Promotion of solid-state lighting for ZnCdSe quantum dot modified-YAG-based white light-emitting diodes

Abstract

In this study, the effect of quantum dot (QD) addition and photoluminescence quantum yield (PL QY) of QDs on the luminous efficacy and color rendering index (CRI) of white light-emitting diodes (LEDs) has been investigated. Two kinds of red-emitting materials, ZnCdSe QDs and commercial CaSiAlN₃:Eu²⁺ (nitride) phosphor, are mixed with commercial Y₃Al₅O₁₂:Ce³⁺ (YAG) phosphor as white LED materials. The as-prepared ZnCdSe QDs have a PL QY of about 44%, which drops to 29% after purification due to the removal of the surfactants. When the as-prepared QDs are used, both the CRI and luminous efficacy of the YAG/QDs-based white LED can be improved to meet the requirements of solid-state lighting (SSL), suggesting that the addition of QDs with high QY can overcome the trade-off property between CRI and luminous efficacy. Besides, the correlated color temperature (CCT) of this white LED can be tuned toward cold light by incorporating YAG with QDs, while it tunes toward warm light for a YAG/CaSiAlN₃-based white LED. This special phenomenon can be solely observed in QD-modified white LEDs. Moreover, the long term operation stability of QD-modified white LEDs is excellent.

---

Introduction

White light-emitting diodes (LEDs), owing to their low level of power consumption and long lifetime, have been applied for lighting and full color displays.^1–3 The United States Department of Energy (DOE) has proposed the following targets for white solid-state lighting (SSL) for general use, such as chromaticity coordinates of pure white light (0.33, 0.33), a color rendering index (CRI) higher than 80, and a luminous efficacy of 200 lm W⁻¹. Because of the trade off effect between CRI and luminous efficacy, in which the CRI and efficacy cannot be improved simultaneously, phosphor converted white LEDs (pc-LEDs) are the most energy efficient option, providing package efficacies greater than 130 lm W⁻¹.⁴

The white light can be generated by red, green and blue (RGB) three chip-based LEDs or pc-LEDs in which the emission from a blue or ultraviolet (UV) LED is combined with phosphors having a longer wavelength.^5–12 In principle, the best strategy is the use of three LEDs because energy is lost in pc-LEDs due to the Stokes shift. However, one drawback of the multi-LED is poor CRI associated with the white light from relatively narrow radiation components.¹³ The UV LED strategy has the advantage of stable colors because changes in the UV emission wavelength will not affect the color output of a device. On the other hand, the pc-LEDs have attracted much interest due to their easy fabrication, low cost, and high brightness.^14,15 Even though, Y₃Al₅O₁₂:Ce³⁺ (YAG)-based white LED suffers from poor color rendering property caused by lack of red emission in the spectrum. This type of white LED has a CRI ranging from 70 to 80.¹⁶ In terms of other yellow phosphors such as Sr₃SiO₅:Eu^2+ 15,17 and Ca-α-SiAlON,¹⁸ since the peak emission from Eu²⁺ is approximately at 570–585 nm, the cool white or daylight cannot be generated or with lower CRI in the Eu²⁺-activated Li-α-SiAlON system.¹⁹ Therefore, a broad-emission phosphor is needed in multi-bands white LEDs in order to have a good CRI. Although mixing several kinds of phosphors will result in high CRI due to the production of broad emission spectrum, the luminous efficacy drops very significantly due to the reabsorption of the emission from other phosphors. Moreover, serious light scattering from phosphors with a particle size in several microns and large Stokes shift, decrease the light extraction efficiency. As a result, the significant energy loss when transferring between different molecular orbitals is noted and the correlated color temperature (CCT) is increased for those pc-LEDs.

Because semiconductor nanocrystals (also called QDs) have excellent properties such as controllable particle sizes, various emission wavelengths and high photoluminescence quantum yields (PL QYs),^20–25 they have emerged as the fourth generation of illumination technology.^26–30 QD blending with YAG phosphor has been reported to improve the CRI value of white LEDs.^10,31 In particular, CdSe and CdS QDs have been regarded as promising materials for white LEDs due to their size-dependent emission tunability and high PL QY.^24–33

For a white LED combined CdSe/ZnS QDs (555 and 613 nm) with InGaN/GaN LEDs, a CRI between 79.6 and 82.4 can be obtained by controlling the concentration QDs.²⁶ Moreover, the hybridization of InGaN/GaN n-UV LEDs with different types of CdSe/ZnS QDs (540, 580, and 620 nm) and conjugated polymer (439 nm) can generate white light. The device has high CRI above 80.²⁷ When CdSe/ZnS QDs dispersed in photosensitive epoxy resins excited by InGaN LED chips, and under a working current of 20 mA, the luminous efficacies and CRI valves of the single (560 nm), dual (540 + 620 nm), and multi (540 + 560 + 580 + 620 nm) hybrid devices were 8.1 lm W⁻¹, 5.1 lm W⁻¹, and 6.4 lm W⁻¹, and 21.46, 43.76 and 66.20, respectively.³² Therefore, white LED with high CRI can be achieved by mixing several kinds of QDs.

For the Sr₃SiO₅-based white LED, Jang et al. pointed out that the luminous efficacy of Sr₃SiO₅-based white LED is 30 lm W⁻¹, it drops to 14 lm W⁻¹ when incorporation of CdSe QDs, but a high CRI value of 90 is achieved.²⁵ On the other side, a bright three-band white light can be generated from the CdSe/ZnSe QD-assisted Sr₃SiO₅:Ce³⁺, Li⁺-based white LED. The CdSe/ZnSe QDs has high PL QY of 79% and the emission wavelength of 623 nm. Its luminous efficiency, CCT, and CRI are 26.8 lm W⁻¹, 6140 K, and 85, respectively.²⁴

In the YAG-modified white LED, Kim et al. pointed out that 120 nm sized yellow emitting nano-YAG:Ce and 4.5 nm-sized orange emitting CdS:Mn/ZnS QDs-based white LEDs exhibited the CRI values of 78 and 62, respectively. YAG:Ce-based white LEDs has poor CRI due to the lack of red spectral component. Blends of YAG:Ce nanophosphors and CdS:Mn/ZnS QDs displays the CRI value of 85.³⁰ In order to improve CRI, Shen et al. pointed out that a white LED combining a blue LED with the blends of nano-YAG phosphors and orange- and red-emission CdSe/CdS/ZnS QDs with a weight ratio of 1 : 1 : 1 showed excellent white light with luminous efficacy, color coordinates, CRI and CCT of 82.5  lm W⁻¹, (0.3264, 0.3255), 91 and 4580  K, respectively.³¹ When CdSe QDs (577 nm, PL QY 40%) is blended with commercial YAG:Ce phosphor under different rations, the CRI (68 to 71) and luminous efficacy (74 to 82 lm W⁻¹) could be improved simultaneous. It is interesting to find that there is a positive effect of QDs on the luminous efficiency and CRI value.²⁹

Wang et al. have prepared ZnCdTe/CdSe QDs by aqueous synthesis, and their emission spectrum ranges from 530 to 650 nm with a PL QY of 2.16%. ZnCdTe/CdSe QDs are blended with green phosphor Ca₈Mg(SiO₄)₄Cl₂ and pumped by blue chip to form white LED. A low CRI is obtained in ZnCdTe/CdSe QDs excited by blue chip. Incorporation of green phosphor with ZnCdTe/CdSe QDs, the CRI is improved up to 88.2 due to the full width at half maximum (FWHM) of ZnCdTe/CdSe and Ca₈Mg(SiO₄)₄Cl₂ is broad.³⁴ Therefore, we know that the CRI value can be increased dramatically by mixing phosphors, which have broad emission spectrum. However, the luminous efficacies of devices are not shown in the paper.³⁴ Moreover, orange- and red-emitting CIS/ZnS QDs with PL QY ranging from 41 to 81% and green phosphors Ba₂SiO₄:Eu²⁺ were blended with silicon resins. The luminous efficacy of this device was 32.7 lm W⁻¹ at 20 mA and with the high CRI of 90.³⁵

Ziegler et al. fabricated white LEDs with a Ra index of 86, which is composed of red InP/ZnS QDs with green Sr_0.94Al₂O₄:Eu_0.06 and yellow YAG:Ce phosphors.³⁶ Aboulaich et al. prepare Ce-doped YAG (YAG:Ce) and CuInS₂/ZnS QDs for application into white LED. YAG:Ce and CIS/ZnS QDs was mixed with silicone to form a film, and were applied on blue InGaN chip as a converter to achieve white LED. Only using YAG:Ce as a converter film shows low CRI and cold white light, and bilayered YAG:Ce and CIS/ZnS QDs film displays higher CRI of about 84 and warm white light with a CCT of 2784 K. However, the luminous efficacy decreases.³⁷ Moreover, the luminous efficacy and CRI of those white LEDs are still not high enough to meet the requirement of DOE.

Zhong et al. pointed out that the stability of alloyed ZnCdSe QDs is better than that of core/shell QDs,³⁸ and the synthesis process is also simple.³⁹ Moreover, we can fabricate red ZnCdSe QDs with PL QY higher than 40%.⁴⁰ Therefore, ZnCdSe alloyed QDs was used to blend with YAG. When applied in white LEDs, the PL QY of QDs has positive effect on the performance but the mechanism has not been fully understood yet. Therefore, in this study, the effect of ZnCdSe QD addition and PL QY of QD on CRI and luminous efficacy of commercial YAG-based white LED has been explored. Three kinds of white LEDs including YAG-based, YAG/CaSiAlN₃-based and YAG/ZnCdSe QDs-based white LEDs are fabricated and the devices properties are also evaluated. By controlling the purification process, ZnCdSe QD with different PL QY can be prepared. The effect of PL QY of ZnCdSe QDs on the performance of these YAG/QDs-based white LED is also studied. The blue LED pumping commercial YAG phosphor, while pumping YAG blend with CaSiAlN₃, and pumping YAG blend with ZnCdSe QD having low or high PL QY is named as YAG, nitride-, and L-QD- or H-QD-modified device, respectively.

Experimental

In this study, all of chemicals and materials were described in this section to prepare ZnCdSe. Cadmium oxide (CdO, 99.998%) was purchased from Alfa Aesar. Zinc oxide (ZnO, 99.99%), lauric acid (LA, 99%), selenium powder (Se, 99.999%), trioctylphosphine (TOP, 90%), and hexyldecylamine (HDA, 90%) were obtained from Aldrich. Hexane (99.7%) and methanol (99%) were provided by Mallinckrodt Chemicals. All chemicals were used as received without further purification.

Most details of the synthetic and characterizing methods were similar to the ones reported previously with some modify.^41–43 Ternary ZnCdSe QDs are prepared as follows. Each of 0.15 mmol of CdO and ZnO were mixed with 8.73 mmol of LA, which used as a complex reagent, in a three-necked flask and then heated it to 230 °C under Ar until a clear solution was formed to prepare the cadmium/zinc-LA precursors. After cadmium/zinc-LA precursor was formed, 13.5 mmol of HDA was added into three-necked flask reheated to 300 °C under Ar flow to form an optically clear solution. At this temperature, the Se solution containing 1.2 mmol of Se dissolved in 1 mL of TOP, was swiftly injected into the reaction flask. The total reaction time was 1 min under 290 °C. The mixed solution was swiftly cooled down to stop reaction. Samples were precipitated with hot anhydrous methanol (65 °C) for purification process. The precipitate was dissolved in hexane to remove unreacted reagents for further measurement.

The high and low PL QY of ZnCdSe QDs was prepared by controlled the volume of hot anhydrous methanol. As-prepared QDs were dissolved in hexane and the PL QY was 44%. 15 mL of the as-prepared QDs solution was evaporated under vacuum to form a powder. 5 mL hot methanol was added into centrifugal tube to remove surfactant. After purification, the PL QY of ZnCdSe was 29%. ZnCdSe QDs with different PL QYs were incorporated into YAG under different weight ratios to form white LED. On the other hand, the commercial red phosphor CaSiAlN₃ (CN-NR630) was purchased from China Glaze CO., LTO, and used for comparison.

Blends of commercial YAG:Ce phosphors with different weight ratios of ZnCdSe QDs or CaSiAlN₃ phosphors (1, 5, 10, 20, and 40 wt%) were prepared (called composite phosphor), and applied for the LED fabrication. White LEDs were fabricated using surface-mounted device (SMD) typed InGaN-based blue emitting LEDs (λ_em = 455 nm, 13 mil) with power efficiency of 19 mW. A mixture of composite phosphor and transparent silicone (1 : 23) is coated on a blue LED, and the subsequent curing is done at 100 °C for 45 min. Blue LED pumping commercial YAG:Ce phosphor is called YAG-based LED, while pumping YAG blend with CaSiAlN₃ is called nitride-modified, pumping YAG blend with low PL QY of ZnCdSe QD is called L-QD-modified, and pumping YAG blend with high PL QY of ZnCdSe QD is called H-QD-modified device, respectively.

The optical properties of samples, phosphors and QDs, were performed by using a fluorescent spectrometer (Hitachi, F-7000) and UV-visible spectroscopy (Jasco V-670). The particle morphology and the sizes distribution were observed by transmission electron microscopy (TEM, JEOL JEM-2010) and scanning electron microscopy (SEM). PL QY of QDs was determined by using fluorescent dye (rhodamine 101 in ethanol) as reference sample. The performance of devices was measured by integrating sphere with 15 cm diameter (Isuzu Optics) under different injection currents between 20 and 100 mA. Long period time test was conducted under 20 mA.

Results and discussion

The excitation and emission spectra of commercial YAG, CaSiAlN₃ phosphors and ZnCdSe QDs are shown in Fig. 1. The wavelengths of excitation peak of YAG are located at 340 and 460 nm, and the band emission located at 534 nm is due to Ce^{3+ 2}D–²F_5/2,7/2 transition.⁴⁴ On the other hand, the excitation wavelength of CaSiAlN₃ is in the range from 300 to 601 nm. The red emission wavelength at 602 nm is due to 4f⁶5d¹ to 4f⁷ transition of Eu²⁺ ions from 550 to 750 nm.⁴⁵ A large stokes shift and FWHM of both YAG and CaSiAlN₃ phosphor owing to the transition between the D and F orbital are also noted. Moreover, because the emission wavelength of CaSiAlN₃ phosphors is longer than that of YAG, the emission energy released from YAG can be reabsorbed by CaSiAlN₃ phosphors.⁴⁵ As a result, when compared with YAG, emission wavelength of CaSiAlN₃ is not so sensitive to the excitation wavelength.

Fig. 1 Emission (solid line) and excitation (dash line) spectra of (a) YAG, (b) CaSiAlN₃ phosphor, and (c) ZnCdSe QDs.

On the other hand, since the purification process does not influence the excitation spectrum significantly, the excitation wavelength of both H-QD and L-QD are noted from 300 to 618 nm. The red emission wavelength peak noted at 619 nm is due to the band to band emission of QD, revealing that the emission wavelength of QDs is non sensitive for the excitation wavelength. Moreover, the PL QY of as-prepared (H-QD) and purified ZnCdSe QDs (L-QD) is 44 and 29% with a FWHM of about 26 nm, respectively, suggesting that during the purification process, the degree of organic passivation changes, thus further affecting the PL QY and luminescence intensity of QD. Based on Fig. 1, it can be observed clearly that the excitation spectrum of ZnCdSe QD is broader, emission spectrum is narrower, and emission wavelength is longer than those of YAG, which suggests that the emission energy of YAG can be reabsorbed by ZnCdSe QDs, too.

Fig. 2 shows the SEM micrographs of YAG and CaSiAlN₃ phosphor, and TEM image of ZnCdSe QDs. The morphology of YAG and CaSiAlN₃ phosphor show an irregular shape and particle size of YAG and CaSiAlN₃ is about 10 and 7 μm, respectively. While a spherical shape with uniform size dispersion can be observed for ZnCdSe QD, and its average particle size is 4.5 ± 0.2 nm. In addition, these pc WLEDs have low device luminous efficacy because micron-sized phosphor induces a scattering effect to lose conversion efficiency from blue to yellow and red light.^25,46 By reducing the particle size of phosphors to nano-scale, the scattering might be ignored.^47,48 Because the ZnCdSe QD is in the nanometer scale, the scattering effect of YAG and CaSiAlN₃ phosphor may be more serious than that of ZnCdSe QD.

Fig. 2 Surface morphologies of (a) YAG, (b) CaSiAlN₃ phosphor, and (c) ZnCdSe QDs.

The XRD pattern of ZnCdSe QD is shown in Fig. 3. We can find that the ZnCdSe QDs seem to have a wurtzite CdSe phase. Previously, it is pointed out that partial substitution of Cd for Zn results in enhancing the PL QY of ZnCdSe QDs, which has CdSe wurtzite with an actual composition of Zn_0.03Cd_0.97Se.⁴⁰

Fig. 3 XRD pattern of ZnCdSe QD.

The EL spectra of YAG-based, nitride and QD-modified white LEDs under 20 mA forward current are shown in Fig. 4. The inset shows the color pictures of white LED with the addition of 10 wt%-red-emitting phosphor for all devices. It can be seen that the luminous intensity of nitride-modified sample is lower than that of YAG-based one shown in Fig. 4(a). The main peak shifts from 534 to 557 nm for YAG and moves from 602 to 616 nm for CaSiAlN₃. The emission wavelength of luminescent materials is changed under different circumstance.⁴⁹ The EL emission intensity coming from YAG decreases as increasing the concentration of CaSiAlN₃, meaning that the yellow light re-absorbs by the CaSiAlN₃. We also find that the EL intensity of nitride-modified phosphor is stronger than that of blue chip. This phenomenon may be due to that the light emitting from blue chip is absorbed or scattered by phosphors. The emission intensity of YAG decreases and is almost suppressed totally when the CaSiAlN₃ phosphor content is higher than 20 wt%. As we have mentioned above that although the range of excitation wavelength for CaSiAlN₃ phosphor is broader than that of YAG and emitting yellow light comings from YAG can be reabsorbed by CaSiAlN₃, increasing the concentration of CaSiAlN₃ will dilute the concentration of YAG and decrease its emission intensity. Moreover, the scattering cannot be avoided owing to those two phosphors are in micrometer-scale size, resulting in increasing energy lost and reducing the light extraction efficiency. Those may be the main reasons resulting in the decreases in the EL intensity of nitride-modified device.

Fig. 4 EL spectra of white LEDs. (a) Nitride-modified, (b) L-QD-modified, and (c) H-QD-modified device. Inset photo shows the device blends with 10 wt% of red phosphor.

On the other hand, the opposite phenomenon is noted in YAG/ZnCdSe QD-based white LED shown in Fig. 4(b) and (c). In L-QD device, the EL intensity coming from blue chip becomes stronger with increasing the QD concentration. This is because the concentration of YAG is diluted, the extraction intensity of blue light increases. In all QD contents, EL intensity of QD-modified phosphor is lower than that of YAG phosphor. EL peak of QD becomes more obvious until the content higher than 10 wt%. Besides, the EL spectrum area ratio of blend phosphor to blue chip (A_p–b) is decreased with increasing QD concentrations, resulting in the enhancement of the CCT for the device. Moreover, the emission wavelength of luminescent materials is changed under different circumstances.⁴⁹

In H-QD device, the R_p–b is higher than that of YAG-based white LED after addition of 5 and 10 wt% QD concentrations. As A_p–b in H-QD is higher than that of YAG, the CCT of H-QD devices are decreased. Moreover, when the concentration of QD is higher than 20 wt%, the CCT is enhanced for the devices, and EL peak of QD becomes more obvious. Based on Fig. 3 we can find that the H-QD modified materials are more effective than nitride- and L-QD modified materials when increasing the luminescence intensity.

The CIE chromaticity coordinates, CRI, CCT, and luminous efficacy for YAG- and composite phosphor-based white LEDs under a forward current of 20 mA are shown in Table 1. Based on Table 1, the luminous efficacies of nitride- and L-QD modified device decrease with increasing red-emitting phosphor contents, due to the particle size and PL QY effects. Most of the blue light emitting from the chip are absorbed by YAG and CaSiAlN₃, resulting in that the CCT of nitride-modified device becomes warm after adding red phosphor.⁴⁵ However, the CCT increases with increasing QD contents for both L-QD- and H-QD modified devices.²⁹ Improving CRI value for all devices is attributed to the longer emission spectrum of red-emitting phosphor as compared to YAG and the extension of the spectrum to red region.^{15,17,18,29,35,37,45} When the CaSiAlN₃ red phosphor content is higher than 40 wt%, the CRI and CCT of device cannot be measured because they are out of the white light region. It is worth mentioning that for QD-modified devices, they still in the white range even adding 40 wt% QDs, implying that their tunability of CCT is higher than that of nitride modified one.

Table 1 Device performances of white LEDs with injected current under 20 mA

Device	Nitride or QD content (wt%)	CIE (x, y)	CRI	CCT (K)	Efficacy (lm W^−1)
YAG-based	0	(0.37, 0.40)	68	4400	72
Nitride	1	(0.38, 0.40)	68	4100	66
	5	(0.38, 0.40)	75	4000	65
	10	(0.42, 0.39)	79	3200	56
	20	(0.45, 0.38)	77	2500	49
	40	(0.55, 0.36)	—	—	36
L-QD	1	(0.37, 0.40)	70	4400	69
	5	(0.36, 0.38)	73	4700	65
	10	(0.36, 0.37)	75	4700	62
	20	(0.33, 0.32)	80	5700	43
	40	(0.29, 0.24)	79	11 100	26
H-QD	1	(0.37, 0.40)	70	4400	73
	5	(0.36, 0.38)	72	4200	75
	10	(0.35, 0.36)	75	4300	68
	20	(0.34, 0.33)	77	4800	59
	40	(0.32, 0.30)	82	6500	43

---

Several reactions, including light absorption (such as by LED chip materials, encapsulates, reflective cup, and phosphor particles), scattering by phosphor particles, and reflection and refraction at interfaces, can happen when propagating light within an LED package. If the energy loss due to the scattering light can be reduced and reused, the efficacy of white LED can be enhanced ideally. Because the excitation wavelength of ZnCdSe QDs and CaSiAlN₃ phosphor is broad and overlaps with the emission wavelength of YAG, they both have potentials to reabsorb those scattering light from YAG phosphor and reabsorb, reflection and refraction light from interface. However, there is a trade-off property between CRI and luminous efficacy. Improving in CRI can be achieved by mixing several phosphors with various wavelengths but the luminous efficacy of devices is decreased accordingly. Previously, we have pointed out that about 11 and 4% improvement of the luminous efficacy and CRI for composite phosphor-based white LED are obtained, overcoming the trade-off property.²⁹ The improvement of device is because the QDs can reabsorb the scattering light which emits from the YAG phosphor, or reabsorb reflection and refraction light which comes from LED chip, then re-emitting light with 577 nm. In this study, in L-QD device, only CRI of device can be improved (about 16%). This may be due to the PL QY of ZnCdSe QDs is not high enough (only 29%) and the utilization of scattering light is not effective. However, increasing PL QY of ZnCdSe QD to 44% (H-QD) will promote the luminous efficacy and CRI about 4 and 6% simultaneously. Moreover, the improvement of CRI is up to 21% for adding 40 wt% of H-QDs.

For white LED application demands, CRI, luminous efficacy, CCT as well as CIE are important. Because the trade-off effect between CRI and luminous efficacy is commonly noted in commercial traditional phosphors in which improving CRI by adding several kinds of phosphors results in the decrease in the luminous efficacy of devices. We try to overcome this disadvantage by adding QD with high PL QY. Because the size of QDs is in the nanometer scale, the scattering can be ignored. Moreover, the excitation window of QDs is quite broad, implying that scattering blue and yellow light can be re-absorbed by QD and then re-emitted red light. In this study, the improvement of CRI of WLED is not obvious due to the small FWHM value, and the trade-off effect can be overcome by adding high PL QY of QD successfully.

Fig. 5 shows the relationship between CRI values and luminous efficacy under various phosphor concentrations. We can find that due to the particle size effect, mixing YAG and CaSiAlN₃ phosphors to fabricate white LED will have low CCT of device, while mixing YAG and QDs, CCT of device can be promoted. The QDs with higher PL QY can improve both luminescence efficiency and CRI under low QD concentration. Denault et al. have pointed out that the devices with Y₃Al₅O₁₂:Ce³⁺ and QDs (590 nm, PL QY > 30%) achieved a CCT of 4000 K, CRI of 81, and luminous efficacy of 57 lm W⁻¹, while devices with Lu₃Al₅O₁₂:Ce³⁺ and the same QDs (590 nm, PL QY > 30%) achieved a CCT of 5700 K, a CRI of 90, and luminous efficacy of 22 lm W⁻¹. The results obtained suggest that the use of QDs may allow for solid state white lighting devices with high color rendition. However, these devices suffered from a decrease in the luminous efficacy compared to devices using only phosphor powder.⁵⁰ This phenomenon is more common for white LED when mixing more than two kinds of phosphor powders. It seems that the CRI and efficacy of devices cannot be improved at the same time. The reasons may be due to the geometry of package and the PL QY of QDs. If reflect cup is removed, the scattering light cannot be reused by QDs, therefore, the luminous efficacy decreases. Chung et al. have pointed out that the high CRI values of the white LED are achieved by efficient energy transfer from YAG:Ce to CdSe/ZnSe (622 nm, PL QY of 58%) QDs in hybrid phosphor system. This device has shown an excellent CRI of 92.3 at 20 mA. However, the luminous efficacy of the hybrid phosphor white LED decreases from 92 to 77 lm W⁻¹.⁵¹ They think that the addition of 622 nm CdSe/ZnSe QDs results in a decrease of the luminous efficacy of the device, because the luminous efficacy critically depends on the intensity of the green to yellow region and aggregation of CdSe/ZnSe QDs in the silicone matrix. The PL QY of QDs is high enough, however they do not use reflect cup to fabricate white LED. Therefore, the enhancement of the CRI and efficacy at the same time cannot be observed in their experiment.

Fig. 5 Relationship between CRI and efficacy of white LEDs under various phosphor concentrations.

In order to understand the color perception of white LED, the related chromaticity diagram is displayed in Fig. 6. The insets show the color pictures of white LED after addition of 40 wt% red-emitting phosphor for all devices. Fig. 6(a) depicts the operating (x, y) coordinates of nitride-modified device. Its CIE coordinate shifts from (0.37, 0.40) to (0.55, 0.36). When the phosphor contents are lower than 20 wt%, white region of the CIE chromaticity diagram can be achieved, while CIE coordinates is out of the white region as the phosphor content higher is than 20 wt%. On the other hand, the operating (x, y) coordinates of QD-modified device can fall all in the white region of the CIE chromaticity diagram shown in Fig. 6(b) and (c). It is interesting to find that the CIE coordinate of QD-modified device can be shifted from (0.37, 0.40) to (0.29, 0.24) and (0.32, 0.30) by adding ZnCdSe QDs with different PL QY. This behavior is different from hybrid yellow and red phosphors devices.

Fig. 6 The CIE chromaticity coordinates of (a) nitride-modified, (b) L-QD-modified and (c) H-QD-modified device with different blend ratios. (1, 2, 3, 4, 5, and 6 is 0, 5, 10, 20, 40, and 100%, respectively.) Inset photo shows the device blends with 40 wt% of red phosphor.

Tran and Shi have concluded that the lumen and CCT of light output strongly depend on the combination of phosphor concentration and phosphor thickness.⁵² The package with lower phosphor concentration and higher phosphor thickness has lower trapping efficiency and less backscattering of light and thus has higher luminous efficacy and CCT. Because the concentration of YAG in L-QD and H-QD-modified device is lower than that of YAG-based white LED, higher CCT of L-QD and H-QD-modified device can be obtained. On the other hand, due to the particle size effect, phosphors in micrometer scale causes lower CCT of nitride-modified device. When compared with YAG-based white LED, the CRI of L-QD and H-QD is also increased, because the light scattering by phosphor particles, reflecting and refracting at interfaces have the possibility to reabsorb by QDs and then emit longer wavelength such as 625 with red light. Therefore, the efficiency of utilization of scattering light is improved, resulting in the promotion of both of the efficiency and CRI. Therefore, we can conclude that using red light materials with longer emission wavelength can improve CRI, but the efficacy of devices depends on the particle size and PL QY of conversion materials.

Table 2 summarizes the optical properties of H-QD white LED. As the applied current was increased, the luminous efficacy of white LEDs fell because of the droop effect.^53,54 The CIE color coordinates and CCT of white LED barely changed. The results demonstrate the high stability of the H-QD white LED. Table 3 summarizes the optical properties of white LEDs after long period test at applied currents 20 mA. In this test, 1 wt% red phosphor is added with QDs. Not only luminous efficacy of blue LED but all of white LEDs also decrease. From this table we can find that the decay rate of QD-modified is 4% per month, while that of YAG-based is 5% per month. CIE, CCT, and CRI for all devices tend to remain constant. This result implies that the long term stability of QD-modified white LEDs is better than that of YAG-based white LED.

Table 2 Optical properties of H-QD white LEDs at applied currents 20 to 100 mA

QD (wt%)	Current (mA)	CIE (x, y)	CRI	CCT (K)	Efficacy (lm W^−1)
0	20	(0.37, 0.40)	68	4400	72
	60	(0.36, 0.39)	68	4450	47
	100	(0.36, 0.39)	68	4490	35
1	20	(0.37, 0.40)	70	4400	73
	60	(0.36, 0.39)	70	4450	49
	100	(0.36, 0.39)	70	4500	36
5	20	(0.36, 0.38)	72	4200	75
	60	(0.36, 0.38)	72	4260	49
	100	(0.36, 0.38)	72	4310	36

---

Table 3 Optical properties of various white LEDs after long term test at applied currents 20 mA^a

Devices	CIE (x, y)		CCT (K)		CRI		Efficacy (lm W^−1)		Decay (% per month)
	A	B	A	B	A	B	A	B
a Applies current: 20 mA; A: as-prepared white LED, B: white LED after long period test (more than one month).
Blue chip	(0.15, 0.03)	(0.15, 0.03)	0	0	0	0	12	11	5
YAG	(0.37, 0.40)	(0.35, 0.38)	4400	4900	68	68	72	66	5
L-QD	(0.37, 0.40)	(0.37, 0.40)	4400	4500	70	70	69	65	4
H-QD	(0.37, 0.40)	(0.37, 0.40)	4300	4300	70	70	73	69	4

---

Conclusion

For an effort to explore the role of PL QY of QDs on efficacy and CRI of YAG phosphor-based white LEDs, ZnCdSe QDs with different PL QYs and CaSiAlN₃ phosphor are blended with YAG phosphors. Compared with the L-QD-modified device, better luminous efficacy values is obtained from H-QD, indicating that the QDs with high PL QY are beneficial for not only reducing the light scattering of YAG phosphor that is inherent in YAG-based phosphor but also increasing the utilization of reflection and refraction light. Both luminous efficacy and CRI value are improved for YAG/H-QDs-based white LEDs. It seems that the PL QY of QDs plays an important role in improving both CRI and luminous efficacy of QD modified-based white LEDs. Moreover, the stability of QD-modified white LEDs is better than that of YAG-based white LED.

Acknowledgements

This work was supported by the Ministry of Science and Technology of ROC under contracts 100-2221-E-150-063, 103-2221-E-150-055 and 104-2221-E-150-049.

Notes and references

1. J. D. Gosnell, S. J. Rosenthal and S. M. Weiss, IEEE Photonics Technol. Lett., 2010, 22, 541.
2. S. Nizamoglu, T. Ozel, E. Sari and H. V. Demir, Nanotechnology, 2007, 18, 065709.
3. H. Wu, X. Zhang, C. Guo, J. Xu, M. Wu and Q. Su, IEEE Photonics Technol. Lett., 2005, 17, 1160.
4. http://energy.gov/eere/ssl/solid-state-lighting.
5. J. K. Sheu, S. J. Chang, C. H. Kuo, Y. K. Su, L. W. Wu, Y. C. Lin, W. C. Lai, J. M. Tsai, G. C. Chi and R. K. Wu, IEEE Photonics Technol. Lett., 2003, 15, 18.
6. C. H. Kuo, J. K. Sheu, S. J. Chang, Y. K. Su, L. W. Wu, J. M. Tasi, C. H. Liu and R. K. Wu, Jpn. J. Appl. Phys., Part 1, 2003, 42, 2284.
7. H. S. Chen, S. J. J. Wang, C. J. Lo and J. Y. Chi, Appl. Phys. Lett., 2005, 86, 131905.
8. K. Sakuma, K. Omichi, N. Kimura, M. Ohashi and D. Tanaka, Opt. Lett., 2004, 29, 2001.
9. L. Zhou, J. Huang, L. Yi, M. Gong and J. Shi, J. Rare Earths, 2009, 27, 54.
10. C. Y. Shen, K. Li, Q. L. Hou, H. J. Feng and X. Y. Dong, IEEE Photonics Technol. Lett., 2010, 22, 884.
11. H. S. Jang and D. Y. Jeon, Opt. Lett., 2007, 32, 3444.
12. H. S. Jang, Y. H. Won and D. Y. Jeon, Appl. Phys. B, 2009, 95, 715.
13. P. F. Smet, A. B. Parmentier and D. Poelman, J. Electrochem. Soc., 2011, 158, R37.
14. J. K. Park, M. A. Lim, C. H. Kim, H. D. Park, J. T. Park and S. Y. Choi, Appl. Phys. Lett., 2003, 82, 683.
15. J. K. Park, C. H. Kim, S. H. Park, H. D. Park and S. Y. Choi, Appl. Phys. Lett., 2004, 84, 1647.
16. M. R. Krames, O. B. Shchekin, R. Mueller-Mach, G. O. Mueller and L. Zhou, J. Disp. Technol., 2007, 3, 160.
17. H. S. Jang and D. Y. Jeon, Appl. Phys. Lett., 2007, 90, 041906.
18. R. J. Xie, N. Hirosaki, K. Sakuma, Y. Yamamoto and M. Mitomo, Appl. Phys. Lett., 2004, 84, 5404.
19. R. J. Xie, N. Hirosaki, M. Mitomo, K. Takahashi and K. Sakuma, Appl. Phys. Lett., 2006, 88, 101104.
20. H. S. Chen, C. K. Hsu and H. Y. Hong, IEEE Photonics Technol. Lett., 2006, 18, 193.
21. M. Molaei, M. Marandi, E. Saievar-Iranizad, N. Taghavinia, B. Liu, H. D. Sun and X. W. Sun, J. Lumin., 2012, 132, 467.
22. C. Y. Huang, Y. K. Su, Y. C. Chen, P. C. Tsai, C. T. Wan and W. L. Li, IEEE Electron Device Lett., 2008, 29, 711.
23. A. W. Tang, F. Teng, Y. H. Gao, D. Li, S. L. Zhao, C. J. Liang and Y. S. Wang, J. Lumin., 2007, 122–123, 649.
24. H. S. Jang, B. H. Kwon, H. Yang and D. Y. Jeon, Appl. Phys. Lett., 2009, 95, 161901.
25. H. S. Jang, H. Yang, S. W. Kim, J. Y. Han, S. G. Lee and D. Y. Jeon, Adv. Mater., 2008, 20, 2696.
26. S. Nizamoglu, G. Zengin and H. V. Demir, Appl. Phys. Lett., 2008, 92, 031102.
27. H. V. Demir, S. Nizamoglu, T. Ozel, E. Mutlugun, I. O. Huyal, E. Sari, E. Holder and N. Tian, New J. Phys., 2007, 9, 362.
28. J. Ziegler, S. Xu, E. Kucur, F. Meister, M. Batentschuk, F. Gindele and T. Nann, Adv. Mater., 2008, 20, 4068.
29. S. R. Chung, K. W. Wang and M. W. Wang, J. Nanosci. Nanotechnol., 2013, 13, 4358.
30. J. U. Kim, Y. S. Kim and H. Yang, Mater. Lett., 2009, 63, 614.
31. C. Shen, J. Chu, F. Qian, X. Zou, C. Zhong, K. Li and S. Jin, J. Mod. Opt., 2012, 59, 1199.
32. H. Wang, K. S. Lee, J. H. Ryu, C. H. Hong and Y. H. Cho, Nanotechnology, 2008, 19, 145202.
33. H. Yang and P. H. Holloway, Adv. Funct. Mater., 2004, 14, 152.
34. R. Wang, J. Zhang, X. Xu, Y. Wang, L. Zhou and B. Li, Mater. Lett., 2012, 84, 24.
35. P. H. Chuang, C. C. Lin and R. S. Liu, ACS Appl. Mater. Interfaces, 2014, 6, 15379.
36. J. Ziegler, S. Xu, E. Kucur, F. Meister, M. Batentschuk, F. Gindele and T. Nann, Adv. Mater., 2008, 20, 4068.
37. A. Aboulaich, M. Michalska, R. Schneider, A. Potdevin, J. Deschamps, R. Deloncle, G. Chadeyron and R. Mahiou, ACS Appl. Mater. Interfaces, 2014, 6, 252.
38. X. Zhong, M. Han, Z. Dong, T. J. White and W. Knoll, J. Am. Chem. Soc., 2003, 125, 8590.
39. S. R. Chung, K. W. Wang and H. S. Chen, J. Nanomater., 2013, 2013, 526862 1.
40. S. R. Chung, K. W. Wang, H. S. Chen and H. H. Chen, J. Nanopart. Res., 2015, 17:101, 1.
41. Z. A. Peng and X. Peng, J. Am. Chem. Soc., 2001, 123, 1389.
42. L. Qu, Z. A. Peng and X. Peng, Nano Lett., 2001, 1, 333.
43. L. Qu and X. Peng, J. Am. Chem. Soc., 2002, 124, 2049.
44. N. Wei, T. Lu, F. Li, W. Zhang, B. Ma, Z. Lu and J. Qi, Appl. Phys. Lett., 2012, 101, 061902.
45. C. C. Lin, Y. S. Zheng, H. Y. Chen, C. H. Ruan, G. W. Xiao and R. S. Liu, J. Electrochem. Soc., 2010, 157, H900.
46. M. Nyman, L. E. Shea-Rohwer, J. E. Martin and P. Provencio, Chem. Mater., 2009, 21, 1536.
47. M. Zachau, D. Becker, D. Berben, T. Fiedler, F. Jermann and F. Zwaschka, Proc. SPIE, 2008, 6910, 691010.
48. N. T. Tran, J. P. You and F. G. Shi, J. Lightwave Technol., 2009, 27, 5145.
49. H. Song and S. Lee, Nanotechnology, 2007, 18, 255202.
50. K. A. Denault, A. A. Mikhailovsky, S. Brinkley, S. P. DenBaars and R. Seshadri, J. Mater. Chem. C, 2013, 1, 1461.
51. W. Chung, H. J. Yu, S. H. Park, B. H. Chun and S. H. Kim, Mater. Chem. Phys., 2011, 126, 162.
52. N. T. Tran and F. G. Shi, J. Lightwave Technol., 2008, 27, 3556.
53. Y. L. Li, Y. R. Huang and Y. H. Lai, Appl. Phys. Lett., 2007, 91, 181113.
54. M. F. Schubert, S. Chhajed, J. K. Kim, E. F. Schubert, D. D. Koleske, M. H. Crawford, S. R. Lee, A. J. Fischer, G. Thaler and M. A. Banas, Appl. Phys. Lett., 2007, 91, 231114.

---