1494

Research Article

Vol. 55, No. 6 / February 20 2016 / Applied Optics

Fast optimization method based on the diffuser dot density for uniformity of the backlight module BING-LE HUANG1,2

AND

TAI-LIANG GUO2,*

1

Information Engineering Department, Fujian Chuanzheng Communications College, Fuzhou, Fujian 350007, China College of Physics and Information Engineering, Fuzhou University, Fuzhou, Fujian 350002, China *Corresponding author: [email protected]

2

Received 3 November 2015; revised 21 December 2015; accepted 11 January 2016; posted 11 January 2016 (Doc. ID 251679); published 19 February 2016

A fast optimization method based on the diffuser dot density (DDD) for uniformity of the backlight module (BLM) is proposed in the paper. First, the relationship between the efficiency of the light emerging and the DDD is analyzed, and then a simulation model that is employed to acquire a serial of simulating data is constructed. Second, a mathematic method to profit the relationship is adopted, and a polynomial relationship is derived. Finally, an algorithm to adjust the DDD and optimize the uniformity of the BLM based on the DDD is constructed. The simulation results prove that only by three times optimization, the uniformity of the BLM can reach 85.6%, and the experimental result indicates that the algorithm proposed in the paper can improve the uniformity rapidly. The final experimental result is that the uniformity of the third optimization reaches 77.4%, which satisfies the target 75% in the phase of designing the BLM. Compared to the conventional optimization method, the method can speed up the procedure and lower the expense of developing the BLM in fabricating the liquid-crystal display. © 2016 Optical Society of America OCIS codes: (150.2945) Illumination design; (220.4830) Systems design; (220.0220) Optical design and fabrication; (230.3720) Liquid-crystal devices; (130.3120) Integrated optics devices. http://dx.doi.org/10.1364/AO.55.001494

1. INTRODUCTION The liquid-crystal display (LCD) is non-self emitting; it needs a backlight module (BLM) that provides a surface light source, which is converted from a point- or line-light source. A typical edge-type BLM, as shown in Fig. 1, includes a light source, a light-guide plate (LGP), and several optical films for scattering and condensing light. An effective way to extract the light out of the LGP to achieve an equal-luminance (EL) condition is to optimize the shape or the pattern of the light-scattering dots on the bottom of the LGP. Owing to the extract ability of the diffuser dot, the dot pattern design plays a crucial role in achieving an EL condition [1–5]. For example, a highluminance level can be achieved for a high-density configuration of the diffuser dots at a specific position or for a configuration of the same dot density at a position closer to the light source. The luminance level of the dot achieved is a function of both its density and its location [6–11]. Thus a new methodology of dot pattern design that utilizes optical design to provide luminance uniformity information and that achieves the EL condition is very useful to backlight products, and it can shorten the time to market. Thus the design of the 1559-128X/16/061494-06$15/0$15.00 © 2016 Optical Society of America

LGP’s light-scattering dots or pattern plays an important role in achieving a BLM at low cost with high-brightness and highluminance uniformity [12–15]. To achieve better luminance uniformity, Li et al. proposed [16]; it adopted a complicated algorithm, and the algorithm spent plenty of time optimizing the uniformity of BLM. Kim proposed [17]; it adopted the pattern density function with a simple exponential function as the method of optimization. Lee et al. proposed [18]; it adopted two separate steps to improve the max luminance and uniformity. Wang proposed [19]; it integrated the BPN and a revised GA to identify the optimal experimental level combination of the optical brightness parameter design. The method mentioned above adopted complicated algorithms, and the procedure of design and optimization took a mass workload. To assist the BLM design, we combined the relevant software, for example, TracePro, Matlab, and BacklightFly. And this paper considers the simulation luminance as the important factor to design and optimize the BLM, and it proposes a fast and effective method to promote the uniformity of the BLM based on the diffuser dot density (DDD).

Research Article

Vol. 55, No. 6 / February 20 2016 / Applied Optics

1495

A. Curve of the Efficiency According to the DDD Average

This paper takes the 4 in. backlight model as an example to construct the model. The structure of the BLM consists of a reflection sheet, LGP, a diffusion sheet, and two brightness enhancement sheets from the bottom to the top. The light bar consists of 13 LEDs at the same interval, and the specification of the LED is 4014. The thickness of the LGP is 1.6 mm. The thickness of the reflector, diffusion film, and the brightness enhancement film (BEF) is 100 μm each, and the optical characteristics of the light source are set as a Lambertian distribution. The sketch of the BLM is shown in Fig. 1. In order to demonstrate the relationship between the DDD average and the efficiency of the emergency light, a serial of the simulation experiments is tested in this paper. And the simulation result is listed in Table 1 as follows. From Table 1, we can see that when the density value is less than 0.4, the efficiency is rapidly improved by increasing the density. Nevertheless, when the density value is over 0.4, the efficiency is not obviously increased. For example, the efficiency is 0.69 when the density is 0.4, and the efficiency is 0.7 when the density is 0.45, and more, when the density value is larger than 0.6, the efficiency invariably maintains at 0.72. On the other hand, when the density value is larger than 0.7, the diffuser dots will easily overlap in the procedure of the printing ink, so there is not obvious significance to keep increasing the density value. In order to reduce the workload of simulation for the relationship, we adopt the polynomial fitting function according to the results of the experiment listed in Table 1. The procedure of fitting the function is as follows:

Fig. 1. Schematic diagram for an LCD backlight module with an edge-type LGP.

2. WORKING PRINCIPLES OF THE OPTIMIZATION Suppose the DDD of the LGP varied according to the coordination as follows: d x; y  sx; y∕a2 ;

(1)

where sx; y is the area of the diffuser dot at the coordination of x; y, and a2 is the area of the grid in the above formulation in Eq. (1). When the size of the light source and LGP are fixed, the emergency light intensity is proportional to the DDD according to the position. We can obtain the relationship as follows: Bx; y∞d x; y;

(2)

where Bx; y and d x; y in Eq. (2) correspond to the emergency light intensity and the DDD at the point of x; y coordination, respectively. The efficiency of the light emerging from the exiting light surface is also proportional to the intensity of the light emerging. We can have the relationship as follows: Ex; y∞Bx; y;

Step 1: Set the initial value for N (the polynomial exponential) to 1. Step 2: Apply the polynomial exponential N to the data points from the list in Table 1 using Eq. (5). Step 3: Calculate the mean square deviation (MSD). If the MSD is less than 0.01, the procedure for the fitting data is over, or we can increase N by one and jump to Step 2.

(3)

where Ex; y in Eq. (3) corresponds to light efficiency at the point of x; y coordination. According to Eqs. (2) and (3), we can have the relationship as follows: Ex; y∞d x; y:

(4)

In the procedure for fitting data, the efficiency of the light extracted can be expressed by the following formulation:

For the convenience of obtaining the formulation between E and d , we can assume they conform to the form of the sums of the polynomial as follows: Ed  

N X

Ai  d ; i

Ed   5.35d ˆ3 − 7.9848dˆ2  3.8318d  0.1069;

where E and d in Eq. (6) correspond to the efficiency and the DDD average, respectively. And we use the interpolation method to plot the fitting function and the original data curve in Fig. 2 above.

(5)

i

where Ai is the proportional constant, and N is the maximum polynomial exponential. In the following section, we can acquire the parameter of Eq. (5) by constructing the simulation model and fitting the relationship between the DDD and the light emergency efficiency. Table 1.

(6)

B. Algorithm of Optimization for the Uniform of LGP Based on Density

In this paper, we adopt the algorithm of optimization based on the DDD. The factor of adjusting the density not only depends on the luminance of the position, but it also depends on the

Experimental Result of the Simulation Efficiency versus the Average Density

Parameter Density (%) Efficiency (%)

Value 1 8

2 14

3 20

4 26

5 30

6 34

7 37

8 40

9 42

10 44

15 53

20 58

25 63

30 66

35 67

40 69

45 70

50 70

55 71

60 72

65 72

70 72

1496

Research Article

Vol. 55, No. 6 / February 20 2016 / Applied Optics

C. Steps of Optimization

The steps of optimization are as follows: Step 1: Set the initial density of the diffuser dots to 0.25. Step 2: Create the density attribution file according to the density in BacklightFly software. Step 3: Import the density attribution file into the TracePro software. Step 4: Calculate the uniformity of the luminance according to the simulation data based on Eq. (9). If the uniformity is less than 85%, adjust the density distribution according to Eq. (8) and then jump to Step 2 to continue optimizing the uniformity of the BLM or else terminate the optimization procedure. The initial design of the diffuser dots should be set to 0.25, because the average density of 0.25 can not only satisfy the required efficiency, but also when the average density is at the level of 0.25, the uniformity of the BLM can easily reach the target. The specification has been explained in [20], which the author has researched previously.

Fig. 2. Profit curve of the efficiency versus DDD.

current density of the region to be adjusted. The relationship between the efficiency and the density has been discussed above. It can be used to improve the uniformity in this section. According to the principles of the optimization, we construct the algorithm as follows: f  Lave ∕Li;j   Ed ave ∕Ed i;j ;

(7)

d i;j  d i;j  f ;

(8)

where f is the factor of adjusting the DDD, and Lave is the average luminance of the whole effective illuminance area. Li;j and Ed i;j  are the average luminance and the efficiency of the density at the i; j cell, respectively. d ave is the average density of the whole DDD at the bottom of the LGP. d i;j is the density of the diffuser dot at the i; j cell. In this paper, we set d ave to 0.25, and the reason for d ave being set to 0.25 is explained below. As we can see from Fig. 2, the efficiency will be improved by increasing the density, indeed, but if the average density is set to the bigger value, such as the maximum value 0.7, the diffuser dots will overlap each other in some local positions, and more uniformity at the level of the average density will not easily be adjusted to satisfy the requirement. To be specific, when the density of the diffuser dot is set to 0.25 and 0.7, the efficiency is 0.63 and 0.72, respectively. The average density of 0.25 will not induce the phenomenon that the diffuser dots overlap each other in the neighborhood, but the average density of 0.7 will emerge with the phenomenon that diffuser dots overlap each other in the neighborhood. Comparing the efficiency between 0.25 and 0.7, the efficiency of the average density of 0.7 is only 0.03 bigger than that of 0.25. So in order to trade off the optical efficiency and the uniformity, the average density is set to 0.25, and the details have been discussed in the previous research paper by the author in [20]. When d i;j is bigger than 0.5, d i;j must be set to 0.5, because if d i;j exceeds 0.5, the diffuser dots will easily overlap each other in some region in the printing ink procedure by using the mask.

3. SIMULATION A. Simulation Condition

The necessary condition of an ideal BLM is to be a plane wave light source. Thus the incident light from LEDs spreads throughout the inside of the LGP and then is uniformly emitted through the emergent surface. Hence, in order to save the cost of the development and improve the efficiency of the design, the design of BLM, whose goal is to maximize the light intensity and optimize the light intensity distribution on the front surface of LGP, requires the assistance of the illumination design simulating software. We have used TracePro (OSA, American) for simulation analysis of output light intensity characteristics for LGP. In this study, we adopted a plate of the size of 120 mm × 69 mm × 1.6 mm with “optical polished” surface as a basic structure of 4 in. LGP, as shown in Fig. 1. The thickness of the LGP is adjusted to an LED with azimuthal radiation. Each of the 13 light sources of 4.0 mm × 1.4 mm in size (LED is typically used) emits one million rays of 550 nm monochromatic wavelength toward the y direction with a Lambertian intensity distribution. A 100% reflector to prevent escape of light rays from LGP surrounded each of the side surfaces of LGP, except the light incident and emergent ones. The material used for the LGP is polymethyl methacrylate (PMMA). The reflective index and light absorption ratio of the LGP were 1.49 and 1.70 × 10−3 mm, respectively [12]. There are diffuser dots at the bottom of the LGP, and the parameters of the diffuser dots are as follows: the shape of the microstructure is a sphere, and the radius and the thickness are 100 and 50 μm, respectively. The angle of the microstructure is 90°. B. Simulation Method and Results

The emergency light surface was divided into 72 × 128 cells at the same interval. Then the density of each cell was adjusted according to the steps of the optimization mentioned above. Figure 3 describes the method of calculating the uniformity of the BLM in the phase of the simulation design. The emergent surface margin of the BLM takes off 1.5 mm, and then the rest of the length H and the width V of the emergent surface are divided into 7 × 7 blocks at the same interval, so the total of

Research Article

Vol. 55, No. 6 / February 20 2016 / Applied Optics

1497

Table 2. Uniformity Calculated by Eq. (9) Optimization Index Uniformity

Fig. 3. Scheme of the method of the uniformity calculated for simulation luminance.

49 points can be acquired. The uniformity is defined in the following formulation: L (9) Uniformity  min × 100%; Lmax where Lmin and Lmax are the minimum and maximum values for the 49 points or 19 points of the simulating luminance or the tested luminance, respectively. In the simulation phase, the simulation data are very abundant, so in order to fully employ the data and promote the efficiency of the optimization, we adopt 49 points to calculate the uniformity; but in the experiment phase, the 19 points are the industry testing standard, so we take 49 points in simulation and 19 points in experiment, respectively. The optimization and simulation results are listed in Fig. 4. Figure 4(a) is the irradiance map of the uniform density that is set to initial 0.25. Figure 4(b) is the irradiance map of the first optimization based on density adjustment according to the algorithm proposed in the paper. Figure 4(c) is the irradiance map of the second optimization. Figure 4(d) is the irradiance map of the third optimization.

The First The Second Optimization Optimization 79.2%

The Third Optimization

80.5%

85.6%

We can find that Fig. 4(c) is fairly uniform for the emergency light surface, obviously. We calculate the uniformity of the three optimization results according to Eq. (8). The final uniformity is listed in Table 2 above. The uniformity of the third optimization is bigger than 85%, so the optimized procedure is terminated. From the optimization results, we can see the method proposed in the paper is effective to quickly acquire uniformity of the BLM. 4. EXPERIMENT In the experiment, the dimensions of the LGP are 120 mm × 69 mm × 1.6 mm, and the tolerance is 0.02 mm. The tolerance of the surface’s smoothness of the LGP is 0.005 mm; the optical polished tolerance is within 20 nm. The height of the diffuser dot is approximately 0.05 mm; the type of ink is WZP, and the ink and the diluent ratio is MP18∶P750  100∶3. The BLM consists of one sheet of reflector (JAR235), one plate of LGP, one sheet of diffuser (TDF107), and two sheets of BEF (AOS-C10B) from the bottom to the top. The current applied to LED is 75 mA. The method of measuring the uniformity and the illuminance adopts a 19-points test method as shown in Fig. 5(a), and the measure device is BM-7, as shown in Fig. 5(b). Calculating the uniformity adopts Eq. (9), and the illuminance is the average luminance of the 19 points. The calculated results of the uniformity and the illuminance are listed in Table 3. The experimental setup is as follows: Step 1: Export the CAD document according to the density distribution of the diffuser dot from the BacklightFly software. Step 2: Make film according to the CAD image document by the outside joint factories. Step 3: Produce the screen based on the film by using the exposure technique. Step 4: Print the ink on the diffuser dot pattern of the LGP using the mask of the screen. Step 5: Dry the ink of the diffuser dot by IR. Step 6: Construct the backlight using the LGP prepared in the previous steps with other optical film in the module. The photometric, electrical, and thermal characteristics of LED sources are highly dependent on one another. With increasing demand for higher power density, high-brightness LED is facing challenging thermal problems that affect optical uniformity and reliability [21–23]. And the package of LED Table 3. Uniformity and Average Luminance Optimization Index

Fig. 4. Irradiance map of the initial, first, second, and third optimized simulation luminance.

Uniformity Luminance (nit)

The First The Second The Third Optimization Optimization Optimization 69.6% 6268

73.5% 6225

77.4% 6275

1498

Research Article

Vol. 55, No. 6 / February 20 2016 / Applied Optics

uniformity rapidly. The final experimental result is that the uniformity of the third optimization reached 77.4%, which satisfies the target 75% in the phase of designing the BLM. Compared to the conventional optimization method, the method can speed up the procedure and lower the expense of developing the BLM in fabricating the LCD. Funding. National Youth Science Foundation (NYSF) (61405037); The 863 Program (2013AA030601-2); Department of Education, Fujian Province (JA13376); Fujian Province Colleges and Universities Outstanding Youth Scientific Research Personnel Training Plan. Acknowledgment. We thank the Admiral Overseas Corporation (FuQing Branch) for the use of their equipment. REFERENCES

Fig. 5. Testing condition of the experiment. (a) Method of 19-points testing. (b) Testing system of BM-7.

used in the paper is 4014, which means 4 mm of length and 1.4 mm of width. From Table 3, we can see that adopting the algorithm proposed in the paper can improve the uniformity rapidly, and the uniformity of the third optimization can reach the target of 75% in the phase of designing the BLM. The experimental results are compared with the simulation results, and the two show good agreement. There is a large gap between the simulation results (85%) and that of the experiment results (77.4%). The reasons are as follows: the optical attribution of the diffuser dot in the simulation is hardly kept the same as the ink in the experiment, because the diffuser dot in the simulation phase only considers the reflective and refractive attributions and ignores the absorptive attribution. And the difference between them will not obviously affect the rule of the optimization. So the simulation model is useful and simple regardless of the particulars from reality. The uniformity will be further promoted by varying the component of the ink in the fabrication phase. 5. CONCLUSION In summary, we have presented a fast optimization method based on the DDD for uniformity of the BLM. By adopting the method proposed in the paper, we only spent three times the optimization, and the uniformity of the BLM can achieve 85.6% in the simulation. Then the three times optimization results were applied to the experiment, with the result indicating that the algorithm proposed in the paper can improve the

1. H.-T. Huang, C.-C. Tsai, and Y.-P. Huang, “Conformal phosphor coating using pulsed spray to reduce color deviation of white LEDs,” Opt. Express 18, A201–A206 (2010). 2. Z. Qin, K. Wang, F. Chen, X. Luo, and S. Liu, “Analysis of condition for uniform lighting generated by array of light emitting diodes with large view angle,” Opt. Express 18, 17460–17476 (2010). 3. K. Wang, F. Chen, Z. Liu, X. Luo, and S. Liu, “Design of compact freeform lens for application specific light-emitting diode packaging,” Opt. Express 18, 413–425 (2010). 4. G. Yating, L. Zhenyue, Z. Ruidong, H. Qi, W. Shin-Tson, L. MingChun, L. Seok-Lyul, and T. Wen-Ching, “A high performance single-domain LCD with wide luminance distribution,” J. Display Technol. 11, 315–324 (2015). 5. B.-T. Chen and J.-W. Pan, “Dual-view angle backlight module design,” Appl. Opt. 54, E80–E87 (2015). 6. J.-G. Chang, Y.-B. Fang, S.-P. Ju, and J.-Y. Hsieh, “Random dot generation scheme using molecular dynamics method for illumination design of a round plane LED source light guide,” Displays 31, 44–53 (2010). 7. K. Wang, D. Wu, Z. Qin, F. Chen, X. Luo, and S. Liu, “New reversing design method for LED uniform illumination,” Opt. Express 19, A830– A840 (2011). 8. B.-l. Huang, T.-l. Guo, and J.-m. Yao, “Design of slanted zigzag staggered barrier for autostereoscopic display,” Acta Photon. Sin. 43, 10050021 (2014). 9. D. Hao, K. Qian, and Y. Luo, “Use of adjusted molecular dynamics method for dot pattern design in large scale light-emitting diode edge-lit backlight unit,” Opt. Eng. 50, 104001 (2011). 10. C.-Y. Li and J.-W. Pan, “High-efficiency backlight module with two guiding modes,” Appl. Opt. 53, 1503–1511 (2014). 11. P. Xu, Y. Huang, Z. Su, and X. Zhang, “Algorithm research on microstructure distribution on the bottom surface of an integrated microoptical light guide plate,” Appl. Opt. 53, 1322–1327 (2014). 12. Y. C. Kim, T.-S. Oh, and Y. M. Lee, “Optimized pattern design of lightguide plate (LGP),” Opt. Appl. 41, 863–872 (2011). 13. S. Park, Y. Shin, E. Choi, H. Ma, and S. Lee, “Improvement of luminance and uniformity of light guide panel using scatterer pattern by laser processing,” Opt. Laser Technol. 44, 1301–1306 (2012). 14. J.-J. Chen, T.-Y. Wang, K.-L. Huang, T.-S. Liu, M.-D. Tsai, and C.-T. Lin, “Freeform lens design for LED collimating illumination,” Opt. Express 20, 10984–10995 (2012). 15. P. Xu, Y. Huang, X. Zhang, J. Huang, B. Li, E. Ye, S. Duan, and Z. Su, “Integrated micro-optical light guide plate,” Opt. Express 21, 20159– 20170 (2013). 16. C.-J. Li, Y.-C. Fang, and M.-C. Cheng, “Study of optimization of an LCD light guide plate with neural network and genetic algorithm,” Opt. Express 17, 10177–10188 (2009). 17. Y. C. Kim, “Optimize pattern design for the thin LGP,” Optik 124, 2171–2173 (2013).

Research Article 18. G. Lee, J. H. Jeong, S.-J. Yoon, and D.-H. Choi, “Design optimization for optical patterns in a light-guide panel to improve illuminance and uniformity of the liquid-crystal display,” Opt. Eng. 48, 024001 (2009). 19. S.-T. Wang, “Optimized light guide plate optical brightness parameter: integrating back-propagation neural network (BPN) and revised genetic algorithm (GA),” Mater. Manuf. Processes 29, 1–8 (2014). 20. B.-L. Huang, T.-L. Guo, E.-G. Chen, and S. Xu, “Study on the optimal scale of the average netted dot density for light guide plate,” Acta Opt. Sin. 35, 5220021 (2015).

Vol. 55, No. 6 / February 20 2016 / Applied Optics

1499

21. K. Neyts, M. Marescaux, A. U. Nieto, A. Elschner, W. Lovenich, K. Fehse, Q. Huang, K. Walzer, and K. Leo, “Inhomogeneous luminance in organic light emitting diodes related to electrode resistivity,” J. Appl. Phys. 100, 114513 (2006). 22. H.-T. Chen, X.-F. Zhou, Y. Zhang, S. Lin, J.-R. Zhou, and Y. Fei, “Study on the staggered array of an LED system for improved thermal behavior,” Appl. Opt. 54, 6752–6757 (2015). 23. H.-T. Chen, W. C. H. Choy, and S. Y. Hui, “Characterization, modeling, and analysis of organic light-emitting diodes with different structures,” IEEE Trans. Power Electron. 31, 581–592 (2016).

Fast optimization method based on the diffuser dot density for uniformity of the backlight module.

A fast optimization method based on the diffuser dot density (DDD) for uniformity of the backlight module (BLM) is proposed in the paper. First, the r...
732KB Sizes 0 Downloads 7 Views