Grailhe and R. Song, J. Phys. Chem. C, 2010, 114, 21077.
7. J. de Wild, A. Meijerink, J. K. Rath, W. G. J. H. M. van Sark and R. E. I. Schropp, Sol. Energy Mater. Sol. Cells, 2010, 94, 1919.
8. J. de Wild, J. K. Rath, A. Meijerink, W. G. J. H. M. van Sark and R. E. I. Schropp, Sol. Energy Mater. Sol. Cells, 2010, 94, 2395.
9. M. Liu, Y. Lu, Z. B. Xie and G. M. Chow, Sol. Energy Mater. Sol. Cells, 2011, 95, 800.
10. D. Chen, Y. Wang and M. Hong, Nano Energy, 2012, 1, 73.
11. S. A. Hilderbrand, F. Shao, C. Salthouse, U. Mahmood and R. Weissleder, Chem. Commun., 2009, 4188.
12. Q. Liu, Y. Sun, T. Yang, W. Feng, C. Li and F. Li, J. Am. Chem. Soc., 2011, 133, 17122.
13. A. D. Ostrowski, E. M. Chan, D. J. Gargas, E. M. Katz, G. Han, P. J. Schuck, D. J. Milliron and B. E. Cohen, ACS Nano, 2012, 6, 2686.
14. F. Wang, Y. Han, C. S. Lim, Y. Lu, J. Wang, J. Xu, H. Chen, C. Zhang, M. Hong and X. Liu, Nature, 2010, 463, 1061.
15. C. Liu, H. Wang, X. Zhang and D. Chen, J. Mater. Chem., 2009, 19, 489.
16. J. Choi, T.-K. Tseng, M. Davidson and P. H. Holloway, J. Mater. Chem., 2011, 21, 3113.
17. G. Han, T. Mokari, C. Ajo-Franklin and B. E. Cohen, J. Am. Chem. Soc., 2008, 130, 15811.
18. H.-M. Lee, D. R. Larson and D. S. Lawrence, ACS Chem. Biol., 2009, 4, 409.
19. K. Szaciłowski, W. Macyk, A. Drzewiecka-Matuszek, M. Brindell and G. Stochel, Chem. Rev., 2005, 105, 2647.
20. X. Chen and Z. Song, J. Opt. Soc. Am. B, 2007, 24, 965.
21. S. Beyazit, S. Ambrosini, N. Marchyk, E. Palo, V. Kale, T. Soukka, B. Tse Sum Bui and K. Haupt, Angew. Chem., Int. Ed., 2014, 53, 8919.
22. D. Chen, Y. Wang, Y. Yu and P. Huang, Appl. Phys. Lett., 2007, 91, 051920.
23. V. Mahalingam, F. Vetrone, R. Naccache, A. Speghini and J. A. Capobianco, Adv. Mater., 2009, 21, 4025.
24. F. Shi, J. Wang, D. Zhang, G. Qin and W. Qin, J. Mater. Chem., 2011, 21, 13413.
25. G. Wang, W. Qin, L. Wang, G. Wei, P. Zhu and R. Kim, Opt. Express, 2008, 16, 11907.
26. J. Shen, G. Chen, T. Y. Ohulchanskyy, S. J. Kesseli, S. Buchholz, Z. Li, P. N. Prasad and G. Han, Small, 2013, 19, 3213.
27. S. Zeng, G. Ren, W. Li, C. Xu and Q. Yang, J. Phys. Chem. C, 2010, 114, 10750.
28. Y. Y. Zhang, L. W. Yang, C. F. Xu, J. X. Zhong and C. Q. Sun, Appl. Phys. B: Lasers Opt., 2010, 98, 243.
29. V. Kale, T. Soukka, J. Hölsä and M. Lastusaari, J. Nanopart. Res., 2013, 15, 1.
30. H. P. Klug and L. E. Alexander, X-ray Powder Diffraction Procedures, Wiley, New York, NY, USA, 1959.
31. K. W. Krämer, D. Biner, G. Frei, H. U. Güdel, M. P. Hehlen and S. R. Lüthi, Chem. Mater., 2004, 16, 1244.
32. E. Harju, I. Hyppänen, J. Hölsä, J. Kankare, M. Lahtinen, M. Lastusaari, L. Pihlgren and T. Soukka, Zeitschrift für Kristallographie Proceedings, 2011, 381.
33. W. Kraus and G. Nolze, PowderCell for Windows v. 2.4, Federal Institute for Materials Research and Testing, Berlin, Germany, 2000.
34. T. Laihinen, M. Lastusaari, L. Pihlgren, L. V. Rodrigues and J. Hölsä, J. Therm. Anal. Calorim., 2015 DOI:10.1007/s10973-015-4609-x.
35. N. C. Dyck, F. C. J. M. van Veggel and G. P. Demopoulos, ACS Appl. Mater. Interfaces, 2013, 5, 11661.
36. W. Qingqing, X. Gang and H. Gaorong, Cryst. Growth Des., 2006, 6, 1776.
37. X. Liang, X. Wang, J. Zhuang, Q. Peng and Y. Li, Adv. Funct. Mater., 2007, 17, 2757.
38. H. Qiu, C. Yang, W. Shao, J. Damasco, X. Wang, H. Ågren, P. Prasad and G. Chen, Nanomaterials, 2014, 4, 55.
39. A. Yin, Y. Zhang, L. Sun and C. Yan, Nanoscale, 2010, 2, 953.
40. F. W. Ostermayer, J. P. van der Ziel, H. M. Marcos, L. G. Van Uitert and J. E. Geusic, Phys. Rev. B: Solid State, 1971, 3, 2698.
41. A. Braud, S. Girard, J. L. Doualan, M. Thuau, R. Moncorgé and A. M. Tkachuk, Phys. Rev. B: Condens. Matter Mater. Phys., 2000, 61, 5280.
42. S. F. Lim, W. S. Ryu and R. H. Austin, Opt. Express, 2010, 18, 2309.
43. J. Shan, M. Uddi, N. Yao and Y. Ju, Adv. Funct. Mater., 2010, 20, 3530.
44. J. Shan, M. Uddi, R. Wei, N. Yao and Y. Ju, J. Phys. Chem. C, 2010, 114, 2452.
45. L. Pihlgren, T. Laihinen, L. C. V. Rodrigues, S. Carlson, K. O. Eskola, A. Kotlov, M. Lastusaari, T. Soukka, H. F. Brito and J. Hölsä, Opt. Mater., 2014, 36, 1698.

---

Footnote

† Electronic supplementary information (ESI) available: Non-normalized intensity spectra and lifetime decay of the NaY_1−xYb_xF₄:Tm (0.5%) and NaYbF₄:Tm (x%) nanocrystal dispersion in oleic acid pumped by 980 nm laser. Table with the summary of decay times of ³H₄, ¹G₄ and ¹D₂ level of NaY_1−xYb_xF₄:Tm (0.5%) and NaYbF₄:Tm (x%) samples with respective amplitude. Low resolution TEM images of the bigger UCNPs. SEM-EDX analysis results of NaYbF₄:Tm (0.8%) nanocrystals. See DOI: 10.1039/c5ra01613d

---