Open Access
Add to BibSonomy
Abstract
We report the development of a high-brightness, high-power Ce:YAG luminescent concentrator pumped by 2240 blue LEDs in quasi-continuous wave operation (10 µs, 10 Hz). Using light confinement and recycling in the three space dimensions, the parallelepiped (1mm×14×mm×200mm) Ce:YAG emits a power of 145 W from a square output surface (1 × 1mm2) corresponding to a brightness of 4.6 kW/cm2/sr. This broadband yellow source has a unique combination of luminous flux (7.6 104 lm) and brightness (2.4 104 cd/mm2) and overcomes many other visible incoherent sources by one order of magnitude. This paper also proposes a deep understanding of the performance drop compared to a linear behavior when the pump power increases. Despite excited state absorption was unexpected for this low doped Ce:YAG pumped at a low irradiance level, we demonstrated that it affects the performance by tripling the losses in the concentrator. This effect is particularly important for small output surfaces corresponding to strong light recycling in the concentrator and to average travel distances inside the medium reaching meters.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
InGaN blue light-emitting diodes (LED) became a mature technology not only at the center of the domestic and automotive lighting market but also for other applications in the field of machine vision, medical lighting, digital projection, display, and metrology. LEDs can be characterized by their power and by their brightness corresponding to the power divided by the emitting area and emitting solid angle (expressed in W/cm2/sr in radiometric units). Both parameters have improved significantly since the design of blue light-emitting diodes in 1994 by Nakamura et al [1]. Starting from a brightness estimated to 0.05 W/cm2/sr [1]. This evolved quickly to 0.3 W/cm2/sr in 1999 [2], 10 W/cm2/sr in 2007 (K2 line by Lumiled), 30 W/cm2/sr in 2015 (Luxeon Z line by Lumiled) and 80 W/cm2/sr in 2020 (Rubix line by Lumiled). With a typical power at the watt level on a 1 mm2 emitting area, LED performance face fundamental limitation caused by the “efficiency drop” at high injected current [3]. Fortunately, LED packaging has considerably progressed. With high density compact chip-scale package, LEDs can be arranged as “pixels” in large arrays for power scaling. However, the brightness of the array can only be degraded compared to a single LED because of the limited filling factor of LEDs emitting surfaces on the array. Despite all LED’s advantages (ruggedness, long lifetime, efficiency, and low price), many applications in illumination like RGB projection, 3D inspection, or treatment in dermatology, require a combination of power and brightness that cannot be addressed by LED even arranged in arrays.
The brightness conservation theorem states that an emitting surface cannot be reimaged to a higher brightness than the source itself using a passive ray-optical system [4]. Hence, to exceed LED’s performance, alternative light sources must be imagined out of the scope of this theorem. An idea coming in the late 1990s is based on wavelength conversion, associating InGaN blue LEDs, as pump sources, with a cerium-doped yttrium aluminum garnet (Ce:YAG) phosphor to generate white light [5,6]. Indeed, Ce:YAG is a highly chemical stable phosphor [7] showing excellent spectroscopic properties with a emission peaking at the maximum human eye sensitivity. Ce:YAG absorption is centered at 460 nm and its emission spectrum peaks at 550 nm and with a FWHM bandwidth of 100 nm. In the mid-2010s, to increase the brightness, research groups proposed to replace blue LEDs by high brightness laser diodes. Laser activated remote phosphor [8] create small fluorescent point sources with highly doped Ce:YAG small plate [9–10]. The brightness reaches 560 W/cm2/sr (3000 cd/mm2) but with a power limited to the watt level (850 lumens) [9] by the power of laser diodes and by thermal management in Ce:YAG. In addition, laser diodes are less robust than LED and their typical lifetime is approximately 10 000 hours whereas LEDs only lose 20% of their optical power after 50 000 hours of use. The next step for brightness and power increases was achieved by pumping large plates of Ce:YAG with blue LED arrays. The parallelepiped, polished on all faces, acts as a waveguide for the light emitted by Ce3+ ions. The largest faces are used to pump the crystal and one of the smallest lateral faces is used as output face. This concept has been used for many years in solar luminescent concentrators [11] offering a brightness enhancement with respect to the sun thanks to the absorption/emission process combined with the geometry of the parallelepiped. Pumped by LED, this setup is referred as “LED-pumped luminescent concentrator” (LED-LC). Such device can deliver peak powers reaching 1 200 W/cm2/sr in quasi-continuous wave operation [12]. In continuous wave operation the brightness reaches the 100 W/cm2/sr range (230 W/cm2/sr [13] and 110 W/cm2/sr [14]). With this unique combination of power and brightness for incoherent sources, LED-LC have been successfully used for solid-state laser pumping [12, 15, 16] medical imaging [13] and digital projection [14].
Recently, it has been proposed to confine the light inside a LC in the three dimensions of space via total internal reflections (TIR) on the 4 faces parallel to the output axis and by using mirrors on the faces perpendicular to the output axis [17]. This setup, referred as 3D luminescent concentrators (3D-LC), has the potential to increase the brightness of the LC by one order of magnitude. The first proof of concept was performed at low pump power: a 1 W blue laser diode was shaped in a large and homogeneous beam corresponding to the size of the largest face of a Ce:YAG luminescent concentrator.
In this paper, we propose a step further by pumping a 3D Ce:YAG luminescent concentrator with high power blue LED arrays. Experiment and performance are described in a first part. Limitations of performance, attributed to excited state absorption (ESA) are experimentally investigated and analytically described in a second part.
2. Experimental setup and performance
2.1 Experimental setup
Blue LEDs with an emission spectrum centered on 450 nm and a spectral full width at half maximum of 20 nm are used in this experiment (Luxeon Z royal blue from Lumiled whose spectrum can be found in [12]). This chip-scale package LED has been chosen for its compacity (1 × 1 mm2 emitting area for a 1.7×1.3 mm2 package). The quasi-continuous wave operation allows to overdrive LED above their nominal driving current (1 A). When the LED is activated during 10 µs at 10 Hz, LED’s driven current can be increased up to 7 A resulting in an optical peak power of 4.2 W (peak brightness of 133 W/cm2/sr). In this quasi-continuous wave operation, the corresponding forward voltage is 5.2 V and the electro/optic efficiency of the LED is 12%. The compactness of the LED package allows to build high-density LED arrays brazed on a water-cooled aluminum printed circuit board (PCB). Each PCB is composed of 280 LEDs divided in 35 lines of 8 LEDs. 8 PCB (gathered by 4) are positioned side by side resulting in two 200×14 mm2 LED arrays of 1120 elements. This corresponds to a filling factor (portion of emitting surface on the array) of 42%.
The luminescent concentrator is composed of two l=100 mm, w=14 mm and t=1 mm Ce:YAG slabs bonded with a UV-curing optical adhesive resulting in a total length of 200 mm (Fig. 1). All the sides of the LC are optically polished to scratch/dig 60/40. The Ce3+ concentration is 9.9 1018 cm-3 corresponding to an average absorption coefficient of 18 cm–1 over the LED emission spectrum. The PCBs are pump close to the large faces of the concentrator (< 1mm). Taking the filling factor of the PCBs (42%) in account, we estimate the average pump intensity on the Ce:YAG large faces at 1.8 W/mm2.
Fig. 1. Schematic representation of the blue LED-pumped Ce:YAG luminescent concentrator used in this experiment.
A HR back mirror is positioned on one of the w×t face which corresponds to the facet opposite to the output. To complete the light recycling process in the third dimension, two D-shaped mirrors are put close to the output face reducing the output surface (area denoted s). The two D-shaped mirrors are fixed on a digital sliding caliper which enable to measure and to control the length of the aperture. The ratio r = wt/s (r ≥ 1) can be varied by a translation of the mirrors in a direction perpendicular to the output direction. All the mirrors used in this experiment have a reflectivity of 99.4% on the emission range of Ce:YAG (E02 coating from Thorlabs).
2.2 Performance of the LED-pumped 3D luminescent concentrator and evidence of ESA limitation
The LEDs operate in quasi-continuous wave operation (10 µs at 10 Hz). An integrating sphere coupled to a spectrometer (LCS-100 from Labsphere) is placed directly in front of the w×t output face to measure the average power emitted by this surface. To evaluate precisely the emitted peak power over the 10 µs pumping window, the duration of the emitted pulses is recorded by a photodiode put on a leakage of the concentrator.
The evolution of the power emitted by the w×t face of the LC as a function of the peak power emitted by the LED array is shown in Fig. 2. The LED are driven with a current from 1 A to 7 A with a step of 1 A. The corresponding pump peak power ranges from 2.9 kW to 9.4 kW. For r=1 (output face without mirrors), the emitted peak power reaches 724 W and the brightness 1.6 kW/cm2/sr. For an output surface close of 1 mm2 (r=14) the peak power and brightness of the LC are respectively 145 W and 4.6 kW/cm2/sr. The smallest aperture achieved is 0.2 mm2 (r=70) for which a power of 33 W and a brightness of 5.3 kW/cm2/sr are reached.
Fig. 2. (a) Power emitted the LC versus the pump power for r=1, r=14 and r = 70 (r = wt/s). In (b) the r=1, r=14 and r = 70 curves are normalized in order to have the same tangent to the origin which describes the linear behavior without ESA.
Compared to the expected proportional behavior, we observe a discrepancy of the output power, more important for higher r values (Fig. 2(b)). For r=1, the power drops by 14% at maximum pump power compared to a linear behavior. For an output surface of 1 mm2 (r=14), the power drops by 50% at maximum pump power. Figure 2 indicates that a pump-intensity-dependent phenomenon is involved. As the repetition rate (10 Hz) and the activation duration (10 µs) are extremely low, this drop cannot be attributed to thermal effects. Previous works related to the spectroscopy of Ce:YAG under high pump intensities [8] suggest that two main loss mechanisms can occur: energy transfer upconversion (ETU) to the conduction band and excited-state absorption (ESA). ETU corresponds to energy transfer involving two Ce3+ ions in the excited state situated in nearby lattices. ETU depends quadratically on the concentration. It is classically observed in Ce:YAG for Ce3+ doping concentration higher than 1% and nearly not detectable for a concentration around 0.2% [8]. As our concentration is estimated to 0.11%, this effect can be neglected. The effect of ESA in Ce:YAG has been studied in previous works [18–21]. The amplitude of ESA is wavelength dependent [18]: being much higher above 600 nm than below. Hence, the distortion of the emitted spectrum for different pump powers and/or different values of r can be a signature of the presence of ESA. To analyze this effect, we compared the emitted spectra for two power levels 2.9 kW (1A driving current) and 9.4 kW (7A driving current) corresponding to the lowest and highest pumping level possible, taking the LED driver in account. The spectra are also recorded for two values of r (r=1 and r=28). For the sake of comparison, the spectra are normalized by their area, like any lineshape function (Fig. 3). One can observe a spectral narrowing on the red part for the highest pumping level compared to the lowest pumping level. This effect could be related to the presence of ESA, the population of the excited state being an increasing function of the pump power. The spectral narrowing is more visible for r=28 than for r=1. This can be attributed to a higher travel distance for r=28 than for r=1 for the rays before finding the output.
Fig. 3. Normalized spectra (lineshape functions) emitted by the LC at two pump powers 2.9 kW and 9.4 kW for r=1 (a) and r=28 (b).
Table 1 gives an overview of the LED array and luminescent concentrator performance. Despite the limitations observed, we obtained a record brightness above 5 kW/cm2/sr in quasi-continuous wave operation overcoming the LED brightness by a factor 40.
Table 1. Characteristics of the light sources used or build in this article. All the sources are driven in the same quasi-continuous wave operation (10 µs, 10 Hz).*Luxeon Z royal blue by Lumiled.
As the concentrator emits in the visible spectrum, it is interesting to estimate its power in lumens, taking the spectrum of the source and the response of the eye versus the wavelength (Fig. 4). Calculations are done for the r=14 performance, presenting the best compromise between power and brightness. The source presents a luminous efficacy of 525 lm/W. It emits a power of 7.6 104 lm with a brightness of 2.4 104 cd/mm2. Compared to laser activated remote phosphors using Ce:YAG, this is one order of magnitude higher for the brightness and nearly two order of magnitude higher for the power.
Fig. 4. Photopic luminosity function (eye’s responsivity) and Ce:YAG luminescent concentrator emission spectrum (r=14).
3. Role of ESA in a 3D LED-pumped Ce:YAG luminescent concentrator
Excited State Absorption (ESA) has two main consequences on the Ce:YAG LC. It reduces the population density of the excited state (n2) and it induces additional propagation losses (αESA=σESAn2 with σESA the ESA cross section). The aim of this section is to describe analytically how ESA affects the output power emitted by the LED-LC and to compare this model with the experimental results of the output powers.
3.1 ESA and trapped rays
In order to simplify the approach, it is assumed that the LC emits at a single wavelength (the mean wavelength of the Ce:YAG LC emission spectrum is 580 nm) and we consider a mean value of the ESA cross section over the Ce:YAG emission spectrum. In a same way, reabsorption effects are integrated over all the emission spectrum and taken in account as part of the loss coefficient. As developed in [18], the stimulated emission cross section is neglected with respect to the value of the ESA cross section. In the ESA process, an electron of the excited state of a Ce3+ ion is sent to the conduction band and then relaxes quickly either in traps, or in the excited state, or in the ground state. The recombination branching ratio (fraction of Ce3+ ion from the conduction band population that recombines non-radiatively into the excited state) is denoted f and estimated to be 0.5 following [8].
As the pump duration (10 µs) is considerably longer than the Ce3+ emission lifetime (64 ns [19]), the calculation of the population density n2 can be carried out in the steady state. Assuming that the decay rate from the conduction band is much higher than the other rates, n2 can be written as :
where, ${\sigma _p}$ is the absorption cross section of Ce:YAG on the LED spectrum, ${n_t}$ is the total population density, A is the radiative decay rate, ${I_p}$ and ${I_e}$ are respectively the photonic intensities at the pump and emitted wavelengths. For simplicity, we assume that ${I_p}$ and ${I_e}$ are homogeneous in all the LC volume, leading to a constant value for ${n_2}$ in all the Ce:YAG. By averaging the pump intensity over all the concentrateur, ${I_p}$ can be written as :Table 2. Parameters used for the analytical description of the power emitted by the LC.
In Eq. (1), one key point is the intensity inside the LC at the emitted wavelength ${I_e}$. Assuming that a total internal reflection is lossless, ${I_e}$ is mostly due to indefinitely trapped rays with no possible escape of faces. They represent a proportion $({3\cos {\theta_{TIR}} - 2} )$ of the total emitted power, corresponding to 51% in the case of Ce:YAG in air (${\theta _{TIR}}\; $ is defined in Table 2). At a given point inside the LC, trapped rays superimpose incoherently after many reflections. In other words, each time that a trapped ray is reflected, the reflection increases the intensity inside the LC. The distance between two reflections can be estimated assuming that reflections occur mostly between the two large faces. Observing that the main angular directions of trapped rays corresponds to an annulus between ${\theta _{TIR}}$ and $90^\circ \; - \; {\theta _{TIR}}$ (Fig. 5), the distance between two reflections is estimated to 1.44t for Ce:YAG, by averaging all the possible distances between the two largest faces for trapped rays.
Fig. 5. Representation of the internal rays in a luminescent concentrator. The dotted cap shows the main location of trapped rays considered in the analytical model.
In practice the number of reflections is not infinite because of the propagation losses and because of LC surface imperfections (imperfect chamfers on the edges, diffusion or dusts on the surfaces and quality of the surfaces). Using the concept of light recycling, ${I_e}$ can be estimated by the sum of the terms of a geometric series as follows (see Supplement 1):
3.2 Output power for a 2D-luminescent concentrator (r=1)
As a first step, we suppose that the output surface is the total w×t face (r=1). This configuration refers as LED-LC operating with a confinement in 2 dimensions. Considering ESA and based on [15], the power emitted by the LC on the w×t face is:
The ESA absorption cross section used in the analytical model is set to σESA = 12 10−18 cm2 for an optimal fit of the measurements. There is a strong fluctuation on the values of σESA in the literature [8,19]. Our analytical model requires a mean value of σESA weighted by the emission spectrum of Ce:YAG which is in the same order of magnitude than the values of the literature.
Equation (4) is used to plot the analytical red curves (r=1) on Fig. 2 and fits the experimental observation. The analytical model shows that ESA gives a good description of the performance deviation compared to a linear evolution. The analytical model can help to estimate the two effects of ESA : the reduction of the population density of the excited state and the additional propagation losses. On Fig. 6(a), we plot the evolution of the excited population density n2 versus the pump power. It shows that ESA is responsible for a small variation of the population. As the output power P2D for the LED-LC operating at full aperture (r=1) is proportional to n2 (see Eq. (4)), this small variation does not affect the output power significantly. On the contrary, Fig. 6(b) shows a strong contribution of ESA to the loss coefficient, tripling the value at maximum pump power. This effect is mainly responsible for the 14% power drop on Fig. 2(b) for r=1.
Fig. 6. Evolution of the population density of the excited state level (a) and the ESA loss coefficient (b) with the pump power.
3.3 Output power for a 3D-luminescent concentrator (r>1)
In this section, light recycling occurs in the three dimensions of space, thanks to the back and the front mirrors. The output surface is reduced following the position of the front mirror, and its area is noted s (r = wt/s). We assume that the light recycled inside the LC by the mirrors has a negligible impact on ${I_e}$ because it undergoes much less reflections than the trapped light. Hence, it is assumed that the value of n2 depends only on the pump power and remains constant with r. Based on [15] and considering ESA, the power emitted by the output surface of the LC is:
Fig. 7. Evolution of the transmission of the additional recycled light with r for the higher and lower pump powers.
3.4 Brightness enhancement of the light recycling process at high pump power
The brightness enhancement (B3D/B2D) by light recycling is described in [15]. The behavior of B3D/B2D corresponds to the sum of the terms of a geometric series as a small amount of light is emitted by the aperture after each roundtrip inside the luminescent concentrator. The equation in [15] is adapted to the presence of ESA in Eq. (8):
Fig. 8. Measured brightness enhancement of LC versus r for several LED pump power. The doted lines correspond to Equ.8.
Here, the brightness enhancements are lower than in [15] performed at low pump power level (best value of 13.5). This difference is explained by three main elements. Firstly, the length of the LC is 200 mm instead of 100 mm, this parameter has a great impact on the travelled distance and thus on the losses. Secondly, the presence of optical adhesive contacting the two Ce:YAG together, induces additional losses (Table 1). Finally, the presence of ESA generates additional losses at high pump power levels.
Including these 3 parameters in our model allows to retrieve the brightness enhancement factors correctly (Fig. 8). It is then obvious that high brightness and high power luminescent concentrators using Ce:YAG are challenging and require to take into account limiting phenomenon such as ESA that was up to now neglected. Compared to previous studies on Ce:YAG limitations using laser diode-pumped Ce:YAG “phosphor” [8], this result could be surprising at a first glance since the average pump intensity with LED is only 1.8 W/mm2 on the Ce:YAG faces, hundred times lower than the pump intensity of blue laser diodes. Likewise, the doping concentration is only of 0.11% in our case, 10 times lower than Ce:YAG phosphors. However, the main difference between a Ce:YAG phosphor and a Ce:YAG concentrator is the propagation length inside the medium : less than 1 mm in phosphor and higher than 1 m in a concentrator. This makes the concentrator very sensitive to loss variation hence to ESA.
4. Conclusion
We demonstrated a 3D Ce:YAG luminescent concentrator pumped by 2 arrays of blue LEDs in quasi-continuous wave operation. Contrary to other 2D LED-LC, this 3D LED-LC can provide a small and squared output surface making the beam shaping easier for applications. For a 1 mm2 output surface, corresponding to a standard LED surface, the source emits a power of 145 W with a brightness of 4.6 kW/cm2/sr, exceeding LED performance by a factor 35. By taking the spectral shape in account, this visible 3D LED-LC emit a luminous flux of 7.6 104 lm and a brightness of 2.4 104 cd/mm2 overcoming other visible sources like arc lamp, laser activated remote phosphor, and 2D luminescent concentrators typically by one order of magnitude and standing among the brightest incoherent light source ever build. Relying on the stability, ruggedness and lifetime of the LED technology, this source has the potential to become a reference for a large panel of applications.
This work gives a deep understanding of the performance limitation of LED-LC versus the pump power. The pump power increases the population density of the excited state and then increases the losses induced by excited state absorption. By this effect, the loss coefficient is tripled at maximum pump power. The analytical model estimating the output power in the presence of ESA is close to the experimental observations validating the role of ESA in LED-LC performance.
We believe that this work gives major keys to pave the way for new broadband sources combining power and exceptionally high brightness.
Funding
Centre National de la Recherche Scientifique (CNRS pre-maturation LEDsGO project); Agence Nationale de la Recherche (ANR-10-LABX-0039-PALM).
Disclosures
The authors declare no conflicts of interest.
Data Availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
Supplemental document
See Supplement 1 for supporting content.
References
1. S. Nakamura, T. Mukai, and M. Senoh, “Candela-class high-brightness InGaN/AlGaN double-heterostructure blue-light-emitting diodes,” Appl. Phys. Lett. 64(13), 1687–1689 (1994). [CrossRef]
2. T. Mukai, M. Yamada, and S. Nakamura, “Characteristics of InGaN-based UV/blue/green/Amber/red light-emitting diodes,” Jpn. J. Appl. Phys. 38, 3976–3981 (1999). [CrossRef]
3. J. J. Wierer, J. Y. Tsao, and D. S. Sizov, “Comparison between blue lasers and light-emitting diodes for future solid-state lighting,” Laser Photonics Rev. 7(6), 963–993 (2013). [CrossRef]
4. R. W. Boyd, “Radiometry and the Detection of Optical Radiation,” Wiley (1983).
5. Y. Shimizu, K. Sakano, Y. Noguchi, and T. Moriguchi, Japanese priority patent applications to U. S. Patent 5, 925 (1996).
6. P. Schlotter, R. Schmidt, and J. Schneider, “Luminescence conversion of blue light emitting diodes,” Appl. Phys. A 64(4), 417–418 (1997). [CrossRef]
7. S. Arjoca, E. G. Villora, D. Inomata, K. Aoki, Y. Sugahara, and K. Shimamura, “Temperature dependence of Ce:YAG sing-crystal phosphors for high-brightness white LEDs/LDs,” Mater. Res. Express 2(5), 055503 (2015). [CrossRef]
8. A. Lenef, M. Raukas, J. Wang, and C. Li, “Phosphor performance under high intensity excitation by InGaN laser diodes,” ECS J. Solid State Sci. Technol. 9(1), 016019 (2020). [CrossRef]
9. Y. Yang, S. Zhuang, and B. Kai, “High brightness laser-driven white emitter for Etendue-limited applications,” Appl. Opt. 56(30), 8321 (2017). [CrossRef]
10. S. Li, L. Wang, N. Hirosaki, and R. Xie, “Color conversion materials for high-brightness laser driven solid-state lighting,” Laser Photonics Rev. 12(12), 1800173 (2018). [CrossRef]
11. Mehran Rafiee, Subhash Chandra, Hind Ahmed, and Sarah J. McCormack, “An overview of various con gurations of Luminescent Solar Concentrators for photovoltaic applications,” Opt. Mater. 91, 212–227 (2019). [CrossRef]
12. P. Pichon, A. Barbet, J-P. Blanchot, F. Druon, F. Balembois, and P. Georges, “Light-emitting diodes, a new paradigm for Ti:sapphire pumping,” Optica 5(10), 1236–1239 (2018). [CrossRef]
13. J. Sathian, J. D. Breeze, B. Richards, N. M. Alford, and M. Oxborrow, “Solid-state source of intense yellow light based on a Ce:YAG luminescent concentrator,” Opt. Express 25(12), 13714–13727 (2017). [CrossRef]
14. D. K. G. de Boer, D. Bruls, and H. Jagt, “High-brightness source based on luminescent concentration,” Opt. Express 24(14), A1069–1074 (2016). [CrossRef]
15. A. Barbet, A. Paul, T. Gallinelli, F. Balembois, J-P. Blanchot, S. Forget, S. Chénais, F. Druon, and P. Georges, “Light-emitting diode pumped luminescent concentrators : a new opportunity for low-cost solid-state lasers,” Optica 3(5), 465–468 (2016). [CrossRef]
16. P. Pichon, A. Barbet, D. Blengino, P. Legavre, T. Gallinelli, F. Druon, J.-P. Blanchot, F. Balembois, S. Forget, S. Chénais, and P. Georges, “High-radiance light sources with LED-pumped luminescent concentrators applied to pump Nd:YAG passively Q-switched laser,” Opt. Laser Technol. 96, 7–12 (2017). [CrossRef]
17. P. Pichon, F. Balembois, F. Druon, and P. Georges, “3D luminescent concentrators,” Opt. Express 29(5), 6915 (2021). [CrossRef]
18. M.J. Weber, “Nonradiative decay from 5d-states of rare earth in crystals,” Solid State Commun. 12(7), 741–744 (1973). [CrossRef]
19. D.S. Hamilton, S.K. Gayen, G.J. Pogatshnik, and D. Ghen, “Optical-absorption and photoionization measurements from the excited states of Ce3+:Y3Al5O12,” Phys. Rev. B 39(13), 8807–8815 (1989). [CrossRef]
20. R.R. Jacobs, W.F. Krupke, and M. J. Weber, “Measurement of excited-state-absorption loss for Ce3+ in Y3Al5O12 and implications for tunable 5d-4f rare-earth lasers,” Appl. Phys. Lett. 33(5), 410–412 (1978). [CrossRef]
21. W. J. Miniscalco, J. M. Pellegrino, and W. M. Yen, “Measurements of excited-state absorption in Ce3+:YAG,” J. Appl. Phys. 49(12), 6109–6111 (1978). [CrossRef]
