OPEN (17) 6, e17043; doi:1038/ls.17.43 Officil journl of the CIOMP 47-7538/17 www.nture.com/ls ORIGINAL ARTICLE Going eyond the limit of n LCD s color gmut Hi-Wei Chen 1, Rui-Dong Zhu 1, Jun He 1, Wei Dun 2, Wei Hu 2, Yn-Qing Lu 2, Ming-Chun Li 3, Seok-Lyul Lee 3, Y-Jie Dong 1,4 nd Shin-Tson Wu 1 In this study, we nlyze how cklight s pek wvelength, full-width t hlf-mximum (FWHM), nd color filters ffect the color gmut of liquid crystl disply (LCD) device nd estlish theoreticl limit, even if the FWHM pproches 1 nm. To overcome this limit, we propose new cklight system incorporting functionl reflective polrizer nd ptterned hlf-wve plte to decouple the polriztion sttes of the lue light nd the green/red lights. As result, the crosstlk etween three primry colors is gretly suppressed, nd the color gmut is significntly widened. In the experiment, we prepre white-light source using lue light-emitting diode (LED) to pump green perovskite polymer film nd red quntum dots nd demonstrte n exceedingly lrge color gmut (95.8% in Commission interntionle de l'éclirge (CIE) 1931 color spce nd 97.3% in CIE 1976 color spce) with commercil high-efficiency color filters. These results re eyond the color gmut limit chievle y conventionl LCD. Our design works eqully well for other light sources, such s 2-phosphorconverted white LED. (17) 6, e17043; doi:1038/ls.17.43; pulished online 8 Septemer 17 Keywords: color gmut; liquid crystl displys; perovskite; polriztion; quntum dot INTRODUCTION The liquid crystl disply (LCD) hs ecome uiquitous in our dily lives. Widespred pplictions rnge from smrtphones, tlets nd computer monitors to TVs, just to nme few 1. An essentil requirement for disply devices is n ccurte representtion of color. To widen the color gmut of n LCD, strightforwrd pproch is to employ cklight with highly sturted red, green nd lue (RGB) primry colors 2. Generlly, light source with nrrower full width t hlf mximum (FWHM) would led to wider color gmut 3.During the pst two decdes, LCD cklights hve evolved from lue-pumped yellow YAG: Ce 3+ phosphor-converted white-light-emitting diodes (1pc-WLED: FWHM ~ 1 nm) to green nd red phosphorsconverted WLEDs (2pc-WLED: FWHM ~ 50 nm) nd to quntum dots (QDs: FWHM ~ 25 30 nm) 4 8. Recently, orgnic inorgnic perovskite (OIP) ws found to hve nrrower FWHM (~18 nm) thn tht of QDs 9,10. As result, it holds potentil to further enhnce the color gmut of n LCD. However, n interesting question to sk is: Is there theoreticl limit to the color gmut of n LCD, even if the FWHM of the light-emitter pproches zero, or o1nm, such tht of lsers? In ddition to cklighting, nrrow-nd color filter (CF) is nother option tht cn e used to enlrge the color gmut, except tht its opticl efficiency is compromised 11. Stte-of-the-rt QDenhnced cklights, long with specilly designed nrrow-nd CFs, could chieve 92% (Ref. 12), ut the trdeoff is ~ 25% reduced efficiency. For ttery-powered disply, this 25% light efficiency loss is hrdly cceptle. Therefore, new pproches to chieve wide color gmut tht mintin high opticl efficiency re urgently needed. In this pper, we first investigte how the FWHM of light source ffects the gmut of LCD colors nd find theoreticl limit, even if FWHM pproches 1 nm. Next, we propose new cklight system incorporting functionl reflective polrizer (FRP) nd ptterned hlf-wve plte to suppress the crosstlk originting from CFs, which in turn significntly widens the color gmut. In the experiment, we prepre white-light source using lue light-emitting diode (LED) to pump green OIP-polymer composite film nd red QD nd demonstrte n exceedingly wide color gmut, with 95.8% in Commission interntionle de l'éclirge (CIE) 1931 color spce nd 97.3% in CIE 1976 color spce, using commercil highefficiency CFs. Our result is comprle to tht of lser projection displys ut with direct-view LCD pnels. In ddition to QDs nd perovskites, our design lso works well for other light sources, such s 2pc-WLED. MATERIALS AND METHODS Before evluting the color gmut of n LCD, let us first elucidte its definition, s this definition is sometimes confusing nd misleding. Severl stndrds hve een proposed to quntify color gmut (for exmple, srgb, Adoe RGB, NTSC nd so on). Here we focus on the most comprehensive stndrd, clled, with RGB lsers. Rec. covers ll the existing stndrds 13 16. Next, we define the 1 College of Optics nd Photonics, University of Centrl Florid, Orlndo, FL 32816, USA; 2 Ntionl Lortory of Solid Stte Microstructures, Collortive Innovtion Center of Advnced Microstructures nd College of Engineering nd Applied Sciences, Nnjing University, Nnjing 210093, Chin; 3 AU Optronics Corp. nd 4 NnoScience Technology Center, University of Centrl Florid, Orlndo, FL 32826, USA Correspondence: ST Wu, Emil: swu@creol.ucf.edu Received 23 Decemer 16; revised 8 Mrch 17; ccepted 8 Mrch 17; ccepted rticle preview online 10 Mrch 17
Going eyond the limit of n LCD s color gmut 2 1.2 Red-QD Green-OIP Blue-LED Gussin fitting 400 450 550 650 700 750 Figure 1 Emission spectr for lue LED nd green perovskite nd red QD. Solid lines re mesured dt nd dshed lines re from Gussin fittings. coverge rtio rther thn the re rtio s the color gmut. Under this definition, the otined color gmut should not exceed 100% Rec.. Another importnt thing to choose is the right color spce, CIE 1931 or CIE 1976 (Refs. 17,18). Although CIE suggests using CIE 1976, s it is color uniform spce, mny disply mkers nd reserch groups re still using CIE 1931 to disply their results. To stisfy oth cmps, here we present color gmut in oth CIE 1931 nd CIE 1976. Gussin-fitting effect Figure 1 demonstrtes the mesured spectr of lue LED, green OIP nd red QD; ll exhiit Gussin-like profiles. Thus, Gussin fitting is commonly conducted to extrct the pek emission wvelength nd FWHM. These fitted curves re then employed to clculte the color gmut y neglecting the fitting discrepncy 3,11,17,19.Throughfittings, we otined the pek wvelength nd FWHM of three primry colors s follows: λ B = 45 nm, Δλ B = 2 nm; λ G = 531.6 nm, Δλ G = 19.7 nm; nd λ R = 63 nm, Δλ R = 24.4 nm, respectively. From Figure 1, noticele discrepncy is oserved, especilly for the lue LED. The red QD nd green OIP present longer emission tils thn 0 w/o Gussin fitting R G w/ Gussin fitting c 0 5 0 5 0 w/o Gussin fitting w/ Gussin fitting B Rel spectrum Gussin fitting 400 450 550 650 700 750 d 5 0 5 0 0 5 0 5 0 400 450 550 650 700 750 w/o Gussin fitting w/ Gussin fitting 5 5 0 400 450 550 650 700 750 0 400 450 550 650 700 750 Figure 2 () Trnsmission spectr for commercil CF rry. Output SPDs for () lue, (c) green nd(d) red primry colors with nd without Gussin fitting. doi:1038/ls.17.43
Going eyond the limit of n LCD s color gmut 3 0.9 0.7 5 540 w/ Gussin fitting w/o Gussin fitting 510 5 530 540 550 570 590 610 630 640 680 y 380 x 700 0.7 v 0 450 440 430 4 u w/ Gussin fitting w/o Gussin fitting Figure 3 Color gmut in () CIE 1931 color spce nd () CIE 1976 color spce. Tle 1 Simulted color gmut for the light source with nd without Gussin fitting CIE 1931 CIE 1976 w/ Gussin fitting 89.4% 89.9% w/o Gussin fitting 82.8% 83.9% 400 450 550 650 700 R G B 1 2 Figure 4 Trnsmission spectr of commercil CFs nd different white-light sources. the fitted Gussin functions. As will e shown lter, these tils ggrvte the color crosstlk, which in turn shrinks the color gmut. In n LCD, the cklight psses through the LC lyer nd CFs efore reching the viewer. Therefore, we hve to incorporte the 750 cklight spectr (Figure 1) into the LCD pnel to clculte the output spectrl power distriution (SPD). In our simultion, we hve chosen fringe field switching LCD with negtive dielectric nisotropy (Δεo0) LC 21,22 nd commercil CF rry 11 s n exmple (Figure 2). Such n LCD hs een widely used in smrtphones nd pds. The wvelength-dependent refrctive indices of the employed LC re lso considered in the simultion. Figure 2 2d depicts the otined SPD for the RGB primry colors with or without Gussin fitting. In oth cses, light lekge for ll three chnnels is oserved clerly, especilly for the lue chnnel, where firly lrge ump leks through the green CF. For the rel spectrum without Gussin fitting, the light lekge is even worse, owing to the long-emission tils, which further deteriorte the color purity. Figure 3 illustrtes the color gmut shrinkge more clerly. In oth color spces, green nd lue color coordintes with Gussin fitting expnd outwrd, representing high purity primry colors. The otined color gmut results re summrized in Tle 1. In the CIE 1931 color spce, the color gmut with Gussin fittings is 6.6% wider thn tht using the rel spectr, wheres in CIE 1976, this difference is 6%. For different CFs nd light sources, this discrepncy vries. However, Gussin-fitted curves led to wider color gmut thn using the rel emission spectr. Therefore, to estlish the theoreticl limit of the color gmut of n LCD, we use the Gussin-fitted spectr. Preto front for the color gmut nd light efficiency From Tle 1, the color gmut without Gussin-fitted spectr is ~ 83%. To improve tht, we fine-tuned the emission spectrum of QD nd OIP to mtch the trnsmission nds of CFs, s Figure 4 demonstrtes. However, tuning the pek emission wvelength influences the light efficiency, ecuse the humn eye hs different sensitivities to different colors. This is governed y the humn eye sensitivity function V(λ), which peks t 555 nm. Therefore, we introduced nother metric, the totl light efficiency (TLE), to chrcterize how much input light trnsmits through the LCD pnel doi:1038/ls.17.43
Going eyond the limit of n LCD s color gmut 4 35 32 30 28 25 TLE (Im W 1 ) 15 10 FWHM = 50 nm FWHM = 40 nm FWHM = 30 nm FWHM = nm FWHM = 10 nm FWHM = 1 nm TLE (Im W 1 ) 24 16 FWHM = 50 nm FWHM = 40 nm FWHM = 30 nm FWHM = nm FWHM = 10 nm FWHM = 1 nm 5 0.7 Colour gmut (CIE 1931) 0.9 12 0.7 Colour gmut (CIE 1976) 0.9 Figure 5 Preto front defined in () CIE 1931 nd () CIE 1976 with different FWHM light sources. 400 450 550 650 CF-1 CF-2 700 750 Figure 6 Trnsmission spectr of two commercil CFs nd white-light source. nd finlly gets converted to the rightness perceived y the humn eye 3 : TLE ¼ 683 R S out ðlþvðlþdl R ð1þ Sin ðlþdl In Eqution (1), S in (λ) is the SPD of the light source (tht is, input cklight) nd S out (λ) represents the output SPD fter the LCD pnel. Both color gmut nd TLE re importnt prmeters for disply device. For such multi-ojective prolem, different ojectives my e mutully exclusive. In short, ny further improvement of one ojective (for exmple, color gmut) is likely to e compromised y the degrdtion of nother ojective (for exmple, TLE). Therefore, Preto front is commonly employed, nd ll solutions will fll either on or elow this Preto front (see Supplementry Informtion for more detils). Figure 5 is plot of the clculted Preto fronts of the LCD using cklight with different FWHM vlues. From Figure 5, severl interesting phenomen re oserved. First, there is n inherent trdeoff etween light efficiency nd color gmut. Thus, delicte lnce should e chosen in prcticl pplictions. Second, for cklight with the sme FWHM, there is indeed theoreticl limit for the color gmut, regrdless of the RGB centrl wvelengths. For exmple, when the FWHM of QD or OIP emission spectrum is 30 nm, the lrgest chievle color gmut is 9% Rec. in CIE 1931 (Figure 5) or 91.5% in CIE 1976 (Figure 5). Next, we find tht s the light source ecomes more sturted (tht is, nrrower FWHM), the mximum color gmut increses nd then grdully sturtes. Currently, the commercil Cdsed QD-enhnced cklight exhiits 30-nm FWHM, nd it could rech 25 nm in the next few yers 7.Assumingthtnmcne otined somedy, ccording to Figure 5, the color gmut improvement is only 2%. Even if the FWHM of the light emitters were lserlike ( 1 nm), the mximum color gmut is ~ 93.5%. This limits the color gmut tht n LCD cn possily chieve. Plese note tht this limit is otined with Gussin-fitted spectr; it would e lowerifrelspectrreused. Next, we exmine the CF effect, s shown in Figure 6. The results shown in Figure 5 re sed on CF-1; it hs high trnsmittnce, ut reltively lrge overlp in the lue/green nd green/red regions. Alterntively, CF-2 possesses smller crosstlk, ut the trnsmittnce is ~ 25% lower in the lue nd green regions. Agin, we crry out the optimiztions using CF-2 nd plot the Preto front results, which re shown in Figure 7. The trend is similr to tht in Figure 5, ut it hs wider color gmut. Due to the suppressed light lekge, the color gmut reches 93.1% when 30-nm FWHM cklight is employed. If we reduce the FWHM to nm, this limit is incresed to 94.9%. However, when the rel spectrum is employed, the color gmut should e nrrower due to the ove-mentioned Gussin-fitting effect. In comprison with CF-1 (Figure 5), CF-2 suffers 25% in opticl efficiency ut only gins 2.7% in color gmut. Such trdeoff my not e worthwhile, especilly for ttery-powered moile disply devices. We hve found the theoreticl limit of the color gmut of n LCD, s shown in Figures 5 nd 7. It is jointly determined y the light source (centrl wvelength nd FWHM) nd the CFs (crosstlk). Even if we reduce the FWHM of the cklight to 1 nm, the color gmut is still limited y the crosstlk of the lue/green nd green/red CFs (Figure 2). Nrrowing the trnsmission nds (CF-2) could mitigte this issue ut not completely. Menwhile, the opticl efficiency is doi:1038/ls.17.43
Going eyond the limit of n LCD s color gmut 5 24 24 22 22 TLE (Im W 1 ) 18 16 14 12 10 8 FWHM = 50 nm FWHM = 40 nm FWHM = 30 nm FWHM = nm FWHM = 10 nm FWHM = 1 nm TLE (Im W 1 ) 18 16 14 12 10 8 FWHM = 50 nm FWHM = 40 nm FWHM = 30 nm FWHM = nm FWHM = 10 nm FWHM = 1 nm 0.75 0 5 0.90 0.95 0 0.70 0.75 0 5 0.90 0.95 0 Colour gmut (CIE 1931) Colour gmut (CIE 1976) Figure 7 Preto front defined in () CIE 1931 nd () CIE 1976 using CF-2. Anlyser ( ) LC lyer CF rry Polrizer ( ) Free region (HG) Hlf-wve region (TN) Ptterned hlfwve plte FRP Blue LED Inverted prism Light guide plte Quntum ril z y x Figure 8 Schemtic digrm nd working principle of the proposed cklight with n FRP nd ptterned hlf-wve plte. Arevitions: HG, homogeneous lignment; TN, twisted nemtic lignment. sustntilly compromised. This trdeoff my not e worthwhile. Therefore, s long s the CF technology mkes no disruptive progress, the color gmut of n LCD will e limited, even if the cklight source exhiits lser-like nrrow emission spectr. RESULTS AND DISCUSSION New device configurtion From the ove discussions, nrrowing the emission spectrum of QD or OIP would eventully rech the theoreticl limit, on the sis of the predictions presented in Figures 5 nd 7. To overcome the rrier resulting from CFs, we propose new cklight configurtion, which is depicted in Figure 8. The key components re FRP nd ptterned hlf-wve plte. They work together to suppress unwnted light lekge cused y CFs. From Figure 8, high power lue LED is used to excite the quntum ril, consisting of red QD nd green OIP-polymer composite film. The light guide plte, together with n inverted prism film, forms directionl cklight 23 tht is comined with front diffuser (not shown here). Such system offers severl dvntges, including wide viewing ngle, negligile color shift nd unnoticele gmm shift 24. Our FRP nd ptterned hlf-wve plte cn e lminted on the top surfce of the inverted prism film. The design principles nd working mechnisms of n FRP hve een reported in Ref. 25. Briefly speking, n FRP is multi-lyer doi:1038/ls.17.43
Going eyond the limit of n LCD s color gmut 6 100 80 0 10 30 100 80 0 10 30 60 40 60 40 0 400 450 550 650 700 0 400 450 550 650 700 Figure 9 Angulr-dependent trnsmission spectr of () x-polrized incident light nd () y-polrized incident light. A A P P c P A d P A 50 μm 50 μm Figure 10 Working principle of ptterned hlf-wve plte for () x-polrized incident light nd () y-polrized incident light. Cptured polrized opticl microscope imges under (c) prllel polrizers nd (d) crossed polrizers. Scle r = 50 μm. Arevitions: A, nlyzer; P, polrizer. structure with lterntive refrctive indices (n 1 nd n 2 ), s shown in Supplementry Fig. S1. Due to the constructive/destructive interferences, it functions s ndpss filter for oth polriztion directions (tht is, x-polrized nd y-polrized). For x-polrized incident light (Figure 9), only short-wvelength light (lue) cn pss, wheres the rest is reflected. For y-polrized incident light (Figure 9), it is reversed: only green nd red lights cn pss through n FRP, wheres lue light is reflected. As result, fter the FRP, the polriztion stte of lue light is orthogonl to tht of green/red lights (Figure 8). The detiled design procedures cn e found in Supplementry Informtion. In generl, our FRP is multi-lyer film with totl thickness of 23.6 μm. We lso studied the ngulr-dependent trnsmission of our FRP, nd the results re shown in Figure 9 nd 9. Within ± 30 incidence, our FRP could still effectively seprte the lue nd green/red regions. For commercil directionl cklight (with FWHM ~, tht is, ± 10 ), the opticl power is mostly confined within ± (Ref. 23). Therefore, our FRP works quite well, regrdless of the smll nd shift. The ove FRP is ptterned phse retrdtion film, which is divided into hlf-wve region nd free region. For light trversing through the λ/2 region, the polriztion is rotted y 90, for exmple, x-polriztion turns into y-polriztion or vice vers. If the light psses doi:1038/ls.17.43
Going eyond the limit of n LCD s color gmut 7 Blue CF 0 5 0 5 0 400 450 550 650 700 750 400 450 550 650 700 750 Blue CF 0 5 0 5 400 450 550 650 700 750 0 400 450 550 650 700 750 Figure 11 Output SPD of lue primry color for () conventionl cklight nd () the newly proposed cklight. through the free region, then the polriztion remins unchnged. To chieve this gol, we employed ptterned LC phse retrdtion film 26 28.Intheλ/2 region, the LC hs 90 twist, wheres in the free region, it hs homogeneous lignment,29. The detiled structure nd working principle re illustrted in Figure 10 nd 10. With the help of photolignment technology, we hve successfully fricted ptterned hlf-wve plte (see Supplementry Informtion for detils). The cptured polrized opticl microscope imges under prllel nd crossed polrizers re shown in Figure 10c nd 10d, respectively. As expected, the ptterned hlf-wve plte performs quite well. For the system configurtion shown in Figure 8, this ptterned hlfwve plte should register with the RGB su-pixels. Here the hlf-wve region is right elow the green/red su-pixels, nd the free region corresponds to the lue su-pixel. After the FRP, the lue light is polrized long the x-xis, wheres the green nd red lights re polrized long the y-xis. After pssing through the ptterned hlfwve plte, no chnge occurs to the light in the free region (corresponding to lue su-pixel), which indictes tht lue light is still polrized long the x-xis. Alterntively, in the hlf-wve region (corresponding to green nd red su-pixels), the polriztion of outgoing light chnges 90, turning the light into y-polrized lue light nd x-polrized green/red lights. Therefore, RGB lights could trverse through the front liner polrizer (with trnsmission xis long x-xis) nd enter the corresponding RGB su-pixels. Plese note tht only lue light cn enter the lue su-pixels nd green/red lights re sored y the liner polrizer due to mismtched polriztion. A similr sitution occurs with the green nd red su-pixels. Therefore, no crosstlk exists etween the lue nd green/red regions, leding to much wider color gmut. Figure 11 illustrtes this effect. Let us ssume the LCD cklight spectrum is shown in Figure 1. When the input white light (lck line in Figure 11) psses through lue CF (lue line), there is noticele light lekge in the green region ( 550 nm), which in turn deteriortes the color purity. However, for our new structure with ptterned hlf-wve plte, lue light nd green/red lights re decoupled in terms of polriztion direction. As discussed ove, only lue light cn enter the lue CF; the crosstlk in the green region is lmost completely eliminted (Figure 11). Similrly, more sturted green nd red primry colors re relized. Thus, the color gmut is enhnced from 82.8% to 89.0%. Going eyond the limit of n LCD s colorgmut With the white-light source shown in Figure 1, we cn otin 89%. To further improve this percentge, we need to crefully choose the RGB centrl wvelengths. In the experiment, we tried multiple comintions of lue LEDs, green OIP nd red QDs to generte the desired white light. For detiled mteril synthesis processes, reders cn refer to Section 3 of the Supplementry Informtion. If we use Gussin-fitted spectr, then it is firly esy doi:1038/ls.17.43
Going eyond the limit of n LCD s color gmut 8 R G B 400 450 550 650 700 750 0.9 0.7 y 5 540 New LCD 700 380 0.7 x c 510 v 5 530 540 550 450 570 440 430 New LCD 0 4 u 590 610 630 640 650 Figure 12 () Mesured spectr for optimized white-light source nd corresponding color gmut in () CIE 1931 color spce nd (c) CIE 1976 color spce. to simulte the color gmut efore serching for suitle light source. However, the ctul emission spectrum often contins long til, nd it is chllenging to ccurtely predict the finl color gmut. After dozens of experiments, we hve found n optiml comintion, s plotted in Figure 12. The RGB color coordintes overlp quite well with the stndrd, s Figure 12 nd 12c shown. The corresponding color gmut is 95.8% in CIE 1931 nd 97.3% in CIE 1976. Here high-efficiency CF (CF-1 in Figure 6) is employed. Compred to the theoreticl limit (93.5% for 1-nm FWHM) shown in Figure 5, we hve successfully overcome this limittion for LCDs. To the est of our knowledge, this percentge is the highest recorded color gmut of n LCD. Furthermore, this result is relized using commercil high-efficiency CFs, intended for TV pplictions, nd the otined luminous efficcy is 74.5 lm W 1. Discussion Our design seeks to rek the color gmut limit of n LCD using QD/OIP-enhnced cklight. Next, we wnt to extend this pproch to other cklight technologies, such s 2pc-WLED 5.Bypckging QD/OIP into n LED chip, oth lifetime nd quntum yield would e scrificed ecuse of the high-junction temperture 30.Therefore,2pc- WLED still holds dvntges in stility, long lifetime, low cost nd simple opticl configurtion. The mjor drwck is reltively rod green nd red spectr, leding to 70% 80%, depending on the CF employed. When incorporting 2pc-WLED (β-silon:eu 2+ s green phosphor nd K 2 SiF 6 :Mn 4+ s red phosphor) into our design, the color gmut cn e enhnced from 80% to 89% using highefficiency CF-1 CFs, s Figure 13 depicts. If CF-2 is employed, the color gmut is oosted to 91.5%, except tht the light efficiency is decresed y ~ 25%. This enles 2pc-WLED to compete doi:1038/ls.17.43
Going eyond the limit of n LCD s color gmut 9 0.9 0.7 5 540 Conventionl LCD New LCD 510 5 530 540 550 570 590 610 630 640 680 y 700 v 380 x 0.7 0 450 440 430 4 u Conventionl LCD New LCD Figure 13 Color gmut defined in () CIE 1931 nd () CIE 1976 using 2pc-WLED cklight. (CF-1 is used s CF). directly with the stte-of-the-rt Cd-sed QD technology in terms of color performnce. CONCLUSION We hve investigted the theoreticl limit of the color gmut of n LCD with respect to two vriles, the light source nd CF. Through our nlyses, we find tht 93.5% is the limit, even if the light source exhiits n extremely nrrow linewidth (FWHM ~ 1 nm). Redesigning the CFs with nrrower trnsmission nds helped to enlrge the color gmut y ~ 3%, ut the opticl efficiency is reduced y 25%. To overcome these shortcomings, we propose new cklight configurtion with n FRP nd ptterned hlf-wve plte. The design decouples the polriztion of lue light nd green/red lights, which effectively suppresses the crosstlk etween these color nds, thus expnding the color gmut. In the experiment, we fricted n LC cell with ptterned hlf-wve plte using the photolignment method. Additionlly, to mtch the RGB primry colors to the Rec. stndrd, we prepred white-light source using lue LED to pump green perovskite nd red QDs nd demonstrted n exceedingly high-color gmut (95.8% in CIE 1931 color spce nd 97.3% in CIE 1976 color spce) with commercil, highefficiency CFs. The results re eyond the color gmut limit tht conventionl LCD cn chieve. Our design works eqully well for other cklight sources, such s 2pc-WLED. CONFLICT OF INTEREST The uthors declre no conflict of interest. ACKNOWLEDGEMENTS The uthors thnk Gunjun Tn for helpful discussions nd AFOSR for prtil finncil support under contrct No. FA9550-14-1-0279. 1 Schdt M. Milestone in the history of field-effect liquid crystl displys nd mterils. Jpn J Appl Phys 09; 48: 03B001. 2 Koyshi S, Mikoshi S, Lim S. LCD Bcklights. New York: John Wiley & Sons; 09. 3 Luo ZY, Chen Y, Wu ST. Wide color gmut LCD with quntum dot cklight. Opt Express 13; 21: 26269 26284. 4 Anndn M. Progress of LED cklights for LCDs. J Soc Inf Disp 08; 16: 287 310. 5 Xie RJ, Hiroski N, Tked T. Wide color gmut cklight for liquid crystl displys using three-nd phosphor-converted white light-emitting diodes. Appl Phys Express 09; 2: 022401. 6 Wng L, Wng XJ, Kohsei T, Yoshimur KI, Izumi M et l. Highly efficient nrrow-nd green nd red phosphors enling wider color-gmut LED cklight for more rillint displys. Opt Express 15; 23: 28707 28717. 7 Jng E, Jun S, Jng H, Lim J, Kim B et l. White-light-emitting diodes with quntum dot color converters for disply cklights. Adv Mter 10; 22: 3076 3080. 8 Steckel JS, Ho J, Hmilton C, Xi JQ, Breen C et l. Quntum dots: the ultimte downconversion mteril for LCD displys. J Soc Inf Disp 15; 23: 294 305. 9 Wng YN, He J, Chen H, Chen JS, Zhu RD et l. Ultrstle, highly luminescent orgnic-inorgnic perovskite-polymer composite films. Adv Mter 16; 28: 10710 10717. 10 Zhou QC, Bi ZL, Lu WG, Wng YT, Zou BS et l. In situ friction of hlide perovskite nnocrystl-emedded polymer composite films with enhnced photoluminescence for disply cklights. Adv Mter 16; 28: 9163 9168. 11 Zhu RD, Luo ZY, Chen HW, Dong YJ, Wu ST. Relizing color gmut with quntum dot displys. Opt Express 15; 23: 23680 23693. 12 Chen J, Gensler S, Hrtlove J, Yurek J, Lee E et l. 14.3: Quntum dots: optimizing LCD systems to chieve color performnce. SID Symp Dig Tech Pp 15; 46: 173 175. 13 ITU-R Recommendtion BT. 709-5. Prmeter vlues for the HDTV stndrds for production nd interntionl progrmme exchnge; 02. 14 Adoe Systems Inc. Adoe RGB (1998) Color Imge Encoding. Adoe Systems Inc.; 05. 15 ITU-R Recommendtion BT. Prmeter vlues for ultr-high definition television systems for production nd interntionl progrmme exchnge; 15. 16 Msok K, Nishid Y, Sugwr M, Nksu E. Design of primries for wide-gmut television colorimetry. IEEE Trns Brodcst 10; 56: 452 457. 17 Msok K, Nishid Y, Sugwr M. Designing disply primries with currently ville light sources for UHDTV wide-gmut system colorimetry. Opt Express 14; 22: 19069 19077. 18 Msok K, Nishid Y. Metric of color-spce coverge for wide-gmut displys. Opt Express 15; 23: 7802 7808. 19 Luo ZY, Xu DM, Wu ST. Emerging quntum-dots-enhnced LCDs. J Disply Technol 14; 10: 526 539. Yng DK, Wu ST. Fundmentls of Liquid Crystl Devices, 2nd edn. New York: John Wiley & Sons; 14. 21 Lee SH, Lee SL, Kim HY. Electro-optic chrcteristics nd switching principle of nemtic liquid crystl cell controlled y fringe-field switching. Appl Phys Lett 1998; 73: 2881 2883. doi:1038/ls.17.43
Going eyond the limit of n LCD s color gmut 10 22 Chen Y, Luo ZY, Peng FL, Wu ST. Fringe-field switching with negtive dielectric nisotropy liquid crystl. J Disply Technol 13; 9: 74 77. 23 Käläntär K. A directionl cklight with nrrow ngulr luminnce distriution for widening the viewing ngle for n LCD with front-surfce light-scttering film. JSoc Inf Disp 12; : 133 142. 24 Go YT, Luo ZY, Zhu RD, Hong Q, Wu ST et l. A high performnce single-domin LCD with wide luminnce distriution. J Disply Technol 15; 11: 315 324. 25 Zhu RD, Tn GJ, Yun JM, Wu ST. Functionl reflective polrizer for ugmented relity nd color vision deficiency. Opt Express 16; 24: 5431 5441. 26 Schdt M, Schmitt K, Kozinkov V, Chigrinov V. Surfce-induced prllel lignment of liquid crystls y linerly polymerized photopolymers. Jpn J Appl Phys 1992; 31: 2155 2164. 27 Schdt M, Seierle H, Schuster A. Opticl ptterning of multi-domin liquid-crystl displys with wide viewing ngles. Nture 1996; 381: 212 215. 28 Chigrinov VG, Kozenkov VM, Kwok HS. Photolignment of Liquid Crystlline Mterils: Physics nd Applictions. Chichester: John Wiley & Sons; 08. 29 Schdt M, Helfrich W. Voltge-dependent opticl ctivity of twisted nemtic liquid crystl. Appl Phys Lett 1971; 18: 127 128. 30 Coe-Sullivn S, Liu WH, Allen P, Steckel JS. Quntum dots for LED downconversion in disply pplictions. ECS J Solid Stte Sci Technol 13; 2: R3026 R3030. This work is licensed under Cretive Commons Attriution- NonCommercil-NoDerivs 4.0 Interntionl License. The imges or other third prty mteril in this rticle re included in the rticle s Cretive Commons license, unless indicted otherwise in the credit line; if the mteril is not included under the Cretive Commons license, users will need to otin permission from the license holder to reproduce the mteril. To view copy of this license, visit http:// cretivecommons.org/licenses/y-nc-nd/4.0/ r The Author(s) 17 Supplementry Informtion for this rticle cn e found on the wesite (http://www.nture.com/ls). doi:1038/ls.17.43