Figure 1(a) shows the XRD patterns of LuAG TCs. All the diffraction peaks could be well indexed as the pure LuAG phase (PDF 97-018-2354), and there is no impurity phase (e.g., LuAM, LuAP, CeO2 and Cr2O3) observed, indicating that the solid-state reaction between Lu2O3 and Al2O3 was completely finished during vacuum sintering. Increasing Cr3+ doping concentration hardly changed the phase composition of TCs, and there was no obvious peak shift observed from the main diffraction peaks of XRD patterns, owing to the low doping concentration of Cr3+ ion. Generally, Cr3+ ion (0.615 Å, CN = 6) would occupy the octahedral Al3+ site (0.53 Å, CN = 6), and Ce3+ ion (1.143 Å, CN = 8) would substitute the dodecahedral Lu3+ site (0.977 Å, CN = 8) in LuAG lattice [7, 27], owing to their similar ionic radius, as is shown in the schematic crystal structure sketch of Ce,Cr:LuAG TC in Fig. 1(b).
Appearances and in-line transmission spectra of the polished LuAG TCs are shown in Fig. 2. All the samples exhibited a transparent appearance, and the words behind them could be clearly recognized by the naked eyes. The color of the TC without Cr3+ doping (Ce01Cr0) was yellowish green, i.e., the intrinsic color of Ce3+ ion. With increasing Cr3+ doping concentration, the color of the TCs was changed from yellowish green to light green, indicating that the emission of the samples was tuned effectively by Cr3+ ion incorporation. From the transmission spectra it could be seen that moderate amounts of Ce3+ and Cr3+ doping hardly affected the transparency of TCs, and their transmittances at 800 nm were close to 70%. Increasing Cr3+ doping concentration deteriorated the transparency of TCs. The variation trend of the transmittances at 800 nm and 400 nm of the prepared TCs could be found in Fig. 2(b). Two broad absorption bands centered at 340 nm and 445 nm were ascribed to the 4f-5d1 and 4f-5d2 transitions of Ce3+ ion, respectively [28, 29]. Besides the intrinsic absorption bands of Ce3+ ion, the absorption centered at 430 nm and 596 nm could be observed from all the Cr3+ ion doped samples, corresponding to the 4A2-4T1 and 4A2-4T2 transitions of Cr3+ ion, respectively .
Figure 3 shows the SEM micrographs of the sintered LuAG TCs. From the fracture surfaces of the samples (Fig. 3(a)-(h)) it could be seen that the fracture modes of all the samples were characterized by both intergranular and transgranular, and residual pores could be observed simultaneously, providing a quasi-densified microstructure. As can be seen from the polished surfaces of TCs (Fig. 3(a’)-(h’)), the amounts of residual pores were increased with increasing Cr3+ doping concentration, indicating that Cr2O3 affected the densification behavior of LuAG TCs during sintering. Also, feature of the residual pores was characterized by intergranular pores. Because the applied sintering temperature in this study was 1800 oC, which was lower than the ideal sintering temperature of LuAG TC, resulting in the intergranular pores in ceramic bulks. Generally, for the real application of TCs as color convertors for white LEDs/LDs, residual pores could act as light scattering centers to increase the light extraction rate [22, 31–33].
The variation trend of grain size as a function of Cr3+ doping concentration could be found in Fig. S1 of the Electronic Supplementary Information (ESI†), and it is obvious that the grain size of TCs was moderately increased with increasing Cr3+ doping concentration. However, the grain size of the fabricated TCs was only increased from 2.79 µm to 3.78 µm, owing to the relative low sintering temperature. Additionally, EDS mapping of the sintered Ce01Cr04 sample is shown in Fig. S2 in the ESI† to investigate its elemental distribution, and it could be found that all the adopted elements (Lu, Al, O, Ce and Cr) were distributed homogeneously inside ceramic bulk, indicating that both Ce3+ and Cr3+ ions were solid soluted into LuAG lattice without segregation [22, 31–33].
PL and PLE spectra of Ce0Cr01 sample is shown in Fig. 4(a). Recording at 687 nm, two broad absorption bands centered at 428 nm and 593 nm could be observed, originating from the 4A2-4T1 and 4A2-4T2 transitions of Cr3+ ion, respectively, leading to the characterized green color of Cr3+ doped LuAG TCs. By exciting Ce0Cr01 TC under 428 nm, an intensive broad emission band covered the orange-red and red color regions centered at 710 nm was obtained from the PL spectra. This emission was corresponded to the spin-allowed 4T2→4A2 transition of Cr3+ ion. Simultaneously, a sharp R line (zero-phonon line) due to the spin-forbidden 2E→4A2 transition could be detected at 690 nm, and a similar observation has been reported by researchers in Cr3+ doped garnet structured materials .
Figure 4(b) shows the normalized PLE spectra of Ce,Cr:LuAG TCs with different Cr3+ doping concentrations (λem = 523 nm). Two excitation bands centered at around 340 nm and 455 nm could be observed, corresponding to the 4f-5d1 and 4f-5d2 transitions of Ce3+ ion, respectively. Notably, a unilateral red shift phenomenon was observed from the left wing of the 410–500 nm band with increasing Cr3+ doping concentration, whereas the right wing of this band was not influenced by Cr3+ doping. Combining with the PLE spectrum of Ce0Cr01 sample shown in Fig. 4(a), it could be deduced that a portion of the emitted photon of Ce3+ ion was absorbed by Cr3+ ion in Ce,Cr:LuAG TCs, resulting in a unilateral red shift of their PLE spectra . Besides, it was speculated that the complicated local crystal environment around Ce3+ ions by doping Cr3+ ions into the [AlO6] octahedron might be another reason that caused the unilateral red shift phenomenon .
Normalized PL spectra of Ce,Cr:LuAG TCs are displayed in Fig. 4(c). The characteristic broad emission band centered at 523 nm could be clearly resolved, corresponding to the 5d-4f transition of Ce3+ ion. The emission of Cr3+ ion could be detected simultaneously from all the Cr3+ doped samples. It could be clearly seen that the emission of Ce3+ ion overlapped with the absorption of Cr3+ ion. Therefore, Ce3+ ions in Ce,Cr:LuAG TC could not only acted as luminescence centers, but also as sensitizers proceeding the energy transfer from Ce3+ to Cr3+ ions, in which a portion of the photons emitted from Ce3+ ions could be absorbed by Cr3+ ions to realize red light emission. The energy transfer process from Ce3+ to Cr3+ ions could be processed through two means, i.e., radiative transition and non-radiative transition, and the detailed schematic diagram of the energy transfer process from Ce3+ to Cr3+ ions could be found in Fig. S3 (ESI†) . The emission intensity of Cr3+ ions was increased with increasing Cr3+ doping concentration, and reached the maximum when the Cr3+ concentration was 0.5 at.% (Fig. 4(c)). Further increasing Cr3+ doping concentration decreased the emission intensity of TC, thanks to the concentration quenching effect. Besides, the right wing of the Ce3+ emission band was blue shifted as increasing Cr3+ concentration, whereas there was no obvious shift observed from the left wing of the band. This unilateral blue shift was owing to the increased absorption at around 593 nm that overlapped the PL spectra of Ce3+ ion as increasing Cr3+ ion doping concentration, which was similar to that of the observed unilateral red shift from the PLE spectra shown in Fig. 4(b). The detailed full width at half maximum (FWHM) values of both PL and PLE spectra of Ce3+ ion are presented in Fig. 4 (d), illustrating that Cr3+ ion doping could regulate the PL spectra of Ce3+ ion in Ce,Cr:LuAG TCs effectively.
In order to further validate the availability of the prepared TCs as fluorescent convertors, TC based white LED devices were constructed using the remote excitation mode. The operating power and the emitting wavelength of the blue LED chips were 20 W and 460 nm, respectively. Figure 5(a) shows the electroluminescent (EL) spectra of the TCs. It was evident that the sample without Ce3+ doping had a strong blue light emission, since the absorption ability of Cr3+ ion at 460 nm was far inferior than that of Ce3+ ion. Besides, the green component corresponding to the emission of Ce3+ ion was decreased with increasing Cr3+ doping concentration, thanks to the enhanced energy transfer from Ce3+ to Cr3+ ions. The variation trend of the chromaticity parameters is displayed in Fig. 5(b), and it could be found that with increasing Cr3+ doping concentration, the luminescence characteristic of white LEDs was changed from greenish to blueish.
Figure 5(c) shows the CRI values of Ce,Cr:LuAG TCs. It was obvious that the CRI values were increased with increasing Cr3+ doping concentration, and reached the maximum value of 75.7 for Ce01Cr03 TC. Further increasing Cr3+ doping concentration decreased the CRI values. The deteriorated CRI value was due to the proportional mismatch among red/green/blue light. Despite the optimized CRI value of 75.7 was not very ideal, it was much higher than that of the sample without Cr3+ doping, and the increment was as high as 46.2%.
The detailed variation trend of the red, green and blue light proportion is shown in Fig. 5(d). According to the system setting of CAS-200 software, the spectral bands of red/green/blue light were 600–780 nm, 500–600 nm and 380–500 nm, respectively, and the corresponding light proportion was obtained by calculating the ratio of the luminous flux of red/green/blue light to the total luminous flux collected by the integrating sphere. It could be seen from Fig. 5(d) that the red component was increased monotonously, whereas the greenish yellow light component was descended simultaneously. Because the blue light emitted from the chip was absorbed by the entire surface of TC under the remote excitation mode, and all the Cr3+ ions in TCs were not reached the saturation status, resulting in the continuous increased red light proportion in Ce,Cr:LuAG TC based white LEDs.
Fluorescence decay curves of Ce,Cr:LuAG TCs are plotted to further explore the energy transfer process from Ce3+ to Cr3+ ions under 460 nm excitation, as is shown in Fig. 6. The decay behavior of all the TCs could be fitted well by the single exponential decay function. From Fig. 6 it was obvious that with increasing Cr3+ doping concentration, lifetimes of Ce,Cr:LuAG TCs at 523 nm presented a monotonically decreasing trend, which was ranged from 55.52 ns (Ce01Cr00) to 33.47 ns (Ce01Cr04), illustrating an effective energy transfer from Ce3+ to Cr3+ ions.
The energy transfer efficiency was determined according to Eq. (1) [17, 37]:
η T = 1-τ/τ0 (1)
where τ and τ0 are the average lifetimes of the donor Ce3+ ions in the presence and without Cr3+ ions, respectively. With increasing Cr3+ doping concentration, the calculated ηT of Ce,Cr:LuAG TCs were 12.76%, 22.42%, 31.93% and 39.82%, respectively. It revealed that the energy transfer efficiency of TCs was promoted effectively by Cr3+ ion doping. The above results were in coincidence with that of the PL and EL spectra, demonstrating that Cr3+ ion doping is an effective approach to regulate the luminescence behavior of Ce:LuAG TC.
Judging from the EL spectra shown in Fig. 5, it was evident that the greenish yellow component of the single structure Ce,Cr:LuAG TC was inadequate. Therefore, scheme of combining Ce,Cr:LuAG TCs with a 0.5 at.% Ce:YAG TC was performed to further improve the luminescence performance of TC based white LEDs. Appearance and transmission spectra of the applied 0.5 at.% Ce:YAG TC could be found in Fig. S4 in the ESI†. Figure 7(a) presents the schematic sketch of TC based white LED device, in which Ce,Cr:LuAG TC was placed at the top of the LED device, and Ce:YAG TC was placed between Ce,Cr:LuAG TC and blue LED chip. From the chromaticity parameters shown in Fig. 7(b) it could be seen that the color hue was regulated effectively by using the “ceramic combination strategy”, indicating the luminescence performance of white LEDs was significantly improved.
Figure 8 indicates the EL spectra and the corresponding CRI values of the white LEDs constructed with Ce:YAG and Ce,Cr:LuAG TCs as color convertors. It was noteworthy to see that the CRI values of the white LEDs were drastically promoted by using the “ceramic combination strategy”, which was in consistence with the optimized chromaticity parameters shown in Fig. 7(b). Surprisingly, by combining Ce:YAG TC with Ce01Cr04 TC, the obtained CRI value of white LEDs was as high as 88.0. Therefore, ratio of red/green/blue emission of white LEDs constructed with the combined TCs was more reasonable, compared with that of the white LEDs constructed with the single structure Ce,Cr:LuAG TCs. So it is obvious that the real sense emitting color of the assembled white LEDs was adjusted effectively by Cr3+ ion doping, as well as combining Ce,Cr:LuAG TCs with Ce:YAG TC, which is shown in the inserts of Fig. 8. Furthermore, the applied Ce3+ doping concentration in Ce,Cr:LuAG TCs could be further optimized, in order to obtain white light emission with a considerable CRI.
With respect to laser lighting test, the as-fabricated Ce,Cr:LuAG/Ce:YAG TCs were fixed on the aluminium alloy support frame with excellent heat dissipation performance to eliminate the possible thermal induced luminescence attenuation, and the appearance of the apparatus for LD measurement is shown in Fig. 9(a). It was composed of an integrating sphere and the constructed LD device. The detailed structure of the applied LD device could be found in our previous work . Similar to the designation of the white LED device shown in Fig. 7, a 0.5 at.% Ce:YAG TC was also placed underneath the Ce,Cr:LuAG TCs to evaluate the luminescence performance of the constructed white LDs, and the applied excitation wavelength of the laser source was 450 nm. Also, the selected pump power of the laser source was 1 W, in order to avoid excessive local temperature that could deteriorate the luminescence performance of TCs during LD operation.
EL spectra of the combined Ce,Cr:LuAG/Ce:YAG TC based white LDs are shown in Fig. 9 (b). The sharp emission peak located at 450 nm was corresponded to the radiation of laser source with a narrow emission band. The emission bands of both Ce3+ and Cr3+ ions were clearly resolved when exciting TCs by a 450 nm blue laser source, and a distinct energy transfer from Ce3+ to Cr3+ ions could be observed simultaneously. Therefore, in addition to LED sources, the prepared LuAG TCs could also be excited effectively by blue laser sources with high energy density and tiny radiation area.
Figure 9(c) indicates the CRI values of the combined Ce,Cr:LuAG/Ce:YAG TCs as a function of Cr3+ doping concentration. Considering a portion of the transmitted blue light is highly directional compared to the emitted light from the TC. In this regard, a 1W blue laser was applied as the exciting source in this study, in order to avoid the side effect that could affect the CRI performance induced by the laser source. It was obvious that the obtained CRI values were regulated effectively by doping Cr3+ ions into Ce:LuAG TC. With increasing Cr3+ doping concentration, the yellow component of the emitted light was reduced, indicating a portion of the yellow light emitted from Ce3+ ions were absorbed by Cr3+ ions to increase the red component of white LDs. It should be noted that the optimized CRI value was as high as 85.5, which was almost identical to that of the TC based white LEDs shown in Fig. 8.
However, from Fig. 9(d) it was found that the luminous flux of white LDs was moderately decreased from 257.3 lm to 218.5 lm with increasing Cr3+ doping concentration. Therefore, there existed a trade-off between CRI and luminous flux. The decreased luminous flux should be attributed to the energy loss originated from the increased yellow light absorption when increasing Cr3+ ion doping concentration. Nevertheless, from Fig. 9(c) it was noteworthy to see that the increment of CRI of white LDs was as high as 59.2%. Consequently, it was worthwhile to sacrifice a small fraction of luminous flux to realize a dramatic increased CRI in Ce,Cr:LuAG/Ce:YAG TC based white LDs. Finally, in addition to white LEDs, Ce,Cr:LuAG TC is also a considerable color convertor for the real applications of white LDs in the future.