Spectral shifting and NIR down-conversion in Bi³⁺/Yb³⁺ co-doped Zn₂GeO₄

Abstract

We report on spectral modification through NIR down-conversion (DC) photoluminescence (PL) in Yb³⁺-Bi³⁺-co-doped Zn₂GeO₄. Energetic downshifting (DS) of UV-A irradiation occurs via intrinsic luminescence of the high-bandgap semiconductor Zn₂GeO₄ as well as via active Bi³⁺ centres. In parallel, both species act as sensitizers for Yb³⁺, strongly extending its excitation region to ∼500 nm. In the absence of Bi³⁺, band-to-band absorption of Zn₂GeO₄ in the UV region results in PL at ∼475–625 nm. Doping with Yb³⁺ initiates energy transfer from trapped defect states to two neighbouring Yb³⁺ ions in a cooperative DC process, resulting in Yb³⁺-related photoemission at ∼1000 nm. The introduction of Bi³⁺ into Zn₂GeO₄:Yb³⁺ greatly extends the absorption band to the visible blue region. Then, energy transfer also occurs through cooperative DC from Bi³⁺ to Yb³⁺. As a result, a strong increase in the absolute Yb³⁺-related PL intensity is observed. This enables ultra-efficient harvesting of UV-A to visible radiation for energy conversion processes.

---

Introduction

The spectral conversion of incident sunlight through photoluminescence has been proposed to enhance the efficiency of various solar energy harvesting processes.^1–7 For example, in the case of crystalline silicon solar cells, a large fraction of the incoming solar energy is lost either because photons do not surpass the energy gap which is necessary to generate a photoelectron (NIR-tail of the solar spectrum), or because their excess energy causes thermalization or electron–hole recombination (near-UV to green part of the solar spectrum). Conversion strategies consequently follow some straightforward schemes: for example, in up-conversion processes (UC), two or more low-energy photons are used to generate one photon of higher energy. In down-shifting (DS) or quantum cutting (QC) processes, on the other side, a high-energy photon is converted into one (DS) or more (QC) photons of lower energy.^1,2,8–12 In the latter, a quantum efficiency which is theoretically higher than unity can be achieved. As a concept, all three processes enable to adjust the incoming solar spectrum through increasing the number of photons within a certain spectral region. Whether or not this is advantageous for a specific solar energy conversion process, however, depends on many other factors, including quantum efficiency, secondary absorption processes, converter properties, system design and cost.

Here, we concentrate on the model system of Zn₂GeO₄:Yb³⁺ and its further sensitization with Bi³⁺ co-dopants. Yb³⁺ has a rather simple band structure with the single excited state of ²F_5/2 at an energy of ∼1 eV. Yb³⁺ photoluminescence therefore occurs at a wavelength of ∼1000 nm, which is very close to the maximum photoelectronic conversion efficiency of silicon. Hence, trivalent ytterbium is an often-sought emitter for solar spectral conversion through QC or DS.^1–4 For this, a suitable sensitizer is required which adds favourable absorption properties to the emission behaviour of Yb³⁺. Most of the trivalent rare earth (RE³⁺) ions, e.g., Pr³⁺, Er³⁺, Nd³⁺, Ho³⁺, Tb³⁺, Tm³⁺ and Dy³⁺ have been considered for this purpose.^1,2,8–19 However, the parity-forbidden 4fⁿ → 4fⁿ transitions in these ions typically result in weak and narrow absorption bands which stand in contrast to the wish for efficient (broadband) harvesting of sunlight. As alternatives, ions such as Ce³⁺, Eu²⁺, Bi³⁺ and Mn²⁺, and/or combination with intrinsic absorption of the host itself (e.g., YVO₄ and ZnO) have therefore been studied for Yb³⁺ activation.^20–24

The orthogermanate of zinc, Zn₂GeO₄, is a wide-bandgap semiconductor with a gap-energy of ∼4.68 eV.²⁵ Without further dopants, it exhibits broadband luminescence in the blue to green spectral region which originates from the recombination of native structural defects.^26,27 Emission from Zn₂GeO₄ occurs at about twice the energy which is necessary to excite Yb³⁺: ²F_7/2 → ²F_5/2. This means that it can be used to sensitize Yb³⁺. On the other hand, absorption covers only the UV region of ∼250 to 350 nm, which prevents efficient solar harvesting. To overcome this, we introduce trivalent bismuth as further activator into the Zn₂GeO₄ host. Bi³⁺ has a 6s² electronic configuration with ground state ¹S₀ and the excited state 6s6p which splits into the ³P₀, ³P₁, ³P₂ and ¹S₁ levels. The emission band of Bi³⁺ is typically located in the blue or green, originating from the ligand-field dependent relaxation of ³P₁ → ¹S₀.^28–36 Although the absorption band of ¹S₀ → ³P₁ is spin-forbidden, it shows reasonably high oscillator strength due to spin–orbit coupling between the ³P₁ and the ¹P₁ level.^12,35 Similar to Zn₂GeO₄, the absorbed energy of Bi³⁺ can probably be transferred to Yb³⁺ through DC processes. Introduction of Bi³⁺ into Zn₂GeO₄:Yb³⁺ is therefore expected to enable broadband activation of Yb³⁺ NIR photoluminescence.

Experimental section

Powder samples and compacted pellets with nominal compositions of Zn_1.998−xGeO₄:Yb_0.002³⁺Bi_x³⁺ (x = 0, 0.006, 0.02, 0.06 and 0.18 and is the number of Zn²⁺ which is displaced by Bi³⁺ in Zn₂GeO₄) were synthesized through conventional high-temperature solid state reaction. For that, ZnO, GeO₂, Bi₂O₃ and Yb₂O₃ (≥4 N) were used as starting materials. Stoichiometric batches of ∼2 g were thoroughly mixed in an agate mortar, pre-calcined at 900 °C for 4 h in air, again ground in an agate mortar, pressed into pellets and finally fired in air at 1300 °C for 12 h.²⁰ Blank (Zn₂GeO₄) and singly-doped (Zn_1.998GeO₄:Bi_0.02) samples were prepared in the same way for reference.

Powder X-ray diffraction (XRD) patterns were obtained on a Siemens Kristalloflex D500 diffractometer over a 2θ range of 10–70°. UV-vis diffuse reflectance (DR) spectra were recorded with a double-beam photospectrometer (Cary 5000) over the spectral range of 200 to 800 nm with 1 nm step size. Static photoluminescence (PL) and PL excitation (PLE) spectra and dynamic decay curves were collected at room temperature with a high-resolution spectrofluorometer (Horiba Jobin Yvon Fluorolog FL3-22), using a static 450 W Xe lamp and a pulsed 75 W Xe flashlamp as excitation sources, respectively. NIR PL was observed with an InP/InGaAs-based thermoelectrically cooled photomultiplier tube (NIR-PMT, Hamamatsu H10330A-75). PLE spectra were measured between 250 to 500 nm with a step size of 1 nm. PL spectra were recorded between 390 and 700, and 930 and 1200 nm with a step size of 1 nm. PLE spectra were corrected over the lamp intensity with a silicon photodiode. PL spectra were corrected with the spectral response of employed PMT.

Results and discussion

Fig. 1a displays ex situ powder XRD patterns of as-prepared Zn_1.998−xGeO₄:Yb_0.002³⁺Bi_x³⁺ for varying values of x. As the major crystal phase in all samples, we identify rhombohedral Zn₂GeO₄ (JCPDS card no. 00-013-0687). A minor amount of secondary Bi₄(GeO₄)₃ (cubic I4̄3d, JCPDS card no. 01-084-0505) can also be indexed at x > 0.06. The unit cell crystal structure of Zn₂GeO₄ and Bi₄(GeO₄)₃ are shown in Fig. 1b and c, respectively. The crystal structure of willemite-type Zn₂GeO₄ belongs to space group R3̄(no. 148). It is comprised of corner-sharing [ZnO₄] and [GeO₄] tetrahedra (Fig. 1b).^20,25 While the incorporation of Bi³⁺ into the crystal lattice is expected to result in a change of lattice parameters and in decreasing symmetry (assumedly due to the parallel formation of oxygen vacancies which is necessary for charge compensation), we do not observe a notable shift in the diffraction peaks. This indicates the very low solubility of Bi³⁺ ions in Zn₂GeO₄ which can be understood on the basis of the large difference in ionic radii between Bi³⁺ and Zn²⁺ or Ge⁴⁺. The ionic radii of Zn²⁺ and Ge⁴⁺ in fourfold coordination are 0.60 and 0.39 Å, respectively.³⁷ There is no clear data available on fourfold-coordinated Bi³⁺, though. For fivefold coordination, it is 0.96 Å, a significant 60% higher than the ^IVZn²⁺ radius. The appearance of the secondary crystal phase is a consequence of this low solubility of Bi³⁺ ions in Zn₂GeO₄. Eulytite-type Bi₄(GeO₄)₃ is composed of [BiO₆] octahedra and [GeO₄] tetrahedra (Fig. 1c).³⁸ It is known as scintillator material and – due to the ^VIBi³⁺-site – is an excellent host for optically active dopants.^38,39

Fig. 1 (a) Ex situ powder XRD patterns of Zn_1.998−xGeO₄:Yb_0.002³⁺Bi_x³⁺ (x = 0, 0.006, 0.02, 0.06 and 0.18) in dependence of Bi³⁺ doping concentration. Tabulated standard diffraction patterns of Zn₂GeO₄ and Bi₄(GeO₄)₃ are shown for reference. The unit cell crystal structure of (b) Zn₂GeO₄ and (c) Bi₄(GeO₄)₃ with [ZnO₄]/[BiO₆] and [GeO₄] tetrahedra is also shown.

Fig. 2 exemplarily shows DR spectra of blank Zn₂GeO₄ and Bi³⁺ (x = 0.02) singly-doped Zn₂GeO₄. The blank Zn₂GeO₄ sample exhibits a broad absorption band in the UV region, i.e., from ∼200 to 360 nm. This absorption band can be deconvoluted in at least two contributions: an intense band at ∼200–290 nm and a weaker shoulder at ∼290–360 nm. The intense absorption band is ascribed to electronic transitions from the valence band to the conduction band of Zn₂GeO₄ host, whereas the origin of the weaker shoulder is not clear.²⁰ From the absorption edge in UV region at ∼263 nm, we estimate a band-gap energy of ∼4.71 eV which is good accordance with reported data.^25,40 Addition of Bi³⁺ adds a strong yellow absorption band (originating from the allowed transition of ¹S₀ → ³P₁ in trivalent bismuth), shifting the UV-edge to ∼450 nm.

Fig. 2 UV-vis diffuse relectance spectra of blank Zn₂GeO₄ (black line) and Bi³⁺ (x = 0.02) singly doped Zn₂GeO₄ (red line). The band gap value of Zn₂GeO₄ is derived by the intersection of tangent (blue line) with abscissa. The blue tangent is shown as a guide for the eye to highlight the band-gap of Zn₂GeO₄.

As already mentioned, Zn₂GeO₄ is a self-activated phosphor. In Fig. 3a and b, we show the room-temperature PLE and PL spectra of blank Zn₂GeO₄. PLE occurs in a strong band from 250 to 280 nm and another, weaker band from 280 to 375 nm, which is consistent with the DR spectra shown in Fig. 2. The excitation maximum locates at ∼270 nm which correspond-well with the band gap energy of pure Zn₂GeO₄. Under UV excitation at 270 nm, PL occurs in the spectral region of ∼475–625 nm with a maximum at ∼540 nm with a full width at half maximum (FWHM) of ∼1647 cm⁻¹ (∼48 nm) (Fig. 3b). This PL band is assigned to radiative defect-recombination, i.e., a donor (V_O˙ and Zn_i˙) and an acceptor defect (V_Zn′ and ionized V_Ge).²⁷ As for effective PL lifetime τ_1/e, we find a range < 3 μs (that is, below the pulse length of the employed Xe flash lamp).

Fig. 3 (a) Room-temperature PLE (λ_em = 540 nm) and (b) PL (λ_ex = 270 nm) spectra for greenish PL in pure Zn₂GeO₄. (c) PLE (λ_em = 976 nm) and (d) PL (λ_ex = 270 nm) spectra of Yb³⁺:²F_2/5 → Yb³⁺:²F_2/7 emission in Yb³⁺ singly doped Zn₂GeO₄. The inset of (d) presents the Stark splitting energy levels of Yb³⁺ from the best fit of the PL spectrum of Yb³⁺. The dashed and solid lines in (d) represent the individual Gaussian functions and total fit of the spectra, respectively, as used for deconvolution.

Room-temperature PLE and PL spectra of Yb³⁺ in Yb³⁺ singly-doped Zn₂GeO₄ are provided in Fig. 3c and d. The PLE spectrum is very similar to the spectrum of pure Zn₂GeO₄ (Fig. 3a) and also corresponds to the DR spectrum (Fig. 2), showing the aforementioned two contributions at 250–290 nm and 290–375 nm. This directly evidences PL activation of Yb³⁺ through energy transfer from the Zn₂GeO₄ host. The corresponding NIR PL covers a broad spectral region of 930–1100 nm with a sharp peak at 976 nm and two red-shifted shoulders. The PL spectrum can best be deconvoluted into three Gaussian functions with maxima at 976, 1001 and 1055 nm (as shown in Fig. 3d). Since neither Bi³⁺ nor the Zn₂GeO₄ host are expected to contribute to the NIR emission at this excitation band, the observed NIR PL is assigned to the transition from the lowest stark level of Yb³⁺:²F_2/5 to the three different stark levels of the ground state of Yb³⁺:²F_2/7 (inset of Fig. 3d). The occurrence of Yb³⁺-related PL after excitation into the conduction band of Zn₂GeO₄ confirms existence of energy transfer from Zn₂GeO₄ to Yb³⁺.

Fig. 4a–c present room-temperature PLE and PL spectra of Bi³⁺ in Zn_1.998−xGeO₄:Yb_0.002³⁺Bi_x³⁺ for varying values of x. In the presence of Bi³⁺, an additional PL band appears in the spectral range of 390 to 500 nm (maximum at 420 nm, FWHM ∼4325 cm⁻¹, excitation at 350 nm). This is ascribed to the transition of Bi³⁺: ³P₁ → ¹S₀ (Fig. 4c). The Stokes-shift of this Bi³⁺-related PL is relatively low, i.e., ∼5200 cm⁻¹.²⁹ With increasing Bi³⁺ concentration, the intensity of the Bi³⁺-related PL is found to decrease as a result of concentration quenching effect (inset of Fig. 4c).^41,42 The corresponding PLE spectra exhibit two distinct bands, i.e., from 250 to 290 nm and from 310 to 390 nm, with peaks at ∼250 and 345 nm, respectively (Fig. 4a). The high-energy part is ascribed to the extrinsic PLE band of the aforementioned transition from the valence to the conduction band of Zn₂GeO₄, which suggests the energy transfer from the Zn₂GeO₄ host to Bi³⁺. The latter contribution is assigned to the intrinsic transition of Bi³⁺: ¹S₀ → ³P₁. Under extrinsic excitation through the Zn₂GeO₄ host (270 nm), the overlapping PL bands the Bi³⁺ active center and the Zn₂GeO₄ host itself can be observed simultaneously with peaks at ∼450 and ∼525 nm, respectively (Fig. 4b and d). The overall PL spectrum therefore spans the very broad spectral region of 390–650 nm with a FWHM of ∼4020 cm⁻¹ (∼160 nm). The maximum PL intensity of Bi³⁺: ³P₁ → ¹S₀ is found for x = 0.06, whereas PL from the Zn₂GeO₄ host is quenched with increasing concentration of Bi³⁺, again confirming the occurrence of energy transfer from Zn₂GeO₄ to Bi³⁺. It is further observed that under extrinsic excitation through Zn₂GeO₄, the PL peak which is attributed to Bi³⁺: ³P₁ → ¹S₀ shifts to longer wavelength, i.e., from ∼420 to 450 nm, as compared to intrinsic excitation of Bi³⁺ centers. This evidences the occurrence of energy transfer from Zn₂GeO₄ to Bi³⁺ through the defect level of Zn₂GeO₄ (V_O˙ and Zn_i˙). As with the intrinsic luminescence from the Zn₂GeO₄ host, the effective lifetime of the Bi³⁺: ³P₁ → ¹S₀ emission is <3 μs.

Fig. 4 Room-temperature PLE (λ_em = 420 nm), PL ((b) λ_ex = 270 nm and (c) λ_ex = 350 nm) and (d) normalized PL (λ_em = 270 nm; normalized to the PL peak at 450 nm) spectra of Zn_1.998−xGeO₄:Yb_0.002³⁺Bi_x³⁺ (x = 0.006, 0.02, 0.06 and 0.18) as a function of Bi³⁺ doping concentration. The insets of (b) and (d) show PL peak intensity in Zn_1.998−xGeO₄:Yb_0.002³⁺Bi_x³⁺ phosphors versus Bi³⁺ doping concentration. The drawn lines in the inset of (b) and (c) are to guide the eye.

Room-temperature PLE (monitoring Yb³⁺: ²F_5/2 → ²F_7/2 PL at 976 nm) and PL (through extrinsic excitation through the Zn₂GeO₄ host at 280 nm and through the Bi³⁺ band at 365 nm, respectively) spectra of Zn_1.998−xGeO₄:Yb_0.002³⁺Bi_x³⁺ are shown in Fig. 5a–c. The PLE bands of the Zn₂GeO₄ host (250 to 315 nm) as well as of the Bi³⁺ active center (345 to 450 nm) can be identified in the PLE spectra of Yb³⁺ (Fig. 5a). This clearly confirms that energy transfer occurs from those entities to Yb³⁺. The typical Yb³⁺-related NIR emission can consequently be observed either through excitation of the Zn₂GeO₄ host (270 nm) or through excitation of Bi³⁺: ¹S₀ → ³P₁ (365 nm, Fig. 5b and c). When excited at 270 nm, the introduction of a small amount of Bi³⁺ (x = 0.06) strongly enhances the PL intensity of Yb³⁺ (here: more than 12 times, Fig. 5b), whereas higher Bi³⁺ concentration results is quenching of the Yb³⁺-related PL. Similarly, when excited at 365 nm (i.e., on the Bi³⁺ band), the PL intensity of Yb³⁺ attains a maximum for x = 0.06 (Fig. 5c).

Fig. 5 Room-temperature PLE (λ_em = 976 nm) and PL (upon excitation of (b) Zn₂GeO₄ host at λ_ex = 280 nm and (c) of Bi³⁺ at λ_ex = 365 nm) spectra for NIR PL from Yb³ emission in Zn_1.998−xGeO₄:Yb_0.002³⁺Bi_x³⁺ (x = 0, 0.006, 0.02, 0.06 and 0.18) phosphors as a function of Bi³⁺ concentration. Insets (b) and (c) show PL peak intensity of Yb³⁺ in Zn_1.998−xGeO₄:Yb_0.002³⁺Bi_x³⁺ phosphors in dependence of Bi³⁺ concentration. Normalized room-temperature decay curves for NIR PL from Yb³ at 976 nm under pulsed excitation of at (d) 280 nm (Zn₂GeO₄ host) and (e) 365 nm (Bi³⁺). Inset of (d) and (e) shows effective lifetime τ_1/e for NIR PL from Yb³ emission as a function of Bi³⁺ concentration. The lines in the insets of (b), (c), (d) and (e) are guides for the eye.

Fig. 5d and e represent normalized decay curves of Yb³⁺: ²F_5/2 → ²F_7/2 PL at λ_em = 976 nm for excitation through the Zn₂GeO₄ host and through Bi³⁺, respectively. All decay curves deviate from a single-exponential form, which is related to the different mechanisms of energy transfer, and eventually also to the presence of various types of Bi³⁺-related emission species (where the above-noted optimal dopant concentration coincides with the onset of Bi₄(GeO₄)₃ formation (Fig. 1)). The effective lifetime τ_1/e of Yb³⁺: ²F_5/2 → ²F_7/2 PL increases with increasing Bi³⁺ concentration from ∼90 to 620 μs for excitation through the host (Fig. 5d). When excited through Bi³⁺ species, it first increases from 290 to 510 μs for a Bi³⁺ concentration up to (x = 0.02). For higher Bi³⁺ dopant concentration, a saturation is found (Fig. 5e and inset of Fig. 5e).

The overall scheme of spectral conversion in Zn₂GeO₄:Bi³⁺,Yb³⁺ is illustrated in Fig. 6 and compared to the solar irradiance spectrum. Without Bi³⁺, only photons in the spectral region below 350 nm (which accounts for only a small portion of the total solar irradiance) can be harvested. In this case, conversion occurs through intrinsic emission as well as through Yb³⁺ emission centers in the spectral regions of 475–625 nm (green) and 930–1100 nm (NIR). The introduction of Bi³⁺ strongly enhances the spectral harvesting efficiency through extending the absorption band to up to 450 nm. DS PL from both Bi³⁺ and Zn₂GeO₄ then spans a broad region in the blue-to-red spectral range. Through energy transfer from Zn₂GeO₄ to Yb³⁺via Bi³⁺ as well as directly from Bi³⁺ to Yb³⁺, notable conversion to the NIR region can be obtained. In this way, Yb³⁺-emission is activated over the excitation region of 250–500 nm.

Fig. 6 PLE spectrum of Yb³⁺ in Yb³⁺ singly (cyan line) and Yb³⁺/Bi³⁺ co-doped (violet line) Zn₂GeO₄, and PL spectrum of Zn_1.998−xGeO₄:Yb_0.002³⁺Bi_0.06³⁺ phosphors in visible (green line) and NIR (red line) region under excitation at 280 nm. The standard solar spectrum for air mass 1.5 according to ASTM G173 - 03(2012) is shown in the background for reference.

This process is summarized in Fig. 7 which depicts the energy level diagrams of the Zn₂GeO₄ host together with the Bi³⁺ and Yb³⁺ emission centers. For Yb³⁺ singly-doped Zn₂GeO₄, the electrons are excited from valence to conduction band, leaving hole centers at the valence band under excitation at 270 nm. These defects are subsequently trapped through non-radiative processes.^20,26,27 Their recombination is accompanied by the intrinsic PL emission of the Zn₂GeO₄ host at ∼ twice the energy-gap of Yb³⁺: ²F_7/2 → ²F_5/2. As a result, besides PL emission, the trapped energy can also be transferred to two neighboring Yb³⁺ entities, resulting in NIR emission at ∼1000 nm. Due to the absence of an intermediate energy level at ∼1000 nm, the energy transfer from Zn₂GeO₄ host to Yb³⁺ is understood as a second order cooperative DC process. With the introduction of Bi³⁺, the absorption band extends to the blue region. Electrons which are excited into the conduction band of Zn₂GeO₄ can relax to the defect level of Zn₂GeO₄ (V_O˙ and Zn_i˙) and then to the excited level of Bi³⁺: ³P₁. Alternatively, the excited electrons at conduction band of Zn₂GeO₄ can directly relax to the excited level of Bi³⁺: ³P₁ through a multi-phonon-assisted relaxation process. At a small Stokes shift of the Bi³⁺: ³P₁ level, DS PL from Bi³⁺ can be observed at an emission energy which is a little bit more than twice the absorption energy of Yb³⁺: ²F_7/2 → ²F_5/2. Similar to host excitation, also here, energy can be transferred to two neighboring Yb³⁺ in a cooperative DC process to yield Yb³⁺ emissions at ∼1000 nm. Under excitation at 365 nm, the electrons are excited from the ground state of Bi³⁺: ¹S₀ to the excited state of Bi³⁺: ³P₁. The following processes are the same as just described.

Fig. 7 Energy level diagrams of Zn₂GeO₄ host, Bi³⁺ and Yb³⁺, and schematic illustration of the energy transfer mechanisms from the Zn₂GeO₄ host to Yb³⁺, from the Zn₂GeO₄ host to Yb³⁺via Bi³⁺ and from Bi³⁺ to Yb³⁺ (CET: cooperative energy transfer; ET: energy transfer; MET: multi-phonon-assisted energy transfer).

Conclusions

In conclusion, we discussed the spectral properties of the high-bandgap semiconductor Zn₂GeO₄ co-doped with Yb³⁺ and Bi³⁺ as a model system for ultra-efficient conversion of the UV-A/blue part of the solar spectrum to the NIR spectral region. We have shown that Yb³⁺-related PL can be activated through intrinsic absorption of the Zn₂GeO₄ lattice itself, and through Bi³⁺ centers. Energy transfer then occurs via cooperative DC from trapped defect states and from the ³P₁ level of Bi³⁺. This results in a strong increase in the absolute intensity of Yb³⁺-related PL. The material enables ultra-efficient harvesting of UV-A to visible radiation for energy conversion processes which rely on NIR irradiation such at c-Si photovoltaics and various photochemical processes.

Acknowledgements

The authors gratefully acknowledge financial support from the German Science Foundation (DFG) through grant no. WO 1220/2-2 and the Department of Education of Guangdong Province (grant no. 2013gjhz0001).

Notes and references

1. X. Huang, S. Han, W. Huang and X. Liu, Chem. Soc. Rev., 2012, 42, 173–201.
2. Q. Y. Zhang and X. Y. Huang, Prog. Mater. Sci., 2010, 55, 353–427.
3. J. de Wild, A. Meijerink, J. K. Rath, W. G. J. H. M. van Sark and R. E. I. Schropp, Energy Environ. Sci., 2011, 4, 4835–4848.
4. B. M. van der Ende, L. Aarts and A. Meijerink, Phys. Chem. Chem. Phys., 2009, 11, 11081–11095.
5. L. Wondraczek, M. Batentschuk, M. A. Schmidt, R. Borchardt, S. Scheiner, B. Seemann, P. Schweizer and C. J. Brabec, Nat. Commun., 2013, 4, 2047.
6. M. Peng and L. Wondraczek, J. Mater. Chem., 2009, 19, 627–630.
7. Q. Xia, M. Batentschuk, A. Osvet, P. Richter, D.-P. Häder, J. Schneider, C. J. Brabec, L. Wondraczek and A. Winnacker, Opt. Express, 2013, 21, A909–A916.
8. D. Chen, Y. Wang and M. Hong, Nano Energy, 2012, 1, 73–90.
9. T. Trupke, M. A. Green and P. Würfel, J. Appl. Phys., 2002, 92, 1668–1674.
10. G. Gao and L. Wondraczek, Opt. Mater. Express, 2013, 3, 633–644.
11. D. Yu, S. Ye, M. Peng, Q. Zhang and L. Wondraczek, Appl. Phys. Lett., 2012, 100, 191911.
12. D. Yu, X. Huang, S. Ye, M. Peng, Q. Y. Zhang and L. Wondraczek, Appl. Phys. Lett., 2011, 99, 161904.
13. J. Zhou, Y. Teng, S. Zhou and J. Qiu, Int. J. Appl. Glass Sci., 2012, 3, 299–308.
14. R. Zhou, Y. Kou, X. Wei, C. Duan, Y. Chen and M. Yin, Appl. Phys. B, 2012, 107, 483–487.
15. L. Aarts, B. van der Ende, M. F. Reid and A. Meijerink, Spectrosc. Lett., 2010, 43, 373–381.
16. J. J. Eilers, D. Biner, J. T. van Wijngaarden, K. Krämer, H.-U. Güdel and A. Meijerink, Appl. Phys. Lett., 2010, 96, 151106.
17. Q. Y. Zhang, G. F. Yang and Z. H. Jiang, Appl. Phys. Lett., 2007, 91, 051903.
18. B. M. van der Ende, L. Aarts and A. Meijerink, Adv. Mater., 2009, 21, 3073–3077.
19. H. Lin, D. Chen, Y. Yu, A. Yang and Y. Wang, Opt. Lett., 2011, 36, 876–878.
20. G. Gao and L. Wondraczek, J. Mater. Chem. C, 2013, 1, 1952–1958.
21. Y. Li, J. Wang, W. Zhou, G. Zhang, Y. Chen and Q. Su, Appl. Phys. Express, 2013, 6, 082301.
22. X. Y. Huang and Q. Y. Zhang, J. Appl. Phys., 2010, 107, 063505.
23. X. Wei, S. Huang, Y. Chen, C. Guo, M. Yin and W. Xu, J. Appl. Phys., 2010, 107, 103107.
24. M. Balestrieri, G. Ferblantier, S. Colis, G. Schmerber, C. Ulhaq-Bouillet, D. Muller, A. Slaoui and A. Dinia, Sol. Energy Mater. Sol. Cells, 2013, 117, 363–371.
25. C. Yan, N. Singh and P. S. Lee, Appl. Phys. Lett., 2010, 96, 053108.
26. G. Anoop, K. M. Krishna and M. K. Jayaraj, J. Electrochem. Soc., 2008, 155, J7–J10.
27. Z. Liu, X. Jing and L. Wang, J. Electrochem. Soc., 2007, 154, H500–H506.
28. G. Blasse and B. C. Grabmaier, Luminescent materials, Springer-Verlag, 1994.
29. G. Blasse and A. Bril, J. Chem. Phys., 1968, 48, 217–222.
30. M. Peng, G. Dong, L. Wondraczek, L. Zhang, N. Zhang and J. Qiu, J. Non-Cryst. Solids, 2011, 357, 2241–2245.
31. M. Ilmer, B. C. Grabmaier and G. Blasse, Chem. Mater., 1994, 6, 204–206.
32. F. Kang, X. Yang, M. Peng, L. Wondraczek, Z. Ma, Q. Zhang and J. Qiu, J. Phys. Chem. C, 2014, 118, 7515–7522.
33. W. Xu, M. Peng, Z. Ma, G. Dong and J. Qiu, Opt. Express, 2012, 20, 15692–15702.
34. N. Niu, F. He, S. Gai, C. Li, X. Zhang, S. Huang and P. Yang, J. Mater. Chem., 2012, 22, 21613–21623.
35. U. Rambabu and S.-D. Han, Ceram. Int., 2013, 39, 701–708.
36. M. Peng, B. Sprenger, M. A. Schmidt, H. Schwefel and L. Wondraczek, Opt. Express, 2010, 18, 12852–12863.
37. R. D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr., 1976, 32, 751–767.
38. R. Nitsche, J. Appl. Phys., 1965, 36, 2358–2360.
39. Y. T. Arslanlar, Z. Kotan, R. Kibar, A. Canımoğlu and N. Can, Spectrosc. Lett., 2013, 46, 590–596.
40. Q. Liu, Y. Zhou, J. Kou, X. Chen, Z. Tian, J. Gao, S. Yan and Z. Zou, J. Am. Chem. Soc., 2010, 132, 14385–14387.
41. G. Gao, R. Meszaros, M. Peng and L. Wondraczek, Opt. Express, 2011, 19, A312–A318.
42. G. Gao, S. Reibstein, E. Spiecker, M. Peng and L. Wondraczek, J. Mater. Chem., 2012, 22, 2582–2588.

---