1.
Introduction

Incandescent and fluorescent lighting were key advances; light-emitting diode (LED) lighting is a similarly crucial advance. As a new type of highly-efficient, cold-light, solid-state lighting, it has the advantages of high luminous intensity, low energy consumption, stable performance, and a long lifespan^[1–3]. In recent years, semiconductor light sources have been suggested for a wide range of applications because of their continuing reductions in price and their improvement of the luminous brightness of arena lighting.



Notably, there are high demand white LEDs, which are compound lights. According to the principle of luminescence, there are three ways to generate white LED light^[4]: combine an ultraviolet LED chip with phosphors of the three primary colors (red, green, blue)^[5–7], assemble an LED chip with the three primary colors or combine a blue LED chip and yellow phosphor. The most common way to produce white light in the LED industry is to mix a red phosphor and a green phosphor, which have corresponding bands, with an equal amount of epoxy or silicone potting around a blue LED chip. The chip is excited by blue light and the phosphors are excited by the formation of the yellow light mixture that forms a white light. However, the phosphor activity can easily subside during the solidification of silica gel^[8], which disrupts the uniform distribution of sediment-produced phosphor and substantially impacts the concentration and coherence^[10] of white LED light^[9].



For a high power white LED, BIN is a division of the Planck black line surrounding the color region, which is in a small area within the division. It can be similar to that for the resolution of the human eye, and there is only one color, this area is called BIN. If the color coordinates of a batch of LED are all in the same BIN, and their color tolerances are relatively concentrated, the consistency of these products is very high．



In the commercial production of small and medium power-type white LEDs, the phosphor content and distribution of the colloidal particles are more likely to affect the color coordinate distribution and control of the Color Tolerance Adjustment because of the small amount of phosphor necessary to package a single LED device. The color tolerance is the difference between the X and Y values calculated from the software of the photoelectric detection system. The smaller the number, the higher the accuracy, and the unit of the tolerance is SDCM. The resulting LED products have poor color concentration and are frequently defective, which both reduces the production efficiency, and increases production costs. For this study, an actual packaging production operation was established to explore the relationships between the phosphor deposition rate and the device’s coordinates and optical properties. The purpose was to find a reasonable packaging technology that could improve the color concentration levels of white LED devices and provide guidance in the actual production through the appropriate control of the deposition rate of phosphor.

2.
Experiment

An LED light source was used to study settlement of a white ceramic substrate. Specifically, a light with a highly concentrated wavelength and brightness was selected to mitigate the influences that the wavelength and luminance of the chip exercised on the color coordinates of the white LED.



The size of the chip was 36 × 80 mil², and the main wavelength range was 450–452.5 nm. In addition, the range of luminance was 480–500 mW and the single voltage range of the chip is 122–124 V. The glue was made of two groups of silicone (one A/B polymer and two methyl silicone components) and the mixing viscosity at room temperature was 5000 cp. The density of silicone after mixing was 1.03 g/cm³. The median particle size of yellow phosphor was 13 μm, and the density was 6.2 g/cm³. A white LED excited only by yellow phosphor produced a low score on the color rendering index (< 80), indicating that it was poor in color; this was due to the lack of a red light component. To remedy this defect, a small amount of red phosphor was added to increase the displayed color index of the white LED, which notably improved the color temperature of the light in the experiment. The median particle size of the red phosphor was 10μm, and the density was 3.1 g/cm³.



Fig. 1 outlines the whole process of the LED package dispensing device, illustrating the solid crystal, welding wire, and round glue components. The ration of phosphors for the white LED at 3000 K was 5351(A) : 5351(B) : GAL-535M : YH-C625EB = 1.8 : 1.8 : 1.3892 : 0.125. To study the sedimentation rate of the phosphors, seven sets of controlled experiments were set up to determine phosphor precipitation at time intervals of 0, 2, 5, 10, 20, 30, and 40 min. The concentration and optical properties of the device were compared by controlling the precipitation times of the phosphors.






onerror="this.onerror=null;this.src='http://www.jos.ac.cn/fileBDTXB/journal/article/jos/2018/12/PIC/18030039-1.jpg'"
class="figure_img" id="Figure1"/>



Download



Larger image


PowerPoint slide






Figure1.
(Color online) Schematic diagram of the LED settlement process.




The phosphorous glue was used within 30 min of mixing and deaerated into a dispensing tube in the packaging technology of the dispensing process. After the phosphorous had been dispensed onto the chips, the semi-finished products were placed in a single 1/4 box and baked in an oven until the phosphorous glue completely covered the chips for approximately 6–10 min. Therefore, it is assumed that the sedimentation of the phosphor mainly occurred during this time, first in the dispensing tube pushed down in the dispensing machine during automatic dispensing^[11], and then on the chip (sedimented in the placement and curing process of precipitation). Because only a single cup of glue was used, the controlled experiment finished in 30 min, and the phosphorous glue in the dispensing machine was at room temperature, therefore the colloid viscosity was remarkably high. Thus, the effect of precipitation in the process of dispensation before completion was almost negligible. In addition, because of the different placements after the completion of each dispensing period, the precipitated phosphor amounts were also different, this affected the concentration of the color coordinates.



The phosphors used in the experiments crystalline with approximately spherical particles (Fig. 2)^[12]. The deposition process of phosphor can be regarded as a fluid movement, guided by self-gravitational force F_g. The buoyancy force of F₀ and the resistance of fare shown in Fig. 3^[13].






onerror="this.onerror=null;this.src='http://www.jos.ac.cn/fileBDTXB/journal/article/jos/2018/12/PIC/18030039-2.jpg'"
class="figure_img" id="Figure2"/>



Download



Larger image


PowerPoint slide






Figure2.
Particle morphology and structure of a phosphor.






onerror="this.onerror=null;this.src='http://www.jos.ac.cn/fileBDTXB/journal/article/jos/2018/12/PIC/18030039-3.jpg'"
class="figure_img" id="Figure3"/>



Download



Larger image


PowerPoint slide






Figure3.
(Color online) Schematic of the force of phosphor particles settling in a stationary fluid.




Although phosphor particles are small, the initial and final settling velocities are not significantly different, hence acceleration can be ignored. When the particles are stable or settling down, f can be expressed as:

$f = {F_{ m{g}}} - {F_0} = frac{1}{6}pi {d^3}left( {{ ho _{{ m{mix}}}} - { ho _0}} ight)g,$

(1)

where d is the mean median particle diameter of phosphor particles, ρ_mix is the density of the phosphor particles, and ρ₀ is the density of silica gel. Specifically,

${ ho _{{ m{mix}}}} = frac{{{m_{{ m{mix}}}}}}{{{V_{{ m{mix}}}}}}.$

(2)

According to Stokes’ law, the resistance fof the phosphor particles during the settling time can be expressed as:

$f = 3pi deta nu ,$

(3)

where η is the mixing viscosity coefficient of silica gel and ν is the settling velocity of the phosphor solid particles.



Based on Eqs. (1) and (3), the particle sedimentation rate can be expressed as:

$silon = frac{{{d^2}left( { ho - { ho _0}} ight)}}{{18eta }}g.$

(4)

3.
Analysis and discussion

The so-called CIE 1931 chromaticity system is the CIE-xyz chromaticity system. The x axis color coordinates equal to the ratio of red primary colors and the y axis color coordinates equal to the proportion of green primaries. So, the z axis color coordinates are 1 ? (x + y).



According to Eq. (4), the settling velocity of the phosphor mass is proportional to its particle size and density, which are inversely proportional to the viscosity of the silica gel in a Newtonian fluid. As indicated through Eq. (2), the density of the mixed phosphor was 5.73 g/cm³, whereas the settling velocity of the particle was 1.69 × 10^?8 m/s, in short, the deposition of phosphor was very slow at room temperature. Notably, however, the particle density is much greater than the fluid density, the particles tend to sink in the Newtonian fluid. Conversely, in a plastic non-Newtonian fluid, the fluid will flow only when the shear force is large enough, the yield value is the minimum shear required for flow. Therefore, it is possible that some particles can be stably suspended in a non-Newtonian fluid with an appropriate yield value^{[14, 15]}. To confirm whether the phosphor was settling at room temperature, the settlement time intervals after dispensation were set at 0, 2, 5, 10, 20, 30, and 40 min. At these times, the coordinate concentration and other photoelectric parameters were tested by a “ZPLED 200” type LED testing machine.



As portrayed in Fig. 4, the color coordinates shift in different degrees and are accompanied by varied sedimentation times following dispensation. Moreover, the coordinates are discrete, which may be caused by the sedimentation of phosphors during the static process. In addition, the standard performance requirements of an LED module for general lighting are specified at a steady-state of less than 7 SDCM. In Fig. 5, the color coordinates are displayed in a range of less than 7 SDCM for settling times ranging from 0 to 20 min. However, the degree of the offset tolerance range differs between 30 and 40 min. Therefore, it can be concluded that the settling time of phosphor occurs between 0 and 20 min, during which the range of colour coordinates are concentrated at (x = 0.4432 ± 0.004, y = 0.4052 ± 0.002). Subsequently, between 30 and 40 min, the color coordinates are more discretely distributed, and the central coordinate is (x = 0.4366 ± 0.003, y = 0.4012 ± 0.003). Thus, energy star standards of color tolerance were not achieved for those products.






onerror="this.onerror=null;this.src='http://www.jos.ac.cn/fileBDTXB/journal/article/jos/2018/12/PIC/18030039-4.jpg'"
class="figure_img" id="Figure4"/>



Download



Larger image


PowerPoint slide






Figure4.
(Color online) Point coordinates diagrams for (a) 0, (b) 2, (c) 5, (d) 10, (e) 20, (f) 30, and (g) 40 min; (h, i) overall placement diagram.






onerror="this.onerror=null;this.src='http://www.jos.ac.cn/fileBDTXB/journal/article/jos/2018/12/PIC/18030039-5.jpg'"
class="figure_img" id="Figure5"/>



Download



Larger image


PowerPoint slide






Figure5.
(Color online) Distribution diagram of SDCM < 7 at different times after static setting.




As revealed in Fig. 6 and Fig. 7, the distribution of phosphor particles in the colloid changes during deposition. In particular, the aggregation of phosphor particles occurs, which leads to a change in the absorption of blue light and the scattering of blue and yellow light, thus, the color temperature also changes. Moreover, because the phosphor particles are deposited on the bottom layer, the concentration of the lower layer rises. The settling velocity is calculated as 1.69 × 10^?8 m/s, and therefore the corresponding decreased distances of the phosphor are 0, 2.028, 5.07, 10.14, 20.28, 30.42, and 40.56 μm. The actual thickness of the LED device is 1.24 mm, as is shown in Fig. 8(a). By contrast, the thickness of the white ceramic substrate is 0.68 mm, so the actual thickness of the colloid is 0.56 mm. When the settling time is 0 min, the phosphor particles are mainly concentrated on the upper surface; thus, more blue light is scattered back to and absorbed by the chip. This means that the blue emission is reduced, and the color temperature is lower. With the subsequent increase in settling time, the settling distance is also increased, and the phosphor particles are mainly concentrated on the lower surface; thus, the leakage of blue light is increased, the whole spectrum ratio is increased, and the color temperature is increased.






onerror="this.onerror=null;this.src='http://www.jos.ac.cn/fileBDTXB/journal/article/jos/2018/12/PIC/18030039-6.jpg'"
class="figure_img" id="Figure6"/>



Download



Larger image


PowerPoint slide






Figure6.
Color temperature trend over time.






onerror="this.onerror=null;this.src='http://www.jos.ac.cn/fileBDTXB/journal/article/jos/2018/12/PIC/18030039-7.jpg'"
class="figure_img" id="Figure7"/>



Download



Larger image


PowerPoint slide






Figure7.
Luminous flux trend over time.




Notably, the luminous flux is equal to a certain band in time and the radiation energy band relates to the vision rate of product. The relative rates of various wavelengths of light are perceived differently by the human eye, which indicates that the radiation power of the different wavelengths is equal, but the light flux is not equal. As depicted in Fig. 8(b), the spectra of different phosphors excited by various peak wavelengths are distinct. The light efficiency of the 555 nm wavelength (the green light flux) is the highest. For all of the experimental white light LEDs in the present study, when the temperature is lower, there is more red light transformation and the luminous flux is lower; by contrast, when the temperature is higher, there is less red light transformation and the luminous flux is higher.






onerror="this.onerror=null;this.src='http://www.jos.ac.cn/fileBDTXB/journal/article/jos/2018/12/PIC/18030039-8.jpg'"
class="figure_img" id="Figure8"/>



Download



Larger image


PowerPoint slide






Figure8.
(Color online) (a) Thickness of the LED device. (b) Spectrum of white LED.




Finally, the color temperature drift caused by phosphor precipitation cannot be ignored. Some researchers have offered a solution to the precipitation process, which consists of the reallocation of the phosphor proportion, and the addition of centrifugal precipitation to artificially control phosphor precipitation. In this scenario, the phosphor glue is evenly distributed on the surface of the chip to enhance the concentration of the color coordinates. However, this method adds a centrifugal sedimentation process after the conventional process; in addition, it considerably increases the time requirement and the cost of equipment, reduces the efficiency of production, and results in the non-uniformity of the whole space. In short, this solution is not feasible for the LED industry.

4.
Conclusion

This research explored the phosphor dispensation process of the LED light sources, it was determined that the deposition of phosphor has a great influence on the concentration distribution of color coordinates in white LEDs. The phenomenon of precipitation was experimentally measured at time intervals of 0, 2, 5, 10, 20, 30, and 40 min. The color coordinate concentrations and optical properties were also tested. The results confirmed that phosphor sedimentation occurs between 0 and 20 min, during which the color coordinate placement is more concentrated; the central coordinates are (x = 0.4432 ± 0.004, y = 0.4052 ± 0.002), and the products have SDCM scores lower than 7. Moreover, the optical properties are more stable, and the luminous flux is maintained at 430 ± 5 lm. Later, between 30 and 40 min, the central coordinates are (x = 0.4366 ± 0.003, y = 0.4012 ± 0.003), the SDCM scores are higher than 7, the color placement is more discrete, and the optical coherence is poorer.