Fig. 1. Blackbody spectrum for CCT of 4000 K that is cropped by the blue LED spectrum on the left side, and the red quantum dot spectrum (represented by a Gaussian distribution) on the right side.

Fig. 2. (a) Image of the “empty” LUXEON 3535 LED module. (b) Simulation model of the LED package: 1, LED chip; 2, bottom reflector; 3, inner side of recycling cavity; 4, diffusing bar; 5, package. (c) Reflectance of the LED package and blue chip.

Fig. 3. (a) Simulated and measured spectral power distribution of resin samples with different phosphor loadings: Sample 1 (1  mL resin+0.32 g YAG:Ce), Sample 2 (1  mL resin+0.082 g YAG:Ce), and Sample 3 (1  mL resin+0.025 g YAG:Ce). (b) Simulated/measured spectra of LED coated with 1 mL resin mixed with 0.048 g YAG:Ce. (c) Simulated/measured radiation pattern of LED package filled with clear resin.

Fig. 4. Emission/absorption spectrum of YAG:Ce (545 nm), LuAG:Ce (516 nm), and red InP/ZnSe quantum dots.

Fig. 5. (a) Ra, (b) R9, and (c) LE values of the cropped backbody spectrum (CCT=4000 K) for different red peak wavelengths and FWHM. The purple isocurve connects the points where Ra has a value of 93, and the R9 values on the curve are approximately 50, while LE changes but only slightly. As an example, the three points in panel (c) indicate the exact values of LE for three different FWHMs; one can notice that red QDs with more narrow FWHM must emit at longer wavelengths to achieve the same Ra/R9 values. As a result, differences in LE values are limited. Graph (d) shows the dependence of LE on FWHM, for the case when Ra is 93.

Fig. 6. LED1, measured spectral power distribution of the white QD-LED with Ra=80. This white LED has a CCT ≈ 4100  K and an LE of 143 lm/W.

Fig. 7. LED2, white QD-LED with the cyan region perfectly filled by using the LuAG:Ce phosphor, resulting in Ra>90 (CCT ≈4200  K, x=0.3755, y=0.3888). LED3, measured spectral power distribution of the white LED using LuAG:Ce (516 nm) and red-emitting QDs (611 nm). It has both high Ra=94 and R9=84 (CCT ≈4400  K, x=0.3618, y=0.3557).

Fig. 8. LED4, measured spectral power distribution of the white LED using LuAG:Ce (516 nm), YAG:Ce (545 nm), and red-emitting QDs (611 nm) as luminescent materials. This LED balances color rendering and LE performance.

Fig. 9. Spectra of different white LEDs with similar color rendering performance (Ra=90/R9≈50) for CCT=3000 K and 4000 K. The white LEDs with red-emitting nitride-based phosphor induce clearly much more lumen loss compared to the QD-based LEDs due to the broad red tail that extends far beyond the human eye sensitivity curve.

Fig. 10. (a) Measured spectral power distribution of the demonstrated InP/ZnSe QD-LED4 (CCT=4000 K, LE=132 lm/W) and corresponding simulated spectrum. (b) Pie chart of the various power losses in the demonstrated QD-LED4. The different loss mechanisms were estimated with the developed simulation model in LightTools.

Fig. 11. Predicted luminous efficacy values of the InP/ZnSe QD-LED with CCT = 4000 K for (a) varying wall-plug efficiency of the blue LED chip and (b) varying quantum yield of the InP/ZnSe quantum dots.