Open Access
Add to BibSonomyAbstract
We report the design and fabrication of a single-cell organic light-emitting diode (OLED) whose emission color shifts inline with the broad black-body radiation and the tunability is controlled by applied voltage. Correlated color temperature (CCT) shift from 2500K to 15000K was intrinsically aroused from carrier mobility change in organic layers. The multi-layer OLED design was optimized by calculation of carrier transport and recombination zone location. Emphasis was made on both color and efficiency. Devices reported here can reach consistently high color rendering index (CRI) 94-97 in broad color range from orange to blue, revealing potential in human centric lighting applications.
© 2016 Optical Society of America
1. Introduction
Presently organic light-emitting diodes (OLEDs) are designed to produce a single color in a single cell at a given unit of luminance. White OLEDs with multi-emission layers can have ideal Commission Internationale de L' Éclairage (CIE) coordinates (0.33, 0.33) and high color rendering index (CRI > 90) [1–3]. However, the daylight color temperature in fact changes from dawn to dusk and our circadian cycle follows this change closely. Recent report on lighting for health and well-being by European Union has shown clearly that light has strong effect on mood, attention and alertness through the light non-visual effect on human body [4]. Thus, attention is now being paid on dynamic change of spectral content, intensity, luminance, duration and timing over the circadian cycle of 24 hours. To mimic this circadian cycle induced mood, broad color tunability is a prerequisite and it can be applied in advanced mood lighting applications (i.e. aircraft cabins). The key is to consistently maintain high CRI when changing the color. Previous tunable OLED includes a carrier-modulating layer among monochromatic multiple emissive layers to control emission by applied voltage [5]. The modulating layer often causes inconsistent and unbalanced carrier transport at different voltages thus reduces color quality. In this paper, we demonstrate a single cell OLED with broad color tunability purely derived from carrier transport properties of functional organic layers. Its voltage dependent emission mimics solar color temperature and closely follows the black-body radiation curve. By carefully designed RGB emissive layers in a single cell, the device has achieved a curved transition in CIE coordinates from orange (0.48, 0.43) to blue (0.26, 0.27), correlated color temperature (CCT) 2500-15000 K, with color rendering index (CRI) 94-97 maintained across broad voltage range 4-10 V.
2. Experimental
Indium tin oxide (ITO) substrates were cleaned by detergent and de-ionized water in ultrasonic bath, and organic materials were deposited by thermal evaporation (400-700K) in high vacuum about 10−6 Torr. 60 nm 4,4,4″-tris(n-(2-naphthyl)-n-phenyl-amino)-triphenylamine (2-TNATA) and 20 nm N’-diphenyl-N,N’-bis(1-naphthyl)(1,1’-biphenyl)-4,4’diamine (NPB) were deposited on top of ITO anode as hole injection layer (HIL) and hole transport layer (HTL), then the emissive layer followed by 20 nm tris(8-hydroxyquinolinato)aluminium (Alq3) as electron transport layer (ETL) and 1 nm lithium fluoride (LiF) as electron injection layer (EIL) with aluminum cathode. For emissive layers (EML), 2-methyl-9,10-di(2-naphthyl)anthracene (MADN) and Alq3 were used as host and N,N’-(4,4’-(1E,1’E)-2,2’-(1,4-phenylene) bis(ethene-2,1-diyl)bis(4,1-phenylene))-bis(2-ethyl-6-methyl-N-phenylaniline) (BUBD-1), 4,4'-bis(9-ethyl-3-carbazovinylene)-1, 1'-biphenyl (BCzVBi), 2,3,6,7- tetrahydro-1,1,7,7, -tetramethyl- 1H,5H,11H-10- (2-benzothiazolyl) quinolizino-[9,9a,1gh] coumarin (C545T), 4-(dicyanomethylene)- 2-t-butyl-6- (1,1,7,7-tetramethyljulolidyl-9-enyl)-4H-pyran (DCJTB) were dopants. By aligning host-guest HOMO and LUMO, we co-deposited blue emissive layer MADN:BUBD-1 or MAND:BCzVBi, green emissive layer Alq3:C545T and red emissive layer Alq3:DCJTB. Each host-guest ratio was controlled and tuned to maximize the color performance. After encapsulation, the device was characterized on its efficiency from current-density-voltage-luminance (J-V-L) as well as spectrum based CIE coordinates, CCT and CRI, all measured by PR650 spectrophotometer with a computer controlled dc source. Instead of a continuous voltage scan, our measurements were taken at interval of 0.5 V for 30 s, followed by another 30 s of open voltage before the next voltage increment was applied. This has prevented hysteresis therefore the measurement is similar to steady state measurement.
3. OLED structure design
To achieve at least good quantum efficiency OLED devices requires that photons are emitted from the carrier recombination zone where holes and electrons formed excitons. The location of recombination zone is governed by carrier transport and should be preferably within emissive layers. At higher applied voltage, all carriers move faster but the amount of increase in drift velocity varies for holes and electrons and is material dependent. Thus potential shift of recombination zone within the device occurs when the applied voltage changes. We can take advantage of this property to produce tunable emission which can be achieved once part or whole of recombination zone shifts from one emissive layer into another. The location of recombination zone was derived from a simplified charge transport model. The electron/hole mobility at certain electric field was defined in Poole-Frenkel effect [6].
where β is Poole-Frenkel constant, F is electric field and μ0 is lower field carrier mobility. Both β and μ0 are material dependent. The drift velocity in a homogeneous electric field is where V is applied voltage and d is organic layer thickness. Multiplying each velocity and corresponding layer thickness, total carrier transit characteristic across the device can be calculated.where t is layer transport time. There is no EIL term in Eq. (4) as LiF layer is only 1 nm thick. The interface of EML and ETL was used as a reference to indicate the location of recombination zone, s. where α is the distance percentage used to compare recombination zone shift in different layer structures. Repeating the calculation across voltage range from 4 to 10 V, we obtained a voltage dependent location of recombination zone α(V) for each device structure, presented in percentage of EML.There are several limitations in this estimation. First, for the ease of calculation, we assume the electric field inside organic layers is homogeneous and equal to applied voltage divided by total thickness. In multi-layer structures, however, the internal potential distribution is often not linear because materials have different resistivity and band bending can occur at layer interfaces, both of which contribute to the uncertainty of the internal field distribution [7,8]. Second, mobility of each material is a function of applied voltage; but mobility in OLED has been proven different from standard mobility measurements (i.e. time-of-flight, space-charge-limited current), whose results are usually highly thickness dependent as well [9,10]. Third, we ignore any disorder in carrier transport, which can be critical in some of the organic semiconductors [11]. Instead, we calculate an average of recombination location, which typically spans 5-15 nm in width inside EML [12]. Finally, the effect of dopant is neglected as they are not more than 5% weight in hosts. Although with those limitations, our calculation is still valid. Mobility data of 2-TNATA, NPB, MADN and Alq3 were obtained from literatures [9, 10, 13, 14] and fed into Eq. (1) to derive the mobility values within our measured voltage range (Table 1). The mobility values obtained across the voltage range would overcome any systematic error to give a correct indication of recombination zone shift.
Table 1. Mobility values of 2-TNATA, NPB, MADN and Alq3 at voltage 4, 7, 10 V
Our tunable OLED design was optimized layer by layer based on previously reported structure ITO / 2-TNATA(HIL) / NPB(HTL) / MADN(EML) / Alq3(EML) / Alq3(ETL) / LiF(EIL) / Al (Fig. 1) [15]. The optimized thickness of each layer was found by single variable method, which varies each layer thickness and keeps other layers unchanged. All results are plotted as α(V) and preferably should increase gradually from EML/ETL interface (0%) to EML/HTL interface (100%) across voltage 4 to 10 V to maximize color tunability. Figure 2 shows the optimization process for 2-TNATA, Alq3 and MADN, and the best design was found to have 60 nm 2-TNATA, 20 nm NPB, 40 nm MADN and 20 nm Alq3.
Fig. 1 Device structure and energy diagram of the color tunable OLED, with 2 options for blue dopant.
Fig. 2 HIL/HTL (a), ETL (b) and EML (c) layer thickness optimization according to calculated recombination zone location across applied voltage 4 to 10 V, presented in percentage between EML/HTL and EML/ETL interfaces.
The calculation reveals that tunable color emission can be achieved purely by matching layer thicknesses with the intrinsic mobility change without any need for carrier blocking or modulating layers. We noticed that the curve α(V) has different gradient characteristics in EML and ETL, which may be due to the bipolar mobility of MADN (the EML) and electron favorable charge transport of Alq3 (the ETL). The logarithmic-like shape of α(V) in EML also gave us hints for configurations of RGB emissive layers – a thicker red emissive layer for faster shift at low voltage and a thinner blue emissive layer for slower shift at high voltage. In addition, we have to consider CIE color matching functions where blue is inferior to red. Meanwhile we also need to take note the difference in dopant efficiency Alq3:3%C545T (14.6 cd/A at 20mA/cm2) > MADN:3%BUBD-1 (13.2 cd/A at 20mA/cm2) > MADN:1%DCJTB (4.7 cd/A at 20 mA/cm2) > Alq3:1%DCJTB (3.5 cd/A at 20 mA/cm2) > MADN:5%BCzVBi (2.5 cd/A at 20 mA/cm2), all of which were calibrated separately in single EML OLEDs. As BUBD-1 and BCzVBi emit sky blue at CIE (0.16, 0.30) and deep blue at CIE (0.16, 0.15) respectively, our design criteria focused on efficiency with BUBD-1 (device A) and color with BCzVBi (device B).
- A. ITO / 2-TNATA(60nm) / NPB(20nm) / MADN:3%BUBD-1(15nm) / Alq3:3%C545T(5nm) / MADN:20%Alq3:1%DCJTB(20nm) / Alq3(20m) / LiF(1nm) / Al(80nm)
- B. ITO / 2-TNATA(60nm) / NPB(20nm) / MADN:5%BCzVBi(20nm) / Alq3:3%C545T(10nm) / Alq3:1%DCJTB(10nm) / Alq3(20nm) / LiF(1nm) / Al(80nm)
In both devices we doped C545T in Alq3 as green emissive layer, whose energy level can confine more holes in blue emissive layer, allowing sufficient thickness for red emissive layer. In device A, we used a co-host system MADN:20%Alq3:1%DCJTB to minimize current induced quenching in DCJTB with a balanced charge transport and enhance the current efficiency (6.8 cd/A at 20 mA/cm2) to outperform either of the two constituent hosts [16–18]. The added 20% Alq3 also promotes red emission at CIE (0.60, 0.40), with luminance intensity close to Alq3:1%DCJTB at CIE (0.60, 0.40) and deeper than MADN:1%DCJTB at CIE (0.55, 0.43). The disadvantage of this device is that greenish blue emission from BUBD-1 limits the thickness of green emissive layer to 5 nm, making its color change similar to a bi-emissive-layer device with linear recorded CIE coordinates. For device B the deep blue dopant BCzVBi allows us to use 10 nm thick green emissive layer, which is essential for a non-linear shift similar to that of black-body radiation. It is noteworthy that during calibration BCzVBi increases charge mobility significantly when doped in MADN, especially in its 5% optimized concentration. This effect becomes more pronounced as the blue emissive layer must be made 5 nm thicker to compensate its inefficient emission. To balance the recombination zone location, we doped DCJTB in Alq3 instead of MADN in device B to produce faster electron transfer into the blue emissive layer. Alq3 has comparable electron mobility to MADN at low electric field but can be an order of magnitude faster at high electric field. The host change should benefit blue emission at high voltage while red emission maintains dominant at low voltage.
4. Results and discussion
Performance from both devices matches our design objectives. In terms of current efficiency and power efficiency, device A (9.9 cd/A at 20mA/cm2, 4.1 lm/W at 1000 cd/m2) is more superior than device B (2.4 cd/A at 20mA/cm2, 0.9 lm/W at 1000 cd/m2), which suffers from current induced quenching in Alq3:DCJTB (Fig. 3). The quenching leads to double in current and half in luminance compared to device A. For color performance, the normalized red spectra from both devices show significant voltage dependent color change, as blue and green emission increase (Fig. 4). Both devices have a broad color tunable range in CIE coordinates and emit orange at 4V, however, the increasing voltage pushes device A towards sky-blue i.e. BUBD-1 emission dominates while device B stays within the black-body radiation curve (Fig. 5). Additional color measurements at applied voltage 4-10 V show CCT range 3000-10000 K and 2500-15000 K, and CRI 75-78 and 94-97, for device A and B, respectively. Those color changes were also verified with our recombination zone calculation, in the form of normalized voltage dependent displacement in CIE coordinates (Fig. 6). Both devices can follow the simulated curve.
Fig. 3 JVL characteristics (a) and current/power efficiency (b) of device A (solid and open circle) and device B (solid and open triangle).
Fig. 4 Voltage dependent spectra of device A (a) and B (b) measured at 4-10 V on a 1 V interval. All spectra were normalized to the red emission peak at 608 nm. Blue and green emission increases with applied voltage.
Fig. 5 Color transitions of device A (solid circle) and B (solid triangle) plotted in CIE 1931 coordinates. The applied voltage is at 4-10 V, delivering color change from orange to blue.
Fig. 6 Calculated recombination zone location (solid black) from 2-TNATA(60nm) / NPB(20nm) / MADN(40nm) / Alq3(20nm) and measured color change of device A (solid circle) and B (solid triangle). The color change is normalized by displacement in CIE coordinates.
We took further investigation in device B to reveal how it managed to follow the solar spectra over a broad color range. First we used color matching functions x ̅,y ̅,z ̅ to calculate CIE XYZ coordinates of black-body radiation which is the Planckian locus Eq. (7),8 [19]. These coordinates were then converted into sRGB tristimulus values by CIE XYZ to sRGB transformation matrix [M]−1 Eq. (9) [20]. The resulting RGB curves were plotted versus CCT in 2500-11000 K (Fig. 7(a)).
Fig. 7 CCT dependent sRGB tristimulus values of black-body radiation curve converted from CIE coordinates (a). Voltage dependent RGB percentage extracted from multiple Gaussian peak fitting (b). (solid triangle for red, solid circle for green, solid square for blue)
Second we extracted similar forms of RGB values by fitting RGB Gaussian peaks to device spectra (Fig. 7(b)). By comparing these two figures in Fig. 7, we found that the our green is about 10% lower while red is 10% higher than standard sRGB values, which is due to our deeper green (0.32, 0.65) and lighter red (0.60, 0.40) than sRGB green (0.30, 0.60) and red (0.64, 0.33). Thus the deeper green basically reinforce the lighter red. Apart from that, the curves have exactly the same shape which confirms the link between CCT and applied voltage. We made extra emphasis on green reinforcement, which is key to the non-linear color shift that allows the color shift to mimic the solar spectrum. Our device was perfectly implemented its slow descent to 19% following an initial rise from 16% to 20%, which produces an accurate black-body radiation path in CIE coordinates. As a result, device B maintains a high CRI 94-97, making it nearly as good as the ideal incandescent light bulb (CRI 100) in color reproduction. Our design provides a potential alternative for applications requiring near-perfect color rendition.
5. Conclusion
In this paper, we show that it is possible to design a white light OLED whose emission color shift can mimic black-body radiation from orange to blue through varying the applied voltage. Accurate color emission with consistently high CRI in full tunable range was achieved in device B. The critical part of the design is to have a blue emissive layer that emits in a deep blue region thus allowing broader blue emission intensity for the tuning. Careful design optimizing layer thicknesses, material matching for both mobility and current induced quenching also need to be taken into consideration. Further work is being done to improve the emission efficiency of device B to bring it closer to solid state lighting that can provide circadian mood control. For example the efficiency of device B can be further enhanced by improving carrier transport layers. For HTL, BCzVBi can be doped in NPB to achieve better Forster energy transfer [21]. For ETL, Alq3, can be replaced by Bphen or it can be added as co-host to increase eletron mobility [22,23].
Acknowledgments
This work is being supported by Innovation and Technology Commission project, UIM/216.
References and links
1. G. Zhou, Q. Wang, C. L. Ho, W. Y. Wong, D. Ma, and L. Wang, “Duplicating “sunlight” from simple WOLEDs for lighting applications,” Chem. Commun. (Camb.) 24(24), 3574–3576 (2009). [CrossRef] [PubMed]
2. C. H. Chang, K. C. Tien, C. C. Chen, M. S. Lin, H. C. Cheng, S. H. Liu, C. C. Wu, J. Y. Hung, Y. C. Chiu, and Y. Chi, “Efficient phosphorescent white OLEDs with high color rendering capability,” Org. Electron. 11(3), 412–418 (2010). [CrossRef]
3. J. H. Jou, S. M. Shen, C. R. Lin, Y. S. Wang, Y. C. Chou, S. Z. Chen, and Y. C. Jou, “Efficient very-high color rendering index organic light-emitting diode,” Org. Electron. 12(5), 865–868 (2011). [CrossRef]
4. “Lighting for health and well-being in education, work places, nursing homes, domestic applications, and smart cities,” Deliverable D3.2 and 3.4, SSL-erate PF7 project (SSL-erate Consortium 2014).
5. J. H. Jou, M. H. Wu, S. M. Shen, H. C. Wang, S. Z. Chen, S. H. Chen, C. R. Lin, and Y. L. Hsieh, “Sunlight-style color-temperature tunable organic light-emitting diode,” Appl. Phys. Lett. 95(1), 013307 (2009). [CrossRef]
6. J. Frenkel, “On pre-breakdown phenomena in insulators and electronic semi-conductors,” Phys. Rev. 54(8), 647–648 (1938). [CrossRef]
7. S. J. Martin, G. L. B. Verschoor, M. A. Webster, and A. B. Walker, “The internal electric field distribution in bilayer organic light emitting diodes,” Org. Electron. 3(3-4), 129–141 (2002). [CrossRef]
8. X. R. Yin, Y. K. Le, X. D. Gao, Z. Y. Sun, and X. Y. Hou, “Internal potential distribution in organic light emitting diodes measured by dc bridge,” Appl. Phys. Lett. 97(15), 153305 (2010). [CrossRef]
9. T. Y. Chu and O. K. Song, “Hole mobility of N,N '-bis(naphthalen-1-yl)-N,N '-bis(phenyl) benzidine investigated by using space-charge-limited currents,” Appl. Phys. Lett. 90, 203512 (2007). [CrossRef]
10. H. Park, D. S. Shin, H. S. Yu, and H. B. Chae, “Electron mobility in tris(8-hydroxyquinoline)aluminum (Alq3) films by transient electroluminescence from single layer organic light emitting diodes,” Appl. Phys. Lett. 90(20), 202103 (2007). [CrossRef]
11. H. Bässler, “Charge Transport in Disordered Organic Photoconductors a Monte Carlo Simulation Study,” Phys. Status Solidi, B Basic Res. 175(1), 15–56 (1993). [CrossRef]
12. B. Ruhstaller, T. Beierlein, H. Riel, S. Karg, J. C. Scott, and W. Riess, “Simulating electronic and optical processes in multilayer organic light-emitting devices,” IEEE J. Sel. Top. Quantum Electron. 9(3), 723–731 (2003). [CrossRef]
13. C. H. Cheung, K. C. Kwok, S. C. Tse, and S. K. So, “Determination of carrier mobility in phenylamine by time-of-flight, dark-injection, and thin film transistor techniques,” J. Appl. Phys. 103(9), 093705 (2008). [CrossRef]
14. M. H. Ho, Y. S. Wu, S. W. Wen, M. T. Lee, T. M. Chen, C. H. Chen, K. C. Kwok, S. K. So, K. T. Yeung, Y. K. Cheng, and Z. Q. Gao, “Highly efficient deep blue organic electroluminescent device based on 1-methyl-9,10-di(1-naphthyl)anthracene,” Appl. Phys. Lett. 89(25), 252903 (2006). [CrossRef]
15. M. F. Lin, L. Wang, W. K. Wong, K. W. Cheah, H. L. Tam, M. T. Lee, and C. H. Chen, “Highly efficient and stable sky blue organic light-emitting devices,” Appl. Phys. Lett. 89(12), 121913 (2006). [CrossRef]
16. R. H. Young, C. W. Tang, and A. P. Marchetti, “Current-induced fluorescence quenching in organic light-emitting diodes,” Appl. Phys. Lett. 80(5), 874–876 (2002). [CrossRef]
17. S. W. Wen, C. K. Yen, T. H. Liu, and C. H. Chen, “Doped RGB organic electroluminescent devices based on a bipolar host material,” Proceedings of the Sixth Chinese Optoelectronics Symposium, 12–14 Sept., 263–265 (2003). [CrossRef]
18. Y. G. Lee, H. N. Lee, S. K. Kang, T. S. Oh, S. Lee, and K. H. Koh, “Fabrication of highly efficient and stable doped red organic light-emitting device using 2-methyl-9,10-di(2-napthyl)anthracene and tris(8-hydroxyquinolinato)aluminum as cohost materials,” Appl. Phys. Lett. 89(18), 183515 (2006). [CrossRef]
19. T. Smith and J. Guild, “The C.I.E. colorimetric standards and their use,” Trans. Opt. Soc. 33(3), 73–134 (1932). [CrossRef]
20. “Multimedia systems and equipment - Color measurement and management - Part 2-1: Color management - Default RGB color space - sRGB,” IEC 61966–2-1 (1999–10) ISBN: 2–8318–4989–6 - ICS codes: 33.160.60, 37.080 - TC 100 - 51 pp.
21. N. H. Kim, Y. H. Kim, J. A. Yoon, S. I. Yoo, K. W. Cheah, F. R. Zhu, and W. Y. Kim, “Efficiency enhancement of fluorescent blue organic light-emitting diodes using doped hole transport and emissive layers,” Chin. Opt. Lett. 14(4), 043001 (2016). [CrossRef]
22. W. C. H. Choy, K. N. Hui, H. H. Fong, Y. J. Liang, and P. C. Chui, “Improving the efficiency of organic light emitting devices by using co-host electron transport layer,” Thin Solid Films 509(1-2), 193–196 (2006). [CrossRef]
23. N. H. Kim, Y. H. Kim, J. A. Yoon, S. Y. Lee, D. H. Ryu, R. Wood, C. B. Moon, and W. Y. Kim, “Color optimization of single emissive white OLEDs via energy transfer between RGB fluorescent dopants,” J. Lumin. 143, 723–728 (2013). [CrossRef]
