1. Introduction

One of the reasons for the low theoretical quantum efficiency of silicon solar cells is related to the Shockley–Quaisser limit, where due to the spectral mismatch between incident photons and the band-gap of crystalline silicon of ~1.1 eV, only about 32% of the solar spectrum can be used for the generation of photoelectrons. In addition, high energy photons are often lost due to recombination effects and/or active filtering to prevent photobleaching [1–3]. For these reasons, several ways have been explored to convert, by photoluminescence, parts of the solar spectrum to regions where the photovoltaic efficiency might be improved. These comprise three principle routes, i.e., downconversion [4–13] where a near-ultraviolet (NUV) to visible (Vis) photon is cut into two or more near infrared (NIR) photons, up-conversion [14,15], where the energy of two or three NIR photons is absorbed to emit one photon of higher energy, and down-shifting [16,17], where the energy of incident photons is shifted into wavelength regions which can be harvested more effectively. Downconversion provides the additional advantage that the energy loss due to the thermalization of hot charge carriers after absorption of a high-energy photon is minimized. A brief look at the solar irradiance spectrum reveals that, in principal, ~15% of excess energy, the UV-Vis part of the spectrum which would otherwise be lost as heat, is available for downconversion [18,19].

Considering the energy level of all lanthanides, Yb³⁺ has been recognized as the most suitable candidate for downconversion. That is, the Yb³⁺ ion has only a single excited state, ²F_5/2 at ~10000 cm⁻¹, above the ground state of ²F_7/2. This means that Yb³⁺ centers may act as efficient recipients for energy quanta of ~10000 cm⁻¹ from any other co-dopant to emit photons with a wavelength of about ~1 µm [4,19]. To effectively sensitize Yb³⁺ to the UV-Vis spectral region, a donor ion with an energy level at ~20000 cm⁻¹ is necessary. For this, Pr³⁺, Er³⁺, Nd³⁺, Ho³⁺, Tm³⁺, Tb³⁺ or Ce³⁺ may be employed. In the present report, we focus on Pr³⁺ to sensitize Yb³⁺ because the absorption bands of Pr³⁺ cover a broad spectral window in the blue region due to the successive energy levels of Pr³⁺:³P_{J(J = 0, 1 and 2)} [Fig. 1(b)]. These absorption bands are located at approximately twice the band level of Yb³⁺:²F_7/2 →²F_5/2 [20–22]. As host material, we chose a glass matrix which exhibits virtually universal forming capability and high compositional flexibility for well-controlled and homogeneous doping. To adjust the ligand symmetry and phonon energy, the glass composition may be selected so that in a subsequent annealing procedure, a crystalline species precipitates from the undercooled melt into which dopants are incorporated during crystallization [23–28]. Here we employ a glass of the composition 20 SrO-20 La₂O₃-10 Al₂O₃-40 B₂O₃-10 SiO₂ [29]. In this system, LaBO₃ crystallites can be precipitated by controlled nucleation in a heat treatment process. Due to the equivalent charge and similar ionic radii of La³⁺ (1.16 Å, CN = 8), Pr³⁺ (1.13 Å, CN = 8) and Yb³⁺ (0.99 Å, CN = 8), we expect that the dopant species can readily be incorporated into the lattice of crystalline LaBO₃ [30].

 

Fig. 1 Absorption spectra of the Pr³⁺/ Yb³⁺ co-doped SLABS glasses dependent on Yb₂O₃ doping concentration.

Download Full Size | PDF

2. Experimental

Precursor glass samples with nominal composition (mol %) 20 SrO-(19-x) La₂O₃-10 Al₂O₃-40 B₂O₃-10 SiO₂-1 Pr₂O₃-x Yb₂O₃ (SLABS, x = 0, 0.1, 0.2, 0.5, 1, 2, and 5) were produced by conventional melting in alumina crucibles at 1400°C for 2 h (air). Glass slabs were obtained after pouring the melt into a preheated (500°C) graphite mold and subsequent annealing at 550°C for 2 h. From these slabs, disks of 20 × 20 × 3.0 mm³ were cut and polished on both sides. To obtain glass ceramic samples, individual specimen were placed on alumina substrates and isothermally annealed at 800°C for up to 32 h (air).

UV-VIS-NIR absorption spectra were recorded from 300 to 2500 nm with a double-beam photo-spectrometer equipped with a 150 mm integration sphere and a PbS detector (Perkin-Elmer Lambda 950). Static and dynamic photoluminescence (PL) were studied with a high-resolution spectrofluorometer and time correlated single photon counting (TCSPC, Horiba Jobin Yvon Fluorolog FL3-22) using a static Xe lamp (450 W) and a Xe flashlamp (75 W) as excitation sources. NIR PL was observed with a thermoelectrically cooled InP/InGaAs-based photomultiplier tube (Hamamatsu H10330A-75). Photoluminescence excitation (PLE) spectra were corrected over the lamp intensity with a silicon photodiode and PL spectra were corrected by the spectral response of the detector using correction spectra of the employed PMT. The crystallization process of each specimen was first analyzed by differential scanning calorimetry (DSC, Netzsch, Ar atmosphere) with a heating rate of 20 K min⁻¹. To identify the crystalline phases after heat treatment, X-ray diffractometry (XRD, Siemens Kristalloflex D500, Bragg-Brentano, 30 kV/30 mA, Cu Kα) was performed with a step width of 0.02°/s and a counting time of 10 s per step over the 2θ range of 5-70°. All analyses were performed at room temperature.

3. Results and discussion

In Fig. 1 the effect of Yb₂O₃ concentration on optical absorption is shown. For Pr³⁺ singly doped SLABS glass, eight absorption peaks centered at 440, 471, 484, 591, 1003, 1416, 1515 and 1915 nm are ascribed to the inhomogeneously broadened 4f–4f transitions from the ground state ³H₄ to the excited states ³P₂, ³P₁, ³P₀, ¹D₂, ¹G₄, ³F₄, ³F₃ and ³F₂ of Pr³⁺ (labels in Fig. 1) [31]. As mentioned in the introduction section, three relatively strong and overlapping absorption bands (³H₄ → ³P_1,2,3) cover a large part of the blue wavelength range of 415-505 nm. For the Pr³⁺/Yb³⁺ co-doped samples, the additional absorption band with a maximum at 976 nm and a shoulder peak at ~935 nm is assigned to the transition from the ground state of Yb³⁺, ²F_7/2, to two different Stark levels of the excited state of ²F_5/2. As expected, the intensity of this NIR absorption band of Yb³⁺ increases linearly with increasing Yb³⁺ doping concentration.

Radiative transitions within the 4fⁿ configurations of trivalent rare earth ions can be predicted by the Judd-Ofelt (J-O) theory. The Judd-Ofelt parameters Ω₂, Ω₄ and Ω₆ as obtained from the measured absorption spectra in Fig. 1 are calculated 1.10 × 10⁻²⁰, 1.06 × 10⁻²⁰ and 0.87 × 10⁻²⁰ cm² respectively [31–33]. Using these values, the radiative decay rate A_JJ', the branching ratio β_JJ' and the radiative emission lifetime τ_rad were estimated. Data are summarized in Table 1.PLE and PL spectra reveal direct evidence of energy transfer from Pr³⁺ to Yb³⁺ [Figs. 2(a)–2(d)]. Room temperature PLE spectra are shown for emission at 608 nm (Pr³⁺) and 976 nm (Yb³⁺), respectively. Consistent with the absorption spectra (Fig. 1), the PLE bands of Pr³⁺ comprise three characteristic peaks at 445, 471 and 483 nm, together with a relatively weak band at 591 nm. The former three bands exhibit similar intensity and strong overlap with each other, what lets expect a relatively high excitation efficiency in this spectral range. The PLE spectra of the Yb³⁺:²F_5/2→²F_7/2 [Fig. 2(d)] are in full agreement with those of the Pr³⁺-related PLE [Fig. 2(a)] and with the absorption spectra of Pr³⁺ (Fig. 1), what is taken as clear evidence for the sensitization of the Yb³⁺ emission center by Pr³⁺ donors in the NUV-Vis spectral range. In the following, the maximum PLE peak at 445 nm is selected to monitor the PL spectra as a function of Yb₂O₃ doping concentration.

Table 1. Predicted Radiative Decay Rates (A_JJ'), Branching Ratio (β _JJ') and Radiative Emission Lifetime (τ_rad) of Pr³⁺ in SLABS Glasses at 300 K

View Table

 

Fig. 2 Steady-state visible (a) PLE (λ_em = 608 nm) and (b) PL (λ_ex = 445 nm) spectra of Pr³⁺, and NIR (d) PLE (λ_ex = 976 nm) and (e) PL (λ_ex = 445 nm) spectra of Yb³⁺ of the Pr³⁺/Yb³⁺ co-doped SLABS glasses as a function of Yb³⁺ doping concentration. (c) PL peak intensity of Pr³⁺ at 608 nm and (f) integrated PL intensity in the spectral range of 920-1200 nm as a function of Yb³⁺ doping concentration. Inset of (b): Energy levels of Pr³⁺. Lines in (c) and (f) are guides for the eye.

Download Full Size | PDF

The PL spectra of Pr³⁺-related emission centers [Fig. 2(b)] consist of six characteristic bands which are located at 487, 530, 608, 646, 701 and 730 nm. Their designation is shown in the inset of Fig. 2(a) and labeled in Fig. 2(b) [34–36]. The PL spectra of all samples are dominated by the red band at ~608 nm which is attributed to the transition of Pr³⁺:³P₀→³H₆. The intensity of all bands of Pr³⁺ decreases gradually with increasing Yb³⁺ concentration over the whole range of observed concentrations [Figs. 2(b) and 2(c)]. Very sharp drop appears to occur in the sample with 5 mol% of Yb₂O₃.

In Fig. 2(e), PL spectra of the NIR region are presented. For the Pr³⁺ singly-doped SLABS sample, the broad NIR PL band centered at ~1047 nm with a full width at half maximum (FWHM) of 73 nm and a shoulder centered at ~998 nm is ascribed to the transitions of ¹D₂→³F₄ and ¹D₂→³F₃ in Pr³⁺, respectively. The relatively weak PL band at ~1475 nm is assigned to ¹D₂→¹G₄. The observation of three bands originating from ¹D₂ implies that relaxation of the ³P_1,2,3 levels occurs via a multi-phonon relaxation to ¹D₂. This indicates that ¹D₂ is efficiently occupied [2]. After co-doping with a small amount of Yb³⁺ (x = 0.1), a sharp PL peak appears at 976 nm, accompanied by a broad shoulder at ~997 nm. These two bands are attributed to the transition of the lowest Stark level of ²F_5/2 in Yb³⁺ to two different Stark levels of the ground state of ²F_7/2 [20]. The latter two overlap with the Pr³⁺-related emission band at 1046 nm. To account for these three bands, all NIR PL spectra were deconvoluted into three Gaussian functions [Figs. 3(a)–3(f)]. Regarding the concentration of Yb³⁺, the intensity of the emission bands at 976 and 997 nm increases for x ≤ 0.5 and decreases sharply for higher amounts Yb³⁺-doping. We conclude that this is caused by concentration quenching and that, in the present matrix material, x = 0.5 represents the optimal doping concentration. As already seen in the excitation spectra, also the decrease of the intensity of Pr³⁺-related PL intensity after co-doping with a small amount of Yb³⁺ and the parallel occurrence of increasingly intense Yb³⁺-related emission bands indicate energy transfer from Pr³⁺ to Yb³⁺. Pr³⁺ acts as a sensitizer by absorbing NUV-Vis photons and partially transferring their energy to Yb³⁺ [Figs. 2(e) and 2(f)] [37–39]. Parallel to the intensity decrease of PL bands of Pr³⁺ in the visible range, also the PL intensity of Pr³⁺:¹D₂→³F₄ [1047 nm, Figs. 3(e) and 3(f)] and Pr³⁺:¹D₂→¹G₄ [1475 nm, Fig. 2(e)] decrease gradually with increasing Yb³⁺ concentration. At the same time, the position of the Pr³⁺:¹D₂→³F₃ band red-shifts from 1047 to 1072 nm with increasing Yb³⁺ concentration [Fig. 3(e)], what is understood as another consequence of energy transfer from Pr³⁺ to Yb³⁺.

 

Fig. 3 Gaussian fitted PL spectra of the Pr³⁺/ Yb³⁺ co-doped SLABS glasses in the NIR region as a function of Yb³⁺ doping concentration, (a) Yb³⁺:²F_7/2→²F_5/2 at 976 nm, (c) Yb³⁺:²F_7/2→²F_5/2 at 997 nm and (e) Pr³⁺:¹D₂→³F₃. (b), (d) and (f) corresponding PL intensity change of the Pr³⁺/ Yb³⁺ co-doped SLABS glasses dependent on Yb₂O₃ doping concentration respectively. Lines in (b), (d) and (f) are guides for the eye.

Download Full Size | PDF

Figures 4(a), 4(c), and 4(e) depict normalized PL decay curves for the Pr³⁺-related transitions of ³P₀→³H₆ (608 nm) and ¹D₂→¹G₄ (1475 nm), and for the Yb³⁺-related transition of ²F_5/2→²F_7/2 (976 nm) as a function of Yb³⁺ concentration after excitation at 445 nm. For Pr³⁺:³P₀→³H₆, all decay curves exhibit multi-exponential behavior [Fig. 4(a)]. In the Pr³⁺ singly-doped sample, this may be due to cross-relaxation between Pr³⁺ ions, whereas in co-doped samples, this is caused by the increasingly faster depopulation of the Pr³⁺:³P₀ level due to energy transfer from ³P_J of Pr³⁺ to ²F_5/2 of Yb³⁺ with increasing Yb³⁺ concentration [39]. The effective PL lifetime of the Pr³⁺:³P₀ state decreases from 7.2 to 1.2 μs when the Yb³⁺ concentration is increased to 5 mol % [Figs. 4(a) and 4(b)]. Similarly, also the effective PL lifetime of the Pr³⁺:¹D₂ state decreases from 7.6 to 2.1 μs with increasing Yb³⁺ concentration [Figs. 4(c) and 4(d)]. This shows that both states contribute to the energy transfer process. The energy transfer from the ³P₀ level results in downconversion of one visible photon into two NIR photons whereas the energy transfer from the ¹D₂ level results in down-shifting of one NIR photon into another NIR photon. A more detailed discussion will be presented in the following paragraphs. The energy transfer efficiency η_ETE is defined as the ratio between the number of depopulation events which result in energy transfer from donor to acceptor and the total population of the excited states in the donor. It can be calculated from the ratio of the lifetimes of the excited state(s) of the donor in the presence of the acceptor as compared to a specimen which contains only the donor species [6,38],

 

Fig. 4 (a) Normalized decay curves of the Pr³⁺/Yb³⁺ co-doped SLABS glasses dependent on Yb₂O₃ doping concentration under excitation at 445 nm by monitoring PL (a) at 608 nm (Pr³⁺:³P₀→³H₆), (c) at 1475 nm (Pr³⁺:¹D₂→¹G₄) and (e) at 976 nm (Yb³⁺:²F_5/2→²F_7/2). Yb³⁺ doping concentration dependent lifetimes of (b) Pr³⁺ PL at 608 nm, (d) Pr³⁺ PL at 1475 nm and (f)Yb³⁺ PL at 976 nm. Lines in (d), (e) and (f) are guides for the eye.

Download Full Size | PDF

Here, τ_x%Yb is the lifetime a Pr³⁺-related state after co-doping with x mol% Yb₂O₃. For both levels, ³P₀ and ¹D₂, the calculated values of η_ET increase with increasing Yb³⁺ concentration, i.e., from 11.5% and 9.8% for x = 0.1 to 83.0% and 88.1% for x = 5.0, respectively. The total theoretical quantum efficiency is then calculated as a function of Yb³⁺ concentration [6,40],

In agreement with the PL intensity change of Yb³⁺:²F_5/2→²F_7/2 (976 nm) upon Yb³⁺ concentration increase [Figs. 4(e) and 4(f)], also the effective lifetime of this emission band first increases up to ~56.6 μs for x = 0.5 and subsequently decreases for higher x, i.e., to 10.7 μs for x = 5.0.

To evaluate the potential of further enhancement of the energy transfer efficiency, glass ceramic samples were considered where LaBO₃ crystallites were precipitated so that Pr³⁺ and Yb³⁺ are incorporated into the crystalline lattice on La³⁺ sites.

Figure 5(a) exemplarily shows a DSC curve of the SLABS glass sample with a heating rate of 20 K min⁻¹ after baseline correction. The onset of glass transition T_g was observed at 679.5 ± 0.5 °C. Three crystallization peaks were found at 852, 926 and 948 ± 0.5 °C. Targeting the non-isothermal crystallization event at 852 °C, controlled crystallization was then performed isothermally at 800 °C for 32 h. XRD [Fig. 5(b)] of the untreated sample did not reveal any diffraction peaks, showing that the as-made glass was - within the accuracy of measurement - free of any crystalline phases. After thermal annealing, multiple intense diffraction peaks were found. These were indexed to the room-temperature orthorhombic phase of LaBO₃ (JCPDS card no. 00-013-0113). In addition, XRD patterns indicate the presence of at least one further, minor crystallite species which we assign as hexagonal SrAl₂B₂O₇ (JCPDS card no. 00-046-0621). For clarity, the tabulated diffraction patterns of both species are shown in Fig. 5(b). The lattice of LaBO₃ is illustrated in Fig. 5(c). It comprises an orthorhombic aragonite-type structure and cell parameters of a = 5.104 Å, b = 8.252 Å, c = 5.872 Å, composed of LaO₉ polyhedra and BO₃ trigonal groups [41]. Due to the aforementioned similarity of ionic radii, Pr³⁺ and Yb³⁺ should readily be incorporated on La³⁺-sites (for comparison, the radius of B³⁺ is about one third, i.e., 0.41 Å, as compared to the three rare earth species) [30].

 

Fig. 5 (a) DSC curve of the SLABS glass sample. (b) ex situ XRD patterns of the 1Pr³⁺/0.5Yb³⁺ co-doped SLABS glass and LaBO₃ glass ceramic annealed at 800 °C for 32 hrs. (c) Crystal structure of LaBO₃. Blue, black and red full sphere illustrates La³⁺, B³⁺ and O²⁺ respectively.

Download Full Size | PDF

Figures 6(a) and 6(d) show PLE spectra (monitoring 608 nm PL from Pr³⁺ and 976 nm PL from Yb³⁺, respectively) of the SLABS glass and glass ceramic for x = 0.5. After crystallization, the characteristic excitation bands of Pr³⁺ are clearly enhanced. Not surprisingly, annealing at elevated temperature leads to a great change in the PL spectra [Figs. 6(b) and 6(e)] and decay curves [Figs. 6(c) and 6(f)], as well. For Vis PL from Pr³⁺ [Fig. 6(b)], the peak intensity at 608 nm increased by a factor of ~5 after crystallization. Moreover, all PL bands apparently split and sharpen notably as a result of crystallization. The latter observations indicate that Pr³⁺ is indeed incorporated into the LaBO₃ crystal phase. A comparison of the decay curves of the Pr³⁺:³P₀→³H₆ PL band of the SLABS glass and the corresponding glass ceramic is shown in Fig. 6(c). Both curves can be best fit by second-order exponential equations, suggesting that in both cases a slow and a fast decay process contributes to PL. The lifetime values are obtained from the best fit. For both decay processes they increase after crystallization, i.e. from 2.4 to 7.1 μs and from 11.9 to 52.3 μs, respectively. This is taken as further evidence for the incorporation of Pr³⁺ into the crystallite phase. If Pr³⁺ is incorporated into the LaBO₃ lattice, we expect that Yb³⁺ would behave similarly. Otherwise, energy transfer between the two species would become highly unlikely due to their spatial and configurational separation. Examining the PL spectra of Yb³⁺ in the NIR range for the crystallized sample [Fig. 6(e)], a strongly enhanced intensity of the peak at 976 nm is visible as compared to the as-made glass. Additionally, the peak intensity ratio of the two peaks at 976 and 997 nm, I₉₇₆/I₉₉₇, decreases from 2.2 to 1.5 after crystallization. We attribute the latter observation to stronger Stark splitting of the ground state of Yb³⁺, ²F_7/2, in the LaBO₃ crystalline environment. Analogous to the behavior in the as-made glass, and also similar to the behavior of Pr³⁺ in the LaBO₃ lattice, decay curves follow a second-order exponential equation [Fig. 6(f)]. After crystallization, the two resulting lifetime values increase from 28.7 to 35.3 μs and from 149.0 to 297.3 μs, respectively, for the fast and for the slow decay process. From these observations, we conclude that also Yb³⁺ is incorporated into the crystallite phase.

 

Fig. 6 Steady-state visible (a) PLE and (b) PL spectra of Pr³⁺, and NIR (d) PLE and (e) PL spectra of Yb³⁺ of the 1Pr³⁺/0.5Yb³⁺ co-doped SLABS glass and LaBO₃ glass ceramic annealed at 800 °C for 32 hrs. (c) Visible (Pr³⁺ PL at 608 nm (³P₀→³H₆)) and (f) NIR (Yb³⁺ PL at 976 nm (²F_5/2→²F_7/2)) decay curves of the SLABS glass and glass ceramic annealed at 800 °C for 32 hrs excited at 445 nm monitoring PL at 608 nm of Pr³⁺ and 976 nm of Yb³⁺ respectively.

Download Full Size | PDF

Although on the first view, the presence of an intermediate level at about 10000 cm⁻¹ in Pr³⁺, suggest a relatively simple energy transfer process, at least four different ways may be considered for the pair of Pr³⁺ and Yb³⁺, illustrated in Figs. 7(a)–7(d). These are

 

Fig. 7 Energy levels diagram of Pr³⁺ and Yb³⁺, and possible energy transfer mechanism from Pr³⁺ to Yb³⁺. (a) Resonant energy transfer from Pr³⁺:³P₀ and Pr³⁺:¹G₄ levels to two Yb³⁺:²F_5/2 level. (b) One step first-order resonant energy transfer from Pr³⁺:³P₀ level to Yb³⁺:²F_5/2 and then a radiative relaxation from Pr³⁺:¹G₀ level to Pr³⁺:³H₄ or Pr³⁺:³H₅ level. (c) Cooperative energy transfer from one Pr³⁺:³P₀ level to two neighboring Yb³⁺:²F_5/2 level. (d) The first-order resonant energy transfer from Pr³⁺:¹P₂ level to Yb³⁺:²F_5/2.

Download Full Size | PDF

The first two mechanisms (a-b) appear rather unlikely due to the following reasons: firstly, the absence of Pr³⁺ PL bands from ³P₀→¹G₄ (which should occur ~950 nm), from ¹G₄→³H₆ (~1850 nm) and from ¹G₄ →³H₅ (~1330 nm), what indicates poor population of the ¹G₄ level of Pr³⁺. Secondly, the ¹G₄ level of Pr³⁺ is ~200 cm⁻¹ lower than the ²F_5/2 level of Yb³⁺. This means that energy transfer should rather occur from Yb³⁺ to Pr³⁺ than vice versa. Thirdly, the absorption band which is assigned to the transition from the ground state of Pr³⁺, ³H₄, to the ¹G₄ level is very weak (Fig. 1). Fourthly, the energy gap between the ¹G₄ and the lower lying ³F₄ level is low (~3000 cm⁻¹) may lead to very high multi-phonon assisted non-radiative transition rates from the ¹G₄ level. Finally, the branching ratio of the ³P₀→¹G₄ radiative transition and the total radiative rates of the ³P₀ level are as low as ~0.06 and 218.07 cm⁻¹, respectively, resulting in a low population of the ¹G₄ level (Table 1).

As briefly discussed in the previous chapters, the decrease in intensity of all PL bands of Pr³⁺ in the presence of Yb³⁺ [Figs. 2(a)–2(f)], and also of PL lifetime of the Pr³⁺:³P₀ and Pr³⁺:¹D₂ levels [Figs. 4(a)–4(d)] suggests that energy transfer occurs through both the ³P₀ and ¹D₂ levels of Pr³⁺. The branching ratio of the ³P₀→¹D₂ radiative transition is close to zero. This indicates the dominating multi-phonon assisted non-radiative transition of ³P₀→¹D₂ (Table 1). The excitation efficiency of Yb³⁺ via the ³P₀ level of Pr³⁺ is 2.7 times higher than via the ¹D₂ level [Fig. 2(e)]. This indicates that also energy transfer from the ³P₀ state is more efficient than from the ¹D₂ state. Additionally, the total radiative rate of the ³P₀ level (10319.43 cm⁻¹) is much higher than that of the ¹D₂ level (924.43 cm⁻¹) which confirms our conclusion (Table 1). The absence of ³P₀→¹G₄ Pr³⁺ PL also indicates that the cooperative energy transfer process cannot be neglected [Fig. 7(c)].

4. Conclusions

In summary, NIR downconversion of one blue photon into two NIR photons (~10000 cm⁻¹) was observed in Pr³⁺/Yb³⁺ co-doped SLABS glasses and corresponding LaBO₃ glass ceramics. The decrease of all Pr³⁺-related PL bands in intensity and lifetime with increasing Yb³⁺ concentration along with the observation of typical Pr³⁺ PLE bands by monitoring Yb³⁺-related PL at 976 nm indicated that energy transfer occurs through both the ³P₀ and the ¹D₂ level of Pr³⁺ to Yb³⁺. Thereby, transfer from Pr³⁺:³P₀ to Yb³⁺ occurs as a downconversion reaction where of one visible photon is cut into two NIR photons, whereas transfer from Pr³⁺:¹D₂ to Yb³⁺ results in a down-shift of the energy of a single NIR. The transfer efficiency from the Pr³⁺:³P₀ level is higher as compared to that from the Pr³⁺:¹D₂ level. A theoretical quantum efficiency of 183% was estimated for this process. In the present matrix material, the optimal doping concentration of Yb₂O₃ for downconversion is ~0.5 mol%. Crystallization of the as-made glass into a LaBO₃ glass ceramic leads to improved emission properties. Thereby, Pr³⁺ ions as well as Yb³⁺ ions enter the crystallite lattice, assumedly on La³⁺ sites.