1. Introduction

Because of the low cost, high color rendering, high reliability, and environmental friendliness, phosphor-converted light-emitting diodes (pc-LEDs) have been used as a promising light source for white LEDs (WLEDs), replacing conventional solid-state lighting [1–8]. White light from pc-LEDs is realized by a mixture of the short-wavelength light emitted from blue LED chips and the down-converted long-wavelength light from phosphors [6,9,10]. For the fabrication of pc-WLEDs, phosphor powers are blended with silicone and solvent, forming a phosphor slurry and then the phosphor slurry is deposited on to LED chips by various methods such as dispensing, conformal, and remote coating [6,11–15], resulting in different phosphor film structures. For conformal structure, the phosphor film is directly located on to the LED chip and so good angular color uniformity (ACU) can be achieved. However, amount of the light experiences scattering by phosphors, and reflection and reabsorption by the chip, resulting in degradation of the device performance [11,14]. These drawbacks were resolved by locating the phosphor film away from the LED chips, namely, employing the remote phosphor structure [16–18]. These phosphor films have been coated by means of various methods such as the slurry method (dispensing coating), the settling method, electrophoretic deposition (EPD), pulsed spray, etc [19–22]. For dispensing coating, because of the flowing nature of the phosphor-silicone slurry and their different densities [23], the phosphor particles are settled in the silicone matrix, leading to change of phosphor distribution. This causes the occurrence of inhomogeneous correlated color temperature (CCT) and ACU [23–28]. For instance, Yu et al. [10,13], experimentally and theoretically investigating the phosphor sedimentation effect on the optical performance of pc-LEDs with dispensed phosphor film, reported that CCT was sensitively affected by the position of the phosphor sediment, namely, either above or outside the LED chips. However, despite these disadvantages, the dispensing coating method is still popular in the industry because of the simple process and low manufacturing cost [23–30]. Thus, to improve the ACU of pc-LEDs with dispensed phosphors, a simple route, namely, use of different slanting angles of the reflector cup-sidewall, was adopted [29,30]. The simulation and experimental results showed that uniform phosphor distribution could be achieved by using the slanting angle that is the same as the contact angle between the phosphor slurry and the reflector cup sidewall. Furthermore, calculation showed that the phosphor layer with inverse-gradient concentrations would enhance the LEE [24]. This was attributed to reduction in the light loss by the absorption of LED chips or substrates. In addition, phosphor sedimentation was found to cause ~20% reduced luminous efficacy in lateral WLEDs compared with vertical WLEDs [23]. These results imply that the location, distribution and filling of phosphors are vital in achieving high luminous efficacy of packaged pc-WLEDs. Thus, in this study, dynamic phosphor sedimentation (forcible settling of phosphor particles) was employed to improve the ACU and reliability of pc-WLEDs for mobile display products (e.g., smartphone) and their results were compared with reference sample with a dispersion-coated phosphor film. Note that ACU is a crucial parameter for the pc-WLEDs for smartphone. The hydrothermal reliability of WLEDs was also examined because it directly affects the reliability of mobile products.

2. Experimental procedure

A 0.2-W LEDs (chip size: 0.8 mm × 0.22 mm and thickness = 110 μm) with a peak wavelength of 446 nm were used for the fabrication of white LEDs. To investigate the effect of dynamic phosphor sedimentation on the luminous flux and reliability of WLEDs, two samples were prepared. Phosphors (a mixture of two monochromatic phosphors, green (78%) and red (22%)) were mixed with silicone solvent to form phosphor slurry. Phosphor particles were manufactured by Denka (nitride green) with averaging diameter of 16 μm and Force4 (K₂SiF₆:Mn⁴⁺: KSF) with averaging diameter of 26 μm, and silicone (phenyl silicone) was produced by Dow Corning. The phosphor slurry was coated onto the LED module using thedispensing coating method. For one sample, the phosphor slurry was dispersed over the LED chip and kept uncured. Then the phosphor was naturally settled by gravity onto the LED chip (referred to here as phosphor dispersion sample). For the other, the phosphor slurry was first dispersed over the LED chip and then the phosphor was forcefully settled by combination of gravity and centrifugal force (referred to here as dynamic phosphor sedimentation sample). The forceful settlement was performed by two step processes. In other words, the samples were first rotated at 1,500 rpm for 150 s, which was followed by 2,000 rpm for 400 s. Figure 1 displays cross-section optical microscopy images obtained from the samples with phosphor films formed by different deposition methods, showing the geometry of the phosphor film, LED chip, and metal reflector. For the package with the phosphor dispersion method, the metal cup is filled up only with the phosphor film, whereas for the package with the phosphor sedimentation, the reflector cup is filled with the phosphor and silicone double films. For both the samples, a quantity of 0.2 mg was dispensed. The total phosphor fraction was 71% and 49% for the phosphor dispersion and sedimentation samples, respectively. The luminous flux of LEDs was measured at 350 mA using a 20-inch integrating sphere detector (ISP500-100). The angular-dependent CCT of WLEDs was measured at driving current of 350 mA using a CAS 140CT-152 Goniophotometer (Instrument system) for view angle varying from −80° to 80°. The hydrothermal dependence of the luminous flux of the samples was characterized at a drive current of 20 mA in a chamber with a humidity of 85% and a temperature of 85 °C. To characterize the thermal characteristics of the packaged samples, their thermal images were taken with a thermal imaging camera (FLIR: A8303sc).

 

Fig. 1 Cross-section optical microscopy images obtained from the samples with phosphor films formed by (a) dispersion (b) and sedimentation methods, showing the geometry of the phosphor film, LED chip, and metal reflector.

Download Full Size | PDF

3. Results and discussion

Figure 2 shows the photoluminescence (PL) emission spectra of packaged white LEDs fabricated with different phosphor deposition methods, which consisted of 446 nm blue LED chip, nitride green, and K₂SiF₆:Mn⁴⁺ yellow phosphors. It is evident that in addition to a luminescence peak at 446 nm from a blue LED chip, there are four peaks from the green and yellow phosphors. It is noted that the nitride green phosphor produced a broad peak at 539 nm while the KSF phosphor yielded three peaks at 613, 631 and 641 nm.

 

Fig. 2 PL emission spectra of packaged white LEDs fabricated with different phosphor deposition methods, which consisted of 446 nm blue LED chip, nitride green, and K₂SiF₆:Mn⁴⁺ yellow phosphors.

Download Full Size | PDF

Figure 3 exhibits the angular distribution of the CCT of the phosphor dispersion and phosphor sedimentation samples in the range of angles from −80° to 80°. At a normal incidence, the phosphor sedimentation sample reveals lower CCT (19,626 K) than thedispersion sample (20,572 K). This behavior can be explained by the different thicknesses of the phosphor layers and distribution of phosphor particles, as will be described below. It is noteworthy that for both the samples, the CCT decreases with increasing viewing angle. This can be attributed to an increase in the portion of down-converted yellow light that was generated as a result of different phosphor layer thicknesses. In other words, as the viewing angle increases, the blue ray travels a longer path (namely, more absorption by the phosphors). The longer path results in the generation of more down-converted yellow light, which hence leads to relatively lower CCT [27–30]. The CCT deviation of the dynamic phosphor sedimentation and dispersion samples is about 7,197 and 13,051 K, respectively. It is shown that although the dynamic phosphor sedimentation sample illustrates much better ACU than the dispersion sample, the CCT deviation is rather high. Kuo et al. [27] adopted a pulse spray coating method to define patterned remote phosphor structure and found that the patterned remote phosphor structure largely improved ACU and gave high chromatic stability in wider operating current range. Unlike this study, the patterned phosphor structure was shown to give remarkably small CCT deviation (266 K). This different deviation may be due to different CCTs, namely, very high CCTs for mobile products and relatively low CCT for solid-state lighting. The CCT deviation may be better controlled in the lower CCT samples. Despite the difference, however, use of such patterned structure may be helpful in reducing CCT deviation.

 

Fig. 3 The angular distribution of the CCT of the phosphor dispersion and phosphor sedimentation samples in the range of angles from −80° to 80°.

Download Full Size | PDF

The current dependence of the luminous flux and the luminous efficacy of the packaged samples (3804 PPA PKG) with different phosphor deposition methods are presented in Fig. 4. The phosphor sedimentation sample shows approximately 2.6% higher luminous flux andluminous efficacy at driving current of 20 mA as compared to the conventional phosphor dispersion sample. This improved lumen behavior can be explained as follows. First, when blue light propagates through the thicker phosphor film (i.e., phosphor dispersion sample), the resultant down-converted yellow light could be re-absorbed by the phosphors [26–31]. This can result in a reduction in the total lumen and luminous efficacy. Second, for the phosphor dispersion sample, the thicker phosphor layer contains well-distributed phosphor particles, by which light experiences back scattering, reducing light extraction efficiency [31]. In addition, unlike the phosphor dispersion sample, the metal reflector cup of the phosphor sedimentation sample is not fully covered by phosphors, as marked by the arrows in Fig. 1(b) and so this can increase light extraction. Furthermore, as will be described later, the phosphor dispersion sample suffers from somewhat inefficient heat dissipation, which could cause the LED temperature to increase and lower its performance.

 

Fig. 4 The current dependence of (a) the luminous flux and (b) luminous efficacy of the packaged samples (3804 PPA PKG) with different phosphor deposition methods.

Download Full Size | PDF

Figure 5 illustrates the lumen maintenance of the packaged samples with different phosphor deposition methods as a function of aging time. To investigate the hydrothermal reliability, the packaged samples were loaded into a chamber where the humidity and the temperature were set to be 85% and 85 °C, respectively. For both the samples, the luminous flux was assessed at 20 mA. It is shown that the luminous flux slowly decreases with increasing aging time. However, the phosphor sedimentation sample exhibits better hydrothermal aging behavior than the phosphor dispersion sample. For example, after aging for 1000 h, the luminous flux of the phosphor sedimentation sample was degraded by 15.3% while the flux of the phosphor dispersion sample was dropped by 28.7%. The phosphor sedimentation method resulted in 13.4% higher lumen maintenance rate than the dispersion method, indicating much better hydrothermal reliability.

 

Fig. 5 The lumen maintenance of the packaged samples with different phosphor deposition methods as a function of aging time. These samples were examined under the hydrothermal condition, namely, a humidity of 85% and a temperature of 85 °C.

Download Full Size | PDF

Figure 6 shows the thermal images of the packages (3804PKG) with different phosphor deposition samples, which were taken at 300 mA using a thermal imaging camera. The typical temperatures were estimated to be 99.0 °C and 103.7 °C for the phosphor sedimentation and dispersion samples, respectively. It is noteworthy that at the same current of 80 mA, the sedimentation sample reveals about 4.7 °C lower package temperature than the dispersion sample. When blue rays travel through the phosphor films, they unavoidably experience Stoke shift, generating thermal heat. For the sedimentation package, the heat could be effectively released through the connected phosphors to metal frame because the phosphors were closely packed by centrifugal force, which is thus somewhat similar to conformal coating method [26,27,30]. On the other hand, for the dispersion sample, the phosphors are well distributed (or well separated from one another) within the silicone encapsulant with poor heatconductivity. Consequently, the heat could not be efficiently dissipated and so the package temperature could increase, leading to somewhat lower luminous efficacy of pc-WLEDs (Fig. 4).

 

Fig. 6 (a) Optical image of a packaged pc-WLED. The thermal images of the packages (3804PKG) with (b) phosphor dispersion and (c) phosphor sedimentation samples, which were taken at 300 mA using a thermal imaging camera.

Download Full Size | PDF

4. Conclusion

We investigated how the phosphor deposition methods affected the CCT, luminance efficacy, and hydrothermal characteristics of packaged samples. For both the dispersion and sedimentation samples, the CCT decreased with increasing viewing angle. However, the phosphor sedimentation package exhibited much better ACU than the dispersion package. The dynamic phosphor sedimentation sample exhibited higher luminous efficacy at 20 mA than the phosphor dispersion sample. The phosphor sedimentation sample exhibited better hydrothermal aging reliability than the phosphor dispersion sample. The measurements showed that the phosphor sedimentation package more efficiently dissipated heat than the dispersion package. These results indicate that the phosphor sedimentation method is a promising deposition technique for the fabrication of high-performance pc-white LEDs for mobile display products.