All non-dopant red–green–blue composing white organic
light-emitting diodes
Shi-Jay Yeh
a, Hung-Yang Chen
a, Min-Fei Wu
a, Li-Hsin Chan
a,
Chih-Long Chiang
a, Hsiu-Chih Yeh
a, Chin-Ti Chen
a,*, Jiun-Haw Lee
b,*aInstitute of Chemistry, Academia Sinica, Taipei 11529, Taiwan, ROC
bGraduate Institute of Electro-Optical Engineering and Department of Electrical Engineering, National Taiwan University,
Taipei 10617, Taiwan, ROC
Received 2 August 2005; received in revised form 10 December 2005; accepted 13 December 2005 Available online 11 January 2006
Abstract
All non-dopant white organic light-emitting diodes (WOLEDs) have been realized by using solid state highly fluores-cent red bis(4-(N-(1-naphthyl)phenylamino)phenyl)fumaronitrile (NPAFN) and amorphous bipolar blue light-emitting 2-(4-diphenylamino)phenyl-5-(4-triphenylsilyl)phenyl-1,3,4-oxadiazole (TPAOXD), together with well known green fluo-rophore tris(8-hydroxyquinolinato)aluminum (Alq3). The fabrication of multilayer WOLEDs did not involve the
hard-to-control doping process. Two WOLEDs, Device B and C, different in layer thickness of Alq3, 30 and 15 nm, respectively,
emitted strong electroluminescence (EL) as intense as 25,000 cd/m2. For practical solid state lighting application, EL inten-sity exceeding 1000 cd/m2was achieved at current density of 18–19 mA/cm2or driving voltage of 6.5–8 V and the devices
exhibited external quantum efficiency (gext) of 2.6–2.9% corresponding to power efficiency (gP) of 2.1–2.3 lm/W at the
required brightness. The thickness of Alq3layer is decisive in color quality of non-dopant WOLEDs. The Commission
Internationale de l’Eclairage (CIE) coordinates of fairly white EL of Device B varied only little from (0.34, 0.39) to (0.34, 0.38) at driving voltage between 6 and 14 V. Device B exhibited relatively high color rendering indexes (CRIs) in the range of 74–81, which were essentially voltage-independent. The other WOLED, Device C, showed even better color purity of white EL (CIEx,y= 0.34, 0.31) along with higher CRI of 83 at 8 V, although higher voltage deteriorated the color
quality of WOLED.
Ó 2005 Elsevier B.V. All rights reserved. PACS: 71.20.Rv; 72.80.Le; 73.61.Ph
Keywords: White; Organic light-emitting diode; Non-dopant; Red; Blue
1. Introduction
Organic light-emitting diodes (OLEDs) have rapidly developed over the last decade and have become the highly competitive alternative of full color flat panel display [1–5], which is currently 1566-1199/$ - see front matter Ó 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.orgel.2005.12.004
* Corresponding authors. Tel.: +886 227898542; fax: +886
227831237.
E-mail addresses: [email protected] (C.-T. Chen),
[email protected](J.-H. Lee).
dominated by the liquid-crystal display (LCD) [6]. Besides glamorous displays, an intriguing and potentially energy-saving application for OLEDs are decorative or domestic solid-state lighting (SSL) and the flat panel display backlight [6–10], OLEDs for SSL or back light for flat panel display have not been considered to be realistic until recent discovery of highly efficient organometallic phosp-horophores [11,12]. The introduction of electroph-osphorescence moves the upper limit of the power efficiency (gP) from20 lm/W for electrofluorescent materials to 80 lm/W, which surpasses 15 lm/W for incandescent light bulb and is comparable with that of a fluorescent lamp. With exceptions of low efficient singlet exciplexes or triplet excimers emit-ting broad-band (white) emission, satisfactory white organic light-emitting diodes (WOLEDs) are always constructed on a multilayer device structure with two (green blue/orange red or blue/yellow) to three (red, green, and blue) light-emitting components that can be fluorophores and/or phosphorophores
[12–31]. Among these color components of
WOL-EDs, many blue, most of yellow, and nearly all orange to red fluorophores suffer from a common problem, namely the concentration quenching of fluorescence in solid state. Consequently, the dop-ant light-emitter of guest–host system becomes a universal method for solving the quenching problem of these long-wavelength, orange to red fluoro-phores[32]. Emission quenching due to the long life-time of the triplet state (phosphorescence) causes severe triplet–triplet annihilation of electrolumines-cence (EL) which is significant even at medium cur-rent density greater than 50 mA/cm2for most cases. Doping is also an inevitable solution to all colors of phosphorophores. Therefore, the doping process seems to be a must in the fabrication of WOLEDs based on small molecular materials, either fluoro-phores or phosphorofluoro-phores. However, in practical OLED manufacture, the doping process is not a trivial task to handle, considering the reproducibil-ity of the optimum doping level, which is normally low and less than 1–2% and let along that it should be carefully controlled in a narrow effective range of ±0.5% for a consistent performance of the devices. Recently, we have successfully developed a few bright (electroluminance (L) of 8000–12,400 cd/m2) and efficient (external quantum efficiency, gext, of 2.4–3.6% or gP of 0.9–1.8 lm/W) non-doped red OLEDs based on extraordinary red light-emitting materials [33–36]. NPAFN (Scheme 1) is one of those rare materials that have an unusual property
of aggregation-induced emission (AIE) and it was adopt here as non-dopant red light-emitting mate-rial for WOLEDs [37]. In addition to NPAFN, we have also explored a new type of blue fluorophore, TPAOXD (Scheme 1), which is bipolar in nature but is an authentic amorphous molecular material [38]. High-performance (maximum brightness 20,000 cd/m2
and maximum external quantum efficiency 2.4%) blue (1931 Commission Interna-tionale de L’Eclairage, CIEx,y of x = 0.16 and y = 0.18) OLEDs containing TPAOXD as the non-dopant blue emitter was found to be relatively stable with little decay of external quantum effi-ciency at 2.3% from low to high current density (10–500 mA/cm2) [39]. In this work, three devices A, B, and C, were fabricated to test the idea of
N O N O N O Al CN NC N N N N O N Si N N N N BCP N N N N N N TPAOXD NPB TPBI Al LiF TPBI 4.1 6.2 2.7 6.4~6.5 2.9~3.0 BCP Alq3 5.95 3.25 5.8 3.6 TPAOXD 5.4 2.3 5.9 3.0 NPAFN NPB ITO 4.7~5.1 Device B: Device C: 30 2 10 40 30 30 0.5 15 2 10 40 30 30 0.5 Device A:10 40 30 5 30 30 0.5 Layers thickness (nm) Alq3 NPAFN Device E:40 0 40 0 10 30 0.5 Device D:40 10 40 0 10 30 0.5
Scheme 1. Chemical structures and their relative energy level (HOMO and LUMO) diagram of NPAFN, Alq3, TPAOXD,
NPB, BCP, and TPBI. Layer structures of Device A–E are also schematically shown here with the layer thickness indicated therein.
bright and efficient all non-dopant WOLEDs using NPAFN, Alq3, and TPAOXD for red, green, and blue emitters, respectively (see photoluminescence, PL, spectra inFig. 1).
There was one preceding report about ‘‘nondoped-type’’ WOLED by Tsuji et al. [40].
However, they were utilizing complementary two-color system, blue emission (445 and 480 nm) from NPB and orange emission (558 nm) from DCM (4-(dicyanomethylene)-2-methyl-6-(p-dimethylami-nostyryl)-4H-pyran), in composing white emission. Furthermore, their WOLEDs adopt an ultrathin layer (only 1 nm thickness) of DCM as one of the complementary colors. Such a thin layer is unlikely to make DCM to be a ‘‘nondoped type’’ emitter (more likely to be a blending of DCM and NPB) and hence unlikely the ‘‘nondoped type’’ WOLED. This is because the authentic ‘‘nondoped type’’ DCM (neat DCM) emitter should have a deep red emission at wavelength greater than 620 nm instead of orange 558 nm [32,41]. Furthermore, their WOLED is weakly electroluminescent (maximum luminance1000 cd/m2) due to the dim blue NPB and concentration quenching orange DCM. The EL efficiency of such WOLED is expected to be low, which in fact is not available from the report. 2. Experimental
Device A–E were fabricated by sequential ther-mal vacuum deposition of thin layer of organic materials and LiF–Al as the final cathode electrode on ITO (indium-thin-oxide)-coated glass substrate. The configuration of these devices is depicted in
Scheme 1. The current, voltage, and light intensity
measurements have been described before [39]. All measurements, including the recording of EL spec-tra, were carried out at room temperature under ambient condition without the encapsulation of the devices. Whereas literature known figures were adopt for Alq3, BCP, and TPBI, HOMO energy lev-els of NPB, NPAFN, and TPAOXD were measured with a low-energy photo-electron spectrometer (Riken-Keiki AC-2). The addition of the absorption on-set energy to the HOMO energy levels led to LUMO energy levels of these materials. The absorp-tion on-set energy was estimated from the soluabsorp-tion UV–visible absorption spectra. In CRI calculation, eight test-color samples were chosen followed the definition of DIN 6169 standards. CIE 1976 (L*u*v*) color-difference formula was used to deter-mine the resultant color shift[42]. In order to dem-onstrate the easy control of the white color EL from these all non-dopant devices, we fixed the thickness of NPB, NPAFN, TPAOXD, and TPBI layers at 10, 40, 30 and 30 nm, respectively, in Device A, B, and C, although these can be the adjustable parameters in optimizing performance of the devices. By just
300 400 500 600 700 800 0.0 0.5 1.0 629 nm NPAFN 520 nm Alq3 443 nm TPAOXD PL Intensity (A.U.) Wavelength (nm) 300 400 500 600 700 800 0.0 0.2 0.4 0.6 0.8 1.0 Device C Device B Device A EL Intensity (A.U.) Wavelength (nm) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 X Y CIE coords CRT coords. 10 nm spacing Device A Device B Device C 520 560 600 480
Fig. 1. Normalized photoluminescence spectra of vacuum ther-mal deposited thin films of TPAOXD, Alq3, and NPAFN; EL
spectra (center) and 1931 CIE coordinates (bottom) of Device A–C at driving voltage of 8 V. CRT (cathode ray tube) coordinates mean the standard red–green–blue color coordinates of CRT screen (traditional color television screen).
varying the thickness of Alq3 and BCP layers we were able to tune the color of the devices showing CIE coordinates as close as to (0.33, 0.33), a stan-dard figure for color-balanced WOLED.
3. Results and discussion
Device A is our testing model and is not a color-balanced WOLED since the 5 nm layer of hole-blocking BCP (2,6-dimethyl-4,7-diphenyl-1,10-phe-nanthroline) is simply too much for WOLED. In current case, 5 nm layer of BCP effectively blocks the hole from entering TPAOXD and hence limits the charge recombination and/or exciton diffusion on it. This greatly prevents the blue emission (kmax 460 nm) from TPAOXD in EL spectra (Device A in Fig. 1). Reducing the thickness of BCP layer to half as in Device B turned out to be just fine for white EL of Device B. CIE coordinates of it is (0.35, 0.39), which is a respectful improve-ment of white EL from (0.41, 0.48) of CIE coordi-nates of yellow–orange Device A (Fig. 1). Even higher color purity of white EL has been achieved by adjusting the thickness of Alq3layer. In Device C, the thickness of Alq3 layer was further reduced to half (15 nm) from 30 nm of Device B. Having both thin layers of green light-emitting Alq3 and hole-blocking BCP, Device C showed significant increase of the blue emission from TPAOXD as well as the decrease of green emission from Alq3(Fig. 1). Such changes of relative contribution of emission color rendered virtually white EL of Device C as indicated by the CIE coordinates (0.34, 0.31)
(Fig. 1). In principle, all non-dopant WOLED with
any CIE coordinates as close as to (0.33, 0.33) can be achieved with the variation of the thickness of Alq3and BCP layers.
WOLEDs reached 1000 cd/m2 of EL intensity, the general requirement for SSL application, at a reasonable current density range of 18–19 mA/cm2
(Fig. 2), which corresponded to 6.5–8 V for Devices
B and C. The maximum EL intensity of 25,000 cd/m2 was observed for Device B at 790 mA/cm2 and 15 V. At low current density near 0.1 mA/ cm2, high gPs over 5 lm/W were observed for Device C and slightly lower 4.0–4.5 lm/W were found for Device B, although both devices display a roll-off of gPat elevated current density (Fig. 2). However, high gexts of 2.9–3.3% remained relatively stable with only slight decay to 2.6–2.9% (corresponding to 2.1–2.3 lm/W) when the current density up to 20 mA/cm2(or 1000 cd/m2lighting intensity). Even
at relatively high current density of 100 mA/cm2
(where WOLED showed EL intensity of
5000 cd/m2
, Fig. 2), gexts stayed reasonably high
around 2.4–2.6%. The rather current-stable external quantum efficiency can be partially attributed to the relatively stable blue emitter TPAOXD. Similar sta-bility of gextwas previously known for blue OLEDs based on amorphous TPAOXD [39]. Containing both electron-poor oxadiazole and electron-rich tri-phenylamine moieties, bipolar TPAOXD possesses electro-transporting characteristic, which is crucial for a good performance WOLEDs here. We have fabricated the same WOLED by using XTPS (trans-4-(bis(2,4-dimethylphenyl)amino)-20,30,40,50 -tetraphenylstilbene), another highly efficient (maxi-mum external quantum efficiency up to 4.1%) non-dopant blue light ðkEL
max 462 nmÞ emitter [43]. However, the XTPS-containing WOLEDs per-formed poorly. Unlike bipolar TPAOXD, XTPS is stilbene-type triarylamine and preferentially trans-ports hole instead of electron. This is not appropri-ate for using XTPS as blue light-emitting mappropri-aterial that is insufficient for electron-transporting in cur-rent devices. 0.1 1 10 100 1000 1 10 100 1000 10000 Electroluminance (cd/m 2 )
Current Density (mA/cm2)
Device B Device C 0.1 1 10 100 1000 0.0 1.0 2.0 3.0 4.0 5.0 6.0
Current Density (mA/cm2)
External Quantum Efficiency (%)
Device B Device C 0.0 1.0 2.0 3.0 4.0 5.0 6.0 Power Efficiency (lm/W)
Fig. 2. Electroluminance (L)–current density (I)–efficiency (gext
Device B was unusually stable regarding CIE coordinates and CRIs. The CIE coordinates of Device B were virtually unchanged with CIE coordi-nates x = 0.34–0.35 and y = 0.38–0.39 at the driving voltage range of 6–14 V (Fig. 3). CRIs of Device B were rather high in the range of 74–81 and they also remained reasonably steady in the driving voltage between 6 and 14 V (Fig. 3). However, we noted that Device C was less stable considering either CIE coordinates or CRIs. With increasing driving voltage, CIE coordinates of Device C shifted to smaller figures and CRIs dropped to smaller num-bers from 83 to less than 50 (Fig. 3). This is in sharp contrast to Device B. Furthermore, Device C exhib-ited complicate up and down of CIE coordinates and CRIs when driving voltage varied from 6 to 14 V. Usually, such nonlinear changes imply that there are at least two different processes that are responsible for the color changes of devices. For Device C, blue EL first increased in intensity relative to the red EL (CIE coordinates changed to smaller figures) when the driving voltage was raised from 6 to 12 V (Fig. 3). Significant decrease of blue EL was then observed when driving voltage was further up to 14 V (Device C inFig. 3). It is not so clear at the moment that why Device B is more stable in general than Device C. We can surmise that it is probably due to the instability of blue TPAOXD because the device having thinner Alq3tends to have more portion of blue EL from TPAOXD (see below).
Among several possible causes for color changes, we can assume that the shifting of the charge recom-bination zone and/or exciton diffusion toward cath-ode side of the device under higher driving voltage as one of the possibilities. Such zone shifting and/ or exciton diffusing takes place in both Device B and C but the thick Alq3layer (30 nm) in between NPAFN and TPAODX layers taking up large por-tion of it that alleviate most of the color changes in Device B. The second possible process that changes the color of the device is the potential deg-radation of blue emitter TPAODX, particularly, under high voltage of 14 V (corresponding to very high current density 1000 mA/cm2), which has been identified before [39]. The lowering CRIs of Device C with rising voltage is mainly due to the diminished relative contribution of green EL from Alq3as the charge recombination zone shifting or exciton diffusing toward blue emitter TPAODX. At high voltage of 14 V, the relative intensity of the blue emission from TPAODX became
weaken-200 300 400 500 600 700 800 0 100 200 300 400 500 600 700 800 76 74 81 81 81 CRI Device B EL Intensity Wavelength (nm) 6 V 8 V 10 V 12 V 14 V 6 8 10 12 14 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 Device B X Y Voltage (V) x 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 y 200 300 400 500 600 700 800 0 200 400 600 800 1000 1200 83 75 59 48 62 CRI Device C EL intensity Wavelength (nm) 6 V 8 V 10 V 12 V 14 V 6 8 10 12 14 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 x y Voltage (V) x 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 Device C y
Fig. 3. The voltage dependency of EL spectra and CIE coordi-nates (x, y), and CRIs of Device B and C under different voltage of 6, 8, 10, 12, and 14 V.
ing and the profile of white EL spectrum became less dented, which turned the CRI to a larger figure. Among several vital issues mentioned above, the success of such WOLED depends on the appropri-ate adjustment of the relative EL intensity of red– green–blue three components via the thickness con-trol of individual light-emitting layer. Through the organ–red dye-doped spacial sensing layer in Alq3, it has been demonstrated before by Tang et al. that charge-recombination is confined about 5 nm near the hole-transport interface in the bilayer OLED [40]. At a glance, observing the blue EL from TPAOXD from our devices, particularly Device B and C, was rather unbelievable because of the thick-ness of Alq3layer is way beyond 5 nm and it is 30 and 15 nm, respectively. However, we have recog-nized that Tang et al. also had reported that EL emission zone in the undoped Alq3system was con-siderably larger than the charge-recombination zone due to the exciton diffusion. The migration of the exciton was carefully estimated by Tang et al. to be more than 20 nm in undoped Alq3[42]. In order to get a clear picture of the thickness-dependent EL of all non-dopant devices here, we have fabricated two more devices, Device D and E. These devices are similar to Device A–C but Alq3layer is thicker (40 nm) and we remove BCP hole-blocking layer to simplify the issue (Scheme 1). Particularly, Device E contains no red NPAFN and we see only strong EL from Alq3 accompanying with weak emission side band around 440 nm, which strongly resembles the blue EL from TPAOXD (Fig. 4). The result of Device E indicates that the charge-recombination zone and/or exciton diffusion is beyond the whole 40 nm thick Alq3layer and entering blue TPAOXD layer. The result is somewhat reconfirmed what Tang et al. found earlier, although 40 nm was a
bit beyond the estimated exciton diffusion length (P20 nm) for Alq3. Here, according to the current results, we have a new proposal that the zone of charge-recombination and/or the range of exciton diffusion of non-doped Alq3depend on the HOMO energy level of hole-transport layer adjacent to the Alq3. This is based on the result of Device D, which has a similar structure to Device E except a thin layer (10 nm) of red NPAFN inserted in between NPB and Alq3. In addition to the strong EL from red NPAFN around 615 nm, when compared with Device D, a significant increase of blue EL from TPAOXD relative to the green EL from Alq3 was observed (Fig. 4). NPAFN has HOMO energy level of 5.8 eV, which is 0.4 eV lower than NPB HOMO energy level (Scheme 1). NPAFN has been proven to be an efficient non-dopant red-light emitter and hole-transport material as well in the non-dopant red OLEDs [34]. We assume that the closer of the HOMO energy level of hole-transport material to that of Alq3is, the deeper of the charge recombina-tion zone is inside Alq3. Assuming the same range of exciton diffusion (a function of diffusion coefficient and the lifetime) [40], the closer charge recombina-tion zone to the interface of Alq3and TPAOXD lay-ers will presumably facilitates the exciton taking place in the blue light-emitting TPAOXD layer. 4. Conclusion
In summary, we have studied the EL properties of a series of multilayer OLEDs containing all non-dopant emitters. Devices with the structure configuration of ITO/NPB/NPAFN/Alq3/BCP/ TPAOXD/TPBI/LiF/Al emit color-balanced white EL. The success of the all non-dopant WOLEDs here hinges on the unusual AIE red emitter of NPAFN as well as the amorphous bipolar blue emit-ter TPAOXD. Green light-emitting Alq3also plays a crucial role of white color purity and stability of non-dopant WOLEDs. Under the condition of sufficient amount of blue EL contributed to the color-balanced WOLED, such as Device B and C, we propose that a thick Alq3-containing device can cover most of the charge-recombination zone and reduce the number of exciton migrating into blue TPAOXD. The relatively unstable blue EL from TPAOXD is limited and CIE- or CRI-stability can thus be enhanced. WOLEDs, particularly the thick (30 nm) Alq3-containing Device B, has been demon-strated to show satisfactory performances, including EL intensity, efficiency, CRIs, and voltage-stable
300 400 500 600 700 800 0.0 0.2 0.4 0.6 0.8 1.0 EL Intensity (A.U.) Wavelength (nm) Device D: Device E:
white EL, which are comparable with or better than conventional electrofluorescence-based WOLEDs that are exclusively based on hard-to-control doping process.
Acknowledgements
This work was supported by Academia Sinica and National Science Council. We thank Prof. Yu-Tai Tao for his instruction in the fabrication and measurement of OLEDs.
References
[1] S.M. Kelly, Flat Panel Display: Advanced Organic Materi-als, Royal Society of Chemistry, Cambridge, 2000. [2] B. Johnstone, Technol. Rev. 104 (September) (2001) 80. [3] O. Gelson, Opt. Laser Eur. 107 (June) (2003) 33. [4] W.E. Howard, Sci. Am. 290 (February) (2004) 76. [5] J.R. Sheats, J. Mater. Res. 19 (2004) 1974. [6] B. Yang, Inf. Disp. 21 (February) (2005) 22.
[7] A. Bergh, D. Craford, A. Duggal, R. Haitz, Phys. Today 54 (2001) 42.
[8] T. Whitaker, Opt. Laser Eur. 126 (March) (2004) 5. [9] B.W. D’Andrate, R.J. Holmes, S.R. Forrest, Adv. Mater. 16
(2004) 624.
[10] T. Whitaker, Comp. Semicond. 10 (October) (2004) 20. [11] S.R. Forrest, Org. Electron. 4 (2003) 45.
[12] B.W. D’Andrade, S.R. Forrest, Adv. Mater. 16 (2004) 1585. [13] J. Kido, M. Kimura, K. Nagai, Science 267 (1995) 1332. [14] R.S. Deshpande, V. Bulovic, S.R. Forrest, Appl. Phys. Lett.
75 (1999) 888.
[15] Z.Y. Xie, J.S. Huang, C.N. Li, S.Y. Liu, Y. Wang, Y.Q. Li, J.C. Shen, Appl. Phys. Lett. 74 (1999) 641.
[16] M. Strukelj, R.H. Jordan, A. Dodabalapur, J. Am. Chem. Soc. 118 (1996) 1213.
[17] R.H. Jordan, A. Dodabalapur, M. Strukelj, T.M. Miller, Appl. Phys. Lett. 68 (1996) 1192.
[18] F. Steuber, J. Staudigel, M. Stossel, J. Simmerer, A. Winnacker, H. Spreitzer, F. Weisso¨rtel, J. Salbeck, Adv. Mater. 12 (2000) 130.
[19] S. Liu, J. Huang, Z. Xie, Y. Wang, B. Chen, Thin Solid Films 363 (2000) 294.
[20] C.W. Ko, Y.T. Tao, Appl. Phys. Lett. 79 (2001) 4234. [21] K.O. Cheon, J. Shinar, Appl. Phys. Lett. 81 (2002) 1738. [22] Y.S. Hunag, J.H. Jou, W.K. Weng, J.M. Liu, Appl. Phys.
Lett. 80 (2002) 2782.
[23] B.W. D’Andrade, M.E. Thompson, S.R. Forrest, Adv. Mater. 14 (2002) 147.
[24] C.H. Chuen, Y.T. Tao, Appl. Phys. Lett. 81 (2002) 4499. [25] G. Li, J. Shinar, Appl. Phys. Lett. 83 (2003) 5359. [26] X.H. Zhang, M.W. Liu, O.Y. Wong, C.S. Lee, H.L. Kwong,
S.T. Lee, S.K. Wu, Chem. Phys. Lett. 369 (2003) 478. [27] S. Tokito, T. Iijima, T. Tsuzuki, F. Sato, Appl. Phys. Lett. 83
(2003) 2459.
[28] T.-H. Liu, Y.-S. Wu, M.-T. Lee, H.-H. Chen, C.-H. Liao, C.H. Chen, Appl. Phys. Lett. 85 (2004) 4304.
[29] G. Cheng, Y. Zhao, Y. Zhang, S. Liu, F. He, H. Zhang, Y. Ma, Appl. Phys. Lett. 84 (2004) 4457.
[30] D. Qin, Y. Tao, Appl. Phys. Lett. 86 (2005) 113507. [31] Y. Xu, J. Peng, Y. Mo, Q. Hou, Y. Cao, Appl. Phys. Lett. 86
(2005) 163502.
[32] C.-T. Chen, Chem. Mater. 16 (2004) 4389.
[33] W.-C. Wu, H.-C. Yeh, L.-H. Chan, C.-T. Chen, Adv. Mater. 14 (2002) 1072.
[34] H.-C. Yeh, S.-J. Yeh, C.-T. Chen, Chem. Commun. (2003) 2632.
[35] H.-C. Yeh, L.-H. Chan, W.-C. Wu, C.-T. Chen, J. Mater. Chem. 14 (2004) 1293.
[36] C.-L. Chiang, M.-F. Wu, D.-C. Dai, Y.-S. Wen, J.-K. Wang, C.-T. Chen, Adv. Funct. Mater. 15 (2005) 231.
[37] J. Chen, B. Xu, X. Ouyang, B. Tang, Y. Cao, J. Phys. Chem. A 108 (2004) 7522.
[38] L.-H. Chan, H.-C. Yeh, C.-T. Chen, Adv. Mater. 13 (2001) 1637.
[39] L.-H. Chan, R.-H. Lee, C.-F. Hsieh, H.-C. Yeh, C.-T. Chen, J. Am. Chem. Soc. 124 (2002) 6469.
[40] T. Tsuji, S. Naka, H. Okada, H. Onnagawa, J. Appl. Phys. 81 (2002) 3329.
[41] C.W. Tang, S.A. VanSlyke, C.H. Chen, J. Appl. Phys. 65 (1989) 3610.
[42] G. Wyszecki, W.S. Stiles, Color Science, second ed., Wiley, New York, 1982, CIE 1931 is commonly used to describe the color of an light-emitting device. However, a uniform color space, such as CIE 1976 Luv or Lab (L*u*v*) space, is needed
for calculating CRI and color difference.
[43] C.-T. Chen, C.-L. Chiang, Y.-C. Lin, L.-H. Chan, C.-H. Huang, Z.-W. Tsai, C.-T. Chen, Org. Lett. 5 (2003) 1261.
