1. Introduction

The rapid development of liquid crystal (LC) panel technology has led to liquid crystal display (LCD) technologies being widely used in various devices, ranging from large products, such as televisions, computer monitors, and notebooks, to small products, such as mobile phones and various consumer electronics. In addition to providing optimal display quality and high functionality, the aforementioned devices have energy-conservation properties, which have been widely researched [1]. Although technological advancements have improved the energy efficiency of LCD technologies, the overall optical efficiency remains relatively low. In general, approximately 6%–9% of the light emitted from an LCD backlight module is transmitted through the LC panel, primarily because of the polarizer and color filter in this panel. The polarizer absorbs over 50% of incident light, and the color filter absorbs approximately 70%. Therefore, both of these components considerably reduce the optical efficiency. Currently, a polarized backlight can be used for effectively reducing absorption by the polarizer.

Absorption by the color filter can be eliminated by using two devices: the field sequential color display (FSCD) and color separation display. The FSCD has many merits, such as not requiring a color filter and subpixels; therefore was expected to increase the transmittance of the LC panel considerably. However, the FSCD could not be easily commercialized because of a problem related to color breakup and the fast scan rate [2]. In the color separation display, the light emerging from a backlight is separated into red, green, and blue light and made incident on the corresponding subpixels, leading to a considerable increase in the transmittance. According to the working principle of the color display, the proposed approaches for realizing color separation can be divided into two types: diffractive optics and geometrical optics. The approaches classified under diffractive optics are described as follows. First, a sandwich grating structure is introduced to deflect the three primaries (RGB colors) into different diffractive orders to spatially separate them [3]. Second, dual gratings and dual lenticular lenses are combined into a hybrid structure to laterally separate the incident light into spatial distributions according to its wavelengths [4]. Third, a combination of a blazed grating and a microlens array is used. The grating separates the incident light into angular distributions according to its wavelengths, and each microlens focuses these angular components into positional distributions in the focal plane [5–10]. The approaches classified under geometrical optics are as follows. First, a huge lenticular lens with angularly positioned LEDs of the three primary colors is used for producing the angularly separated collimated primaries, and a lenticular lens array is then used for focusing them onto the corresponding pixels [11]. Second, LEDs of the three primary colors are positioned at the thin end of a wedge-shaped light-guide plate (LGP) with a spherical thick end to angularly separate the primaries, and a lenticular lens array is subsequently used for focusing them onto the corresponding pixels [12]. Third, a lenticular lens array with grouped lens elements is used for directly focusing linear light sources of red, green, and blue on the corresponding subpixels [13]. The first two approaches are for an edge-lit backlight, while the third approach is for a direct-lit backlight. Although the literature on color separation involving geometrical optics approaches is less extensive than that on color separation involving diffractive optics approaches, the approaches using geometrical optics are superior to those using diffractive optics. The diffractive optics approaches have the following drawbacks: first, precisely manufacturing large gratings is difficult. Second, a grating with a small pitch reduces the optical transmission efficiency, although it can facilitate angular separation of the incident light. This contradiction cannot be resolved. Third, the incident light must be highly collimated. By contrast, manufacture of geometrical optics is comparatively easy. Most importantly, the geometrical optics approaches have no contradiction between angular separation and optical efficiency. Although these approaches also require collimated incident light, their tolerance to collimation is higher than that of diffractive optics approaches. Therefore, using geometrical optics approaches is relatively more feasible for color separation. However, the aforementioned geometrical optics approaches also have some disadvantages. The two approaches applicable to an edge-lit backlight have two disadvantages: First, they separate the color in the transverse direction, and the primary colors are angularly distributed in the sequence R–G–B–G–R because of reflection from both sidewalls of the LGP. Consequently, there are five primary beams in the transverse direction, implying that the primary color beams are close to each other. Therefore, highly collimated incident light is required for avoiding an overlap between two primaries, hindering color separation. Moreover, the five primary beams composing four-subpixel arrangement is different from the traditional three-subpixel arrangement, resulting in a lower resolution. Second, high spatial uniformity of the backlight is difficult to achieve. The angular distribution of the light emitted from the LED must be precisely specified so that the light entering the LGP is collimated and uniformly distributed. This concern becomes more severe when the backlight area is large. Although an improved approach combining spatial and temporal color separation has been proposed for overcoming the problem of low resolution, it is complex and costly [14]. Moreover, its spatial uniformity is not satisfactory. In the approach applicable to a direct-lit backlight, accurately aligning the backlight and the LC panel is difficult. In addition, the backlight thickness increases considerably. Furthermore, high uniformity is difficult to achieve.

On the basis of the advantages and disadvantages of the aforementioned approaches, this paper proposes an all-new design for an edge-lit color-separation backlight to collocate with an LCD panel, as shown in Fig. 1. The proposed apparatus features a composite LGP that induces the three primaries (R, G, and B) to uniformly emerge with angular separation in the longitudinal direction; the three primaries are then separately focused onto the R, G, and B subpixels by a lenticular array attached on the bottom of the LC panel. Because only three primaries are angularly distributed in the longitudinal direction, the consequent three-subpixel arrangement is identical to the traditional subpixel arrangement and beneficial for electric layout. Moreover, the overlap between the primaries also decreases, beneficial for the color gamut of the LCD. In addition, this design can be applied to a large-area backlight.

 

Fig. 1 Scheme of a LC panel operating with a color-separation backlight.

Download Full Size | PDF

2. Design concept and model principles

The key to realizing the proposed apparatus is to design (1) a composite LGP that can induce the three primaries to uniformly emerge with angular separation and (2) an LED coupler for providing uniform and collimated incident light, which is necessary for the composite LGP. The design of (1) is necessary for uniformly illuminating an LC panel for each primary, and the design of (2) is required for suppressing the overlap between the primaries. The composite LGP and LED coupler is detailed in the following paragraphs.

2.1 Composite LGP

To induce the three primaries to uniformly emerge from the apparatus with angular separation in the longitudinal direction, the three primaries were separately coupled into an LGP at their respective angles with respect to the horizontal. They were reflected by the micro V-grooves on the bottom surface of the LGP, resulting in their emergence from the LGP at three distinct angles with respect to the normal. However, the three primaries propagated in the LGP at various angles with respect to the horizontal, and therefore, the spatial distribution of the illuminance of the three primaries above the top surface of the LGP must be different. To evaluate the effect of the propagation direction on the luminance distribution, a simulation was conducted. In the simulation, the half-angle of the three primaries was assumed to be 5°, and their propagation angles (measured in air) with respect to the horizontal were set to 0°, ± 10°, and ± 10° (R, G, and B, respectively). Moreover, the LGP was composed of polymethylmethacrylate (PMMA), and its dimensions were 480 mm × 80 mm × 2 mm. Furthermore, the micro V-grooves were initially optimally distributed on the bottom surface of the LGP to induce the red primary to uniformly emerge from the top surface of the LGP, and the simulation was then conducted for the other two primaries. The simulation results are shown in Fig. 2. The spatial luminance of the red primary was highly uniform as anticipated, but the uniformities of the other two were low. The greater the propagation angle deviates from the horizontal, the lower the uniformity. The main reason for this decrease in uniformity is that the counts of the propagating light hitting the V-grooves increase with the propagation angle. Thus, the primary with a large propagation angle tended to emerge in the region near the front end of the LGP and became absent in the region near the rear end. Therefore, it is crucial to induce the three primaries to uniformly emerge at their respective angles in the longitudinal direction.

 

Fig. 2 Longitudinal spatial distributions of illuminance above the top surface of the LGP for light propagating at various angles with the horizontal in the LGP.

Download Full Size | PDF

To induce the three primaries with their respective propagation angles to uniformly emerge from the top surface of the LGP, the counts of the three primaries hitting the micro V-grooves must be substantially equalized. We propose a design for a composite LGP for achieving this objective. The design concept is as follows. The hitting counts of light propagating in an LGP depend on both the propagation angle and LGP thickness (Fig. 3(a)), and the relationship can be expressed as

 

Fig. 3 Design concept of a composite LGP for two primaries propagating in different directions.

Download Full Size | PDF

Similarly, a composite LGP for three primaries separately propagating in different directions can be designed on the basis of the same concept, as shown in Fig. 4. The key to realizing this design is to appropriately arrange the three primary light sources by using a well-designed LED coupler.

 

Fig. 4 Design of a composite LGP accommodating three primaries that separately propagate and emerge in different directions.

Download Full Size | PDF

2.2 LED coupler

The LED coupler used in this design must satisfy two requirements: it should convert the light emitted from an LED into uniform collimated light and guide each of the three primaries into the composite LGP at prescribed angles to ensure that the three primaries propagate separately in different directions within the LGP. To obtain uniform collimated light, we adopted a V-cut compound parabolic concentrator (CPC) with longitudinally extended V-groove microstructures on its bottom surface, as shown in Fig. 5(a). The V-cut CPC can convert the light emitted from an LED into uniform collimated light in a thin space, as demonstrated in our previous study. Moreover, multiple V-cut CPCs can be integrated in a side-by-side configuration to form a laterally extended light outlet that emits the uniform collimated light. This light outlet can be used as the light source for a large monolithic LGP, as shown in Fig. 5(b). Therefore, the width of the LGP is theoretically unlimited. For additional details, the reader can refer to [15]. V-cut CPCs were stacked in three layers in the LED coupler so that each layer of the V-cut CPCs converted one primary. The V-cut CPCs were the first major component of the LED coupler.

 

Fig. 5 V-cut CPC for mixing and collimating light: (a) top view; (b) multiple V-cut CPCs are arranged side by side to form a laterally extended light outlet.

Download Full Size | PDF

Next, we used a coupling prism to guide each primary emerging from V-cut CPCs in the same layer into the composite LGP at a prescribed angle. Therefore, the LED coupler was designed as shown in Fig. 6. The apparatus had a two-level folded structure: the upper level accommodated the composite LGP, while the lower level accommodated the three layers of V-cut CPCs. Moreover, the coupling prism connected the composite LGP and the V-cut CPCs. When the light entering the LGP is highly collimated, it must be angularly distributed with symmetry to the horizontal so that the light can be uniformly extracted out of the top surface of the LGP. To guide the collimated light emerging from the V-cut CPCs into the composite LGP at the prescribed angle θ, the lower surface of the coupling prism (CP) was a 45° sloping facet; the upper surface of both CP 1 and CP 2 consisted of two sloping facets so that the collimated light could be guided into the composite LGP at angles of $\pm \theta °$ with respect to the horizontal. The upper surface of CP 3 was a 45° sloping bare facet that reflected the collimated light onto the prism array at the front end of the LGP, leading to the collimated light entering the LGP at angles of $\pm \theta °$ with respect to the horizontal. When the light entering an LGP is highly collimated, the light must enter at angles of $\pm \theta °$ with respect to the horizontal; illuminate the V-grooves on the bottom surface of the LGP; and then uniformly emerges from the top surface of the LGP. In addition, both CP 1 and CP 2 partially protrude into the LGP, and their upper surfaces have dichroic coating layers to ensure that the three primaries separately illuminate the prescribed entrance regions of the LGP. In Fig. 6, Coating 1 reflects the red primary but transmits the other two primaries, and Coating 2 reflects the green primary but transmits the blue primary. Therefore, the blue primary emerging from CP 3 can be transmitted through Coatings 1 and 2. The CP is the second major component of the LED coupler.

 

Fig. 6 Two-level folded structure of the apparatus comprising a composite LGP and three layers of V-cut CPCs.

Download Full Size | PDF

3. Simulation results and discussion

In this section, we discuss a simulation conducted to demonstrate the new backlight design for angularly separating the primaries. The proposed apparatus features a composite LGP that induces the three primaries to uniformly emerge at prescribed angles in the longitudinal direction. The three primaries are then separately focused on the R, G, and B subpixels by using a lenticular array. The apparatus has a two-level structure: the upper level accommodates the composite LGP, while the lower level accommodates three layers of V-cut-CPCs; the coupling prisms connect the LGP and the V-cut CPCs. If the apparatus were to be used as the backlight of a 16:9 LC panel, then its total thickness and length would be 15.5 mm and 500 mm, respectively, for a 23-in. (LEDs arranged on the shorter edge) or 39-in. panel (LEDs arranged on the longer edge). To verify whether the proposed design functions as anticipated, we constructed an optical model and performed a series of simulations. The parameters used in the model are as follows. Six pieces of 3.8 mm × 0.6 mm × 1.0 mm LEDs were used as the original primary sources (two pieces for each of the primaries R, G, and B), which showed an approximately Lambertian luminous intensity distribution with a light-emission area of 2.8 mm × 0.4 mm. Both Fresnel loss and material absorption were taken into consideration in the simulation model. To reduce the simulation time, the optical model comprised only six LEDs, six V-cut CPCs (two V-cut CPCs arranged side by side for each layer), a composite LGP with an area of 500 mm × 80 mm, and three coupling prisms; both its side boundaries were assumed to have completely specular reflective properties, as shown in Fig. 7. Therefore, the simulation results of the model were fully equivalent to those of an identical model containing more pieces of LEDs and V-cut CPCs. In addition, we verified whether the lenticular array separately focused the three primaries on the R, G, and B subpixels. Furthermore, both the optical efficiency and color gamut of the color-filter-free LCD system were evaluated and compared with those of the traditional LCD system. The following paragraph details the simulations and presents a discussion of the results.

 

Fig. 7 Scheme of the optical model of the apparatus: (a) perspective view and (b) side view (not proportionally scaled).

Download Full Size | PDF

3.1 Simulation of the LED coupler

The simulation of the LED coupler is discussed in this subsection. The V-cut CPCs in the first, second, and third layers collimated and uniform the light emitted from a red LED, a green LED, and a blue LED, respectively (Fig. 5). The length and width of the V-cut CPC were set to 460 mm and 40, mm respectively. The pitch of the V-grooves on the bottom surface of the CPC was set to 100 μm, and the apex angle of the V-groove was set to 150°. The V-cut CPC in the first layer had an inlet with dimensions of 2.8 mm × 0.4 mm and an outlet with dimensions of 40 mm × 1.9 mm; the V-cut CPCs in the second and third layers had an inlet with dimensions of 2.8 mm × 0.4 mm and an outlet with dimensions of 40 mm × 4.5 mm. The CP collocated with the first-layer V-cut CPC (CP 1) had an upper surface comprising two sloping facets at angles of 40° and 50° with respect to the horizontal. The coupling prism collocated with the second-layer V-cut CPC (CP 2) had an upper surface with two sloping facets at angles of 36° and 54° with respect to the horizontal, and the coupling prism collocated with the third-layer V-cut CPC (CP 3) had an upper surface with a sloping facet at an angle of 45° with respect to the horizontal. All three coupling prisms had lower surfaces with a sloping facet at an angle of −45° with respect to the horizontal. An array of isosceles prisms with an apex angle of 55° was formed at the front end of the composite LGP to induce the incident light emerging from CP 3 to propagate at angles of $\pm 26°$ with respect to the horizontal in the LGP. Moreover, the upper surfaces of CP 1 and CP 2 had dichroic coating layers (Coatings 1 and 2, respectively). The dichroic coating layers were assumed to be ideal low-pass filters: Coating 1 had a reflection of 1.0 at wavelengths above 600 nm, but a transmission of 1.0 at wavelengths below 600 nm; Coating 2 had a reflection of 1.0 at wavelengths above 490 nm, but a transmission of 1.0 at wavelengths below 490 nm. The spectral reflection of the dichroic coating and the spectrum of three primaries emitted from LEDs are presented in Fig. 8. The peak wavelengths of the three primaries were 465, 525, and 625 nm.

 

Fig. 8 Spectrum characteristics of the dichroic coating layers and primaries from LEDs.

Download Full Size | PDF

The simulation results for the propagation directions of three primaries guided into the composite LGP from the LED coupler on the lower level are presented in Fig. 9. The propagation direction of the light within the composite LGP was represented by the angle with respect to the horizontal. Figure 9 shows that the three primaries angularly separated; thus, the red, green, and blue primaries propagated at the prescribed angles of $\pm 10°$, $\pm 18°$, and $\pm 26°$, respectively. Because the collimation degree of the light emerging from the outlet of the V-cut CPC depends on the ratio of the outlet area of the V-cut CPC to the inlet area, the width of the angular distribution of the red primary was greater than those of the green and blue primaries. Moreover, the primaries partially overlapped one another, causing the color gamut of the display to decrease.

 

Fig. 9 Propagation directions of the light guided into the composite LGP by the LED coupler.

Download Full Size | PDF

3.2 Simulation of the composite LGP

To induce the primaries propagating in different directions to uniformly emerge from the top surface of the composite LGP, the parameters of the composite LGP in Fig. 4 were set using Eqs. (1)–(3) for use in the simulation; the set values are presented in Table 1. Because the red, green, and blue primaries propagate at angles of $\pm 10°$, $\pm 18°$, and $\pm 26°$, respectively with respect to the horizontal, the ratio of the thicknesses of the LGPs in which the red, green, and blue primaries propagate is 1:1.84:2.77 according to Eq. (1). The LGP in which the red primary propagated comprised layer PMMA 1, and its thickness was set to 1.6 mm in the simulation. Consequently, the LGP in which the green primary propagated comprised layers PMMA 1, OA 1, and PMMA 2, and its thickness was set to 2.95 mm. Similarly, the LGP in which the blue primary propagated comprised layers PMMA 1, OA 1, PMMA 2, OA 2, and PMMA 3, and its thickness was set to 4.45 mm. The thickness of layer OA was set to 0.05 mm, and the thickness of the PMMA layers was then determined. Because the boundaries between two primaries were set at $\pm 14°$ and $\pm 22°$, the refractive indices of layers OA 1 and OA 2 were determined to be 1.45 and 1.39, respectively. To ensure that the green primary emerged in the normal direction, micro V-grooves with an apex angle of $108°$ were formed on the bottom surface of the composite LGP. Moreover, the bottom surface had a mirror coating with a reflection of 0.98 to ensure that the blue primary emerged at a prescribed angle with respect to the normal.

Table 1. Parameters of the composite LGP in the simulation.

View Table | View all tables in this article

After the related parameters were determined, we simulated the composite LGP and determined whether the three primaries uniformly emerged from the top surface of the composite LGP at the prescribed angles. The intensity of the three primaries emerging from the composite LGP is shown in Fig. 10. The central directions (i.e., intensity peak) of the red, green, and blue primaries were at angles of $11°$, $0°$, and $-12.5°$, respectively; the angular distributions of the three primaries in the longitudinal direction are plotted in Fig. 10(d). The directions of the red and blue primaries emerging from the composite LGP were substantially symmetrical with that of the green primary, in accordance with our design objective. The angular distribution of the red primary was wider than those of the other two primaries because the V-cut CPC in which the red primary propagated was thinner than the other two V-cut CPCs.

 

Fig. 10 Intensity of the three primaries emerging from the composite LGP: intensity diagrams for the (a) red, (b) green, and (c) blue primaries; (d) angular distributions of the three primaries in the longitudinal direction.

Download Full Size | PDF

Next, the simulation results regarding the illuminance of the three primaries above the top surface of the composite LGP are shown in Fig. 11. Figure 11(a) indicates that no partition line appeared in the center, although two V-cut CPCs were arranged side by side for each layer of the LED coupler. Figure 11(b) indicates the longitudinal spatial distributions of the normalized illuminance of the three primaries along the center of the diagrams in Fig. 11(a). The spatial luminance of the three primaries had good uniformity and performed the substantially identical trend. Therefore, the apparatus can provide good spatial uniformity in both the illuminance and color, which is a key to realize a backlight for color separation. The spatial uniformity was defined as the percentage ratio of the minimal illuminance to the maximal illuminance; thus, the spatial uniformities of the red, green, and blue primaries were 86%, 84%, and 87%, respectively.

 

Fig. 11 Illumination of the three primaries above the top surface of the composite LGP: (a) illumination diagrams for the three primaries and (b) normalized illuminance along the central longitudinal line of the diagrams.in (a)

Download Full Size | PDF

3.3 Simulation of the three primaries separately focused onto subpixels

We designed a lenticular array attached to a color-filter-free LC panel to separately focus the three primaries onto their corresponding subpixels. Each lenticular lens was aligned with one pixel that comprised three subpixels with a black matrix intervening, as shown in Fig. 12. Because the three collimated primaries emerging from the backlight with angular separation would be focused on the focal plane of a lenticular lens, we determined the related parameters to ensure that the focal plane of the lenticular array would substantially coincide with the pixel plane; thereby, the three primaries were focused onto their respective subpixels. The related parameters described in Fig. 12 are detailed in Table 2. In the simulation, the pitches of both the pixel and lenticular lens were 300 μm, the pitch of the subpixels was 80 μm, and the width of the black matrix intervening between subpixels was set to 20 μm to ensure that it absorbed color-mixed light and prevent the color gamut of the LCD system from decreasing. The determination of the black matrix width is detailed in the next paragraph. For simplicity, the refractive indices of the materials in the LC panel and lenticular lenses were set to 1.5.

 

Fig. 12 Structure of a unit cell of the LC panel operating with a color-separation backlight.

Download Full Size | PDF

Table 2. Parameters of the LC panel in the simulation.

View Table | View all tables in this article

We simulated the illumination of the pixels of the LC panel operating with a color-separation backlight, and the results are shown in Fig. 13. The figure shows that the primaries partially overlapped one another; the region illuminated by the red primary is the widest and that illuminated by the blue primary is the narrowest. To access the wide color gamut, the color-mixed light must be absorbed as much as possible; therefore, the black matrix (gray bars in Fig. 13) was positioned in the overlap region for absorbing the color-mixed light. However, the wider the black matrix is, the less the amount of transmitted light and the lower the optical efficiency. Therefore, we made a compromise between the color gamut and optical efficiency. Thus, the width of the black matrix was set to 20 μm, and the pitch of the subpixel was 80 μm. Furthermore, because the propagation directions of the three primaries emerging from the subpixel were around the normal, a diffuse film would be attached to the panel screen to appropriately widen the angular distributions of the three primaries emerging from the subpixels for more observers [16].

 

Fig. 13 Illumination on the pixels of the LC panel: (a) Normalized illuminance along the longitudinal; (b) illumination on RGB subpixels.

Download Full Size | PDF

3.4 Analysis of the color gamut of the LCD with the color-separation backlight

On the basis of the parameters listed in Table 2, we evaluated the color gamut of the color-filter-free LCD operating with the color-separation backlight. The coordinates of the white point in the CIE 1931 diagram were set to (0.312, 0.329), and therefore, the ratio of the luminous flux of the three primaries emerging from the subpixels was determined to be 4.95:11.72:1 (R:G:B) so that the light emerging from the three subpixels could be precisely mixed to obtain white light. Subsequently, we modulated the ratio of the luminous flux of the LEDs (light source) of the three primaries in the simulation to realize the target ratio of the luminous flux of the three primaries emerging from the subpixels. The ratio of the luminous flux of the LEDs of the three primaries was determined to be 6.32:11.94:1 (R:G:B). Furthermore, in the simulation, the red, green, and blue subpixels were activated in turns to attain the coordinates (CIE 1931) of the red, green, and blue colors of the LCD, and the color gamut of the LCD was determined, as shown in Fig. 14. The triangle drawn using solid white lines is the color gamut defined by the National Television System Committee (NTSC), while the triangle drawn using dashed white lines is the color gamut of the LCD operating with the color-separation backlight. The coordinates of the three primaries emitted from the LEDs used as light sources for the V-cut CPCs are also marked. The figure shows that the area of the color gamut of the LCD was smaller than that of the three primaries because of color mixing in the subpixels; the area of the color gamut of the LCD was 99.5% of that defined by the NTSC. Therefore, the color filter can be removed for such an LCD, preserving only the black matrix. Related data are presented in Table 3.

 

Fig. 14 Evaluation of the color gamut of the color-filter-free LCD with the color-separation backlight.

Download Full Size | PDF

Table 3. Coordinates of the color gamut in the CIE 1931 diagram.

View Table | View all tables in this article

3.5 Efficiency analysis

Although the color-filter-free LCD system exhibited a wide gamut, the energy efficiency is also crucial for assessing the performance of nonimaging optical devices. Because of recent concerns regarding energy conservation, devices must exhibit high performance and energy efficiency. Therefore, the energy efficiency of the major components of the entire system was further analyzed. The LCD system included a color-filter-free LC panel and a color-separation backlight, which included the LED coupler and composite LGP. The simulation results for the optical efficiency for each component are listed in Table 4. The optical efficiency of the entire system involving the red primary was lower than that for the other two primaries. The main reason is that the angular distribution of the red primary emerging from the V-cut CPC was not sufficiently narrowed, leading to substantial loss of the red primary in the coupling prism and LC panel. The wider angular distribution of the red primary resulted from the corresponding V-cut CPC having a finite thickness. If the thickness of the V-cut CPC were to be increased, then the angular distribution of the emerging light would be more concentrated. However, the thickness of the entire apparatus would proportionally increase. Moreover, the optical efficiency of the composite LGP was lower than that of other components regardless of the primary because the reflective coating layer on the bottom surface of the composite LGP absorbed light. The average optical efficiency of the entire system was 41%. Absorption by a pair of polarizers attached to the LC panel was not considered in the computation of the optical efficiency.

Table 4. Optical efficiency of the color-filter-free LCD system regarding the three primaries.

View Table | View all tables in this article

To compare the optical efficiency of the color-filter-free LCD system with that of the traditional LCD, we simulated the white light from a traditional backlight transmitted in an LCD with a color filter thereon. The light sources of the traditional backlight were composed of red, green, and blue LEDs; the transmissive spectrum of the color filter is shown in Fig. 15. Both the LEDs and black matrix used in the traditional LCD system were assumed to be identical to those in the color-filter-free LCD system. To obtain the same white point, the ratio of the luminous power of the LEDs of the three primaries was determined to be 4.45:8.88:1 (R:G:B). The total optical efficiency of the traditional LCD can be approximately expressed as

 

Fig. 15 Transmissive spectrum of the color filter.

Download Full Size | PDF

3.6 Summary

The simulations demonstrated that the proposed planar apparatus as a color-separation backlight could operate with a color-filter-free LC panel to build a color-filter-free LCD system, which performed the high optical efficiency and wide color gamut. The planar apparatus was a two-level folded structure comprising an upper composite LGP and lower LED coupler; the LC panel had a lenticular lens array on its bottom surface and a black matrix on its top surface, but no color filters. The LED coupler had three layers of V-cut CPCs to collimate the light separately emitted from the red, green, and blue LEDs respectively into three primary beams; then the primary beams were guided into the composite LGP at their respective prescribed angles of $\pm 10°$, $\pm 18°$, and $\pm 26°$ with the horizontal by the coupling prisms. The composite LGP received the primary beams and induced them to uniformly emerge from its top surface with three angularly separated primaries (R, G, and B) in the longitudinal direction. The intensity peaks of the red, green, and blue primaries were at the angles of $11°$, $0°$, and $-12.5°$, respectively with the normal of the top surface; the uniformities of the spatial illumination of the red, green, and blue primaries above the top surface were 87%, 84%, and 86%, respectively. The lenticular lens array on the bottom surface of the LC panel separately focused the three primaries on their respective subpixels on the top surface of the LC panel. The LCD system performed the color gamut 99.5% of the NTSC and optical efficiency of 41% that two times the traditional LCD system. The proposed apparatus has the following advantages due to its two major components, the composite LGP and LED coupler. First, the composite LGP can induce the primaries to uniformly emerge with three-only angularly separated primaries in the longitudinal direction, the lenticular lens array focuses them into their respective subpixels, and thereby three subpixels make up one pixel (R-G-B). Therefore, the resolution is higher than the design with five angularly separated primaries in the transverse. As the resolution of the LCD product trends higher (e.g. 4k2k product) today, this superiority becomes more obvious. Moreover, because the angular space is shared by only three angularly separated primaries, the required collimation degree of the emerging primaries is lower, leading the backlight more easily to satisfy the requirement. Further, the layout of the R-G-B pixel is identical to that of the traditional LCD, so is the circuit layout of the LC panel. Second, the LED coupler contains the V-cut CPCs that can highly collimate and sufficiently uniform the light emitted from the LED at the same time, so it can provide better lateral spatial uniformity than the design using a huge lenticular lens [10, 13]. Moreover, multiple V-cut CPCs can be integrated in a side-by-side configuration to form a laterally extended light outlet that emits the uniform collimated light, which can be used as the light source for a large monolithic LGP applied to a large backlight. Furthermore, the light emitted from the outlet is highly collimated in both the lateral and vertical directions, which aid the light emerging from the composite LGP be more accurately focused onto the subpixels.

Although the color-filter-free LCD system performed the wide color gamut and high optical efficiency, there are two aspects worthy of further improvement in the future: optical efficiency and thickness. The optical efficiency can be improved by the methods such as further collimating the primary emitted from the LED coupler [17], reducing the size of the LED, tapering the composite LGP, narrowing the width of the black matrix, and polarizing the emerging light from the LGP. Reducing the thickness by integrating three layers of the V-cut CPCs is a potential way.

4. Conclusion

This study proposes a slim planar apparatus for uniformly emitting three primaries from the top surface at their respective angles differing in the longitudinal direction, which can be used as a color-separation backlight for the color-filter-free LCD system. The proposed apparatus has a two-level folded configuration comprising a composite LGP and LED couplers. In the simulation, the LED coupler had a V-cut CPC for collimating the primary emitted from an LED and a coupling prism for coupling the primary into the composite LGP at a prescribed angle depending on color. The three coupled primaries propagating at their respective angles in the composite LGP were uniformly extracted out of the top surface of the LGP at three angles differing in the longitudinal direction. Furthermore, we designed a lenticular lens array attached to a color-filter-free LC panel for focusing the three primaries onto their respective subpixels. Each pixel of the LC panel had three subpixels, identical to the traditional layout, so this layout does not suffer from decrease in resolution. In the optical model, the proposed planar apparatus can be applied in backlight modules with a maximal diagonal dimension of 39 in. and a thickness of 15.5 mm. The simulation results indicated that the spatial uniformities of the three primaries ranged between 84% and 87%, the color gamut of the LCD system was 99.5% of that defined by the NTSC, and the total optical efficiency of the LCD system was 41%, as the double of a traditional LCD system. Therefore, the proposed apparatus not only improved the total optical efficiency of the LCD system markedly but also provided the LCD system with a wide color gamut.