Emission Spectral Stability Modification of Tandem Organic Light-Emitting Diodes through Controlling Charge-Carrier Migration and Outcoupling Efficiency at Intermediate/Emitting Unit Interface.

An intermediate unit (IMU) of tandem organic light-emitting diode (TOLED) can affect the emission spectral stability because of the different charge-carrier migration and outcoupling efficiency at intermediate/emitting unit interface. In this work, an IMU consisting of carbon 60 and copper(II) phthalocyanine, inserted between red-emitting unit and blue-emitting unit, were constructed and fabricated. We focus on the effect of the relative location of the red-emitting unit and blue-emitting unit in the device configuration on the emission spectral stability of TOLED. In the color-tunable TOLED, a maximum current efficiency of 20.4 cd/A, along with light-emitting spectra tuning from red emission at (0.63, 0.31) according to Commission Internationale de L'Eclairage coordinates at a bias of 10 V to white emission at (0.34, 0.27) at 22 V, is realized. Meanwhile, in the color-stable TOLED, a maximum current efficiency of 15.5 cd/A and a stable white emission at (0.31, 0.27) in a wide range of biases from 12 to 22 V are also obtained. This work might inspire a promising approach for the fabrication of color-tunable and color-stable OLEDs for both information display and solid-state lighting.

Introduction

Owing to the remarkable performance of organic light-emitting diodes (OLEDs), including wide color gamut, low power consumption, high refresh rate, and mechanical flexibility, extensive studies on them have been a research highlight in the field of display and lighting technologies for decades.[1−5] Notably, there are different technical requirements for OLED display and lighting technologies. For display demand, future high-resolution OLED display requires red–green–blue (RGB) pixel size as small as possible. Thus, it would be highly promising to utilize a novel type of emitting pixel, which vertically stacks RGB electroluminescence (EL) units in a single device, other than three different pixels. Therefore, compared to conventional OLEDs, color-tunable OLEDs hold great potential in the commercialization of next-generation ultra-high-resolution OLED displays. For lighting demand, the color stability should be taken into consideration when evaluating the lighting source. The color shift needs to be strictly controlled during operation, meaning the variation of Commission Internationale de L’Eclairage (CIE) coordinates should be as small as possible.[6]

Diverse approaches for the realization of color-tunable OLEDs have been reported. Initially, the color-tunable OLEDs adopt plurality of ways to control the emitting intensity of different emitting units. For example, Forrest et al. used Mg–Ag–indium tin oxide (ITO) as intermediate layer to connect the blue-emitting unit and red-emitting unit and realized the first color-tunable device.[7] After that, Chen et al. adopted a side-by-side red, green, and blue subcells arrangement scheme with two alternate current driving sources to fabricate color-tunable OLEDs.[8] Then, to simplify the structure of devices, a color-tunable OLED consisting of red-, green-, and blue-emitting layers (EMLs) without tandem structure was also achieved by controlling the voltage to change the position of the electron–hole recombination. For instance, Meng et al. fabricated a solution-processed OLED with four adjacent EMLs.[9] Xu et al. reported a fluorescent OLED by using two nondoped ultrathin yellow and blue EMLs.[10] Recently, inserting an exciton adjusting interlayer in two or three complementary EMLs is used as an effective route to attain color-tunable OLEDs, which further simplifies the device structure.[11] On the other hand, lighting technology, unlike the color-tunable OLEDs with their voltage-dependent color shifts, requires stable luminescent property.[12,13] According to previous research, single-emitting layer structure is a convenient approach to white light by incorporating multiple emitters (blue/orange or red/green/blue) into a single layer.[14,15] Moreover, to obtain superior efficiency and stable color spectra simultaneously, the luminophores of different emitting colors can be placed in individual units to produce white light, called as the tandem organic light-emitting diode (TOLED).[16]

As mentioned above, previous works on color-stable and color-tunable OLEDs have met the requirements of lighting and display technologies, respectively. Several theories, such as charge injection and hopping model,[17] exciton density equation,[18] and a triplet diffusion model for analyzing energy transfer,[19] have been applied to simulate charge transport and exciton transfer in OLEDs. Nevertheless, the above-mentioned mechanisms only considered electrical properties, with no optical properties included. Moreover, transformation between color-tunable and color-stable OLEDs is still challenging in current research. Furthermore, the mechanism behind this transformation may not only promote the development of both color-tunable and color-stable OLEDs but also realize the integration of multifunctional devices.

In this work, we demonstrate the effect of intermediate unit (IMU) consisting of carbon 60 (C60) and copper(II) phthalocyanine (CuPc) on device emission spectral stability using two tandem OLEDs (TOLEDs) with a blue-emitting unit and another red-emitting unit. By simply switching the relative location of the red-emitting unit and blue-emitting unit in the device configuration, both color-tunable and color-stable light emission can be simultaneously obtained, which is much different from the previous studies of either color-stable or color-tunable tandem OLED. A maximum current efficiency of 20.4 cd/A and various color tuned from red at (0.63, 0.31) to white at (0.34, 0.27) are achieved in color-tunable TOLED. On the other hand, the color-stable TOLED demonstrates a stable white at (0.31, 0.27) under a 12–22 V driving voltage with a maximum current efficiency of 15.5 cd/A. Furthermore, to gain deep understanding of light outcoupling mechanisms of OLEDs, a finite-difference time-domain (FDTD) method is used to calculate the field intensity distribution of the blue and red emission,[20,21] and the distribution of the optical characteristics of both the color-tunable and color-stable TOLEDs is analyzed in detail.

Experimental Section

Device Fabrication

Transparent indium tin oxide (ITO)-coated glasses with a sheet resistance of 15 Ω/sq were cleaned in an ultrasonic bath with acetone, deionized water, and isopropanol for 15 min. After drying in an oven (80 °C), the ITO glasses were treated with UV/ozone (SunMonde, UV-O3 Cleaner, 40 W, Shanghai, China) for 15 min before loading into the vacuum chamber. Subsequently, ITO glass substrates were transferred into a high-vacuum evaporation chamber with a basic pressure of less than 3 × 10–6 Torr to deposit each functional layer. Figure shows the schematic diagram of the resulting OLEDs. In blue/red-emitting unit, 40 nm N,N′-bis-(3-naphthy)-N,N′-biphenyl-(1,1′-biphenyl)-4,4′-diamine (NPB) was used as hole-transporting layer. The 30 nm host materials N,N′-dicarbazolyl-4-4′-biphenyl (CBP) doped with bis[2-(2,4-difluorophenyl)-pyridyl]picolinate iridium(III) (FIrpic) acted as blue EML, whereas for red-emitting unit, 30 nm host material CBP doped with bis(1-phenyisoquionoline)(acetylacetonate)iridium(III) (Ir(piq)2acac) acted as EML.[22] The 30 nm 4,7-diphenyl-1,10-phenanthroline (Bphen) acted as electron-transporting layer. Red- and blue-emitting units were connected in series via a planar heterojunction, which consisted of C60 and CuPc. To modify the energy level of C60 and CuPc, lithium fluoride (LiF) and molybdenum oxide (MoO) buffers are introduced, respectively.[23] The evaporating rates of these materials were controlled at 0.1–0.2 nm/s, and the Al cathode was controlled at 0.5–1 nm/s, respectively. The active emission area of each OLED was 3 × 3 mm2.

Figure 1

(a) Schematic structure of tandem OLED with C60/CuPc IMU as intermediate layer. (b) Energy-level diagram of materials in tandem OLED.

Properties Characterization

The current density–voltage-luminance characteristics and EL spectra of the TOLEDs were simultaneously characterized by a computer-controlled programmable Keithley model 2400 power source and a Photo Research PR655 luminance meter/spectrometer in ambient environment. Surface morphologies were characterized by an atomic force microscope (AFM) (Agilent, AFM 5500) operating in tapping mode. Optical transmittance and absorption spectra of C60 and CuPc were recorded by a HITACHI U-3900 UV–vis scanning spectrophotometer. Ultraviolet photoemission spectroscopy (UPS) measurement was conducted to characterize the highest occupied molecular orbital (HOMO) level of organic semiconductors by a Thermo Scientific Escalab 250Xi ultra-high-vacuum system. The UPS measurements were performed with an unfiltered He I (hv = 21.22 eV) gas discharge lamp and a total instrumental energy resolution of 100 meV.

Optical Modeling and Simulation

Optical simulation was conducted using the FDTD simulation (Lumerical FDTD Solutions 8.17.1157) with a trial license to calculate the field intensity distributions (including near-field Poynting vector distribution) in the devices. In the optical modeling, frequency-dependent refractive index (n) and extinction coefficient (k) of ITO, C60, CuPc, and emitting unit were experimentally determined by the VUV-VASE Ellipsometer (J.A. Woollam Co., Inc.) and then the Al cathode and glass were fitted with the complex refractive index in the FDTD simulation.[24]

Results and Discussion

Electrical Properties of C60/CuPc IMU

One important issue in tandem OLEDs is the effective charge injection and good transport for the intermediate connector.[25] According to previous research works, an optimizing thickness of C60/CuPc IMU is used in tandem OLED.[23,34] To further investigate the role of C60/CuPc IMU between two emitting units, we examined the charge-generation, charge-injection, and transport characteristics of the intermediate unit. To clearly elucidate the ability of charge generation and charge injection in the LiF/C60/CuPc/MoO3 IMU, we fabricated the charge-generation-only devices (B-1, B-2, and B-3). The detailed layer structures of the B-series devices are shown in Table . As shown in Figure a, in device B-1 (without C60/CuPc IMU), current density under a forward bias of 10 V is 5.7 × 10–4 mA/cm2, indicating that a huge charge-transport barrier exists in device B-1. On the other hand, the current response in device B-2 is 16.4 mA/cm2 at a forward bias of 12 V, whereas in device B-3, the introduction of an ultrathin LiF layer between Bphen and C60, along with a thin layer of MoO3 in between CuPc and NPB, greatly enhances the device current to 24.5 mA/cm2 under the bias of 12 V. This indicates that the generated electrons in the IMU cannot effectively inject to Bhpen from C60 in device B-2. Electron-injection barrier between C60 and Bphen and hole-injection barrier between CuPc and NPB are effectively repressed by introducing LiF and MoO3, respectively. It is worth noting that the HOMO of CuPc (−5.3 eV) is only 0.1 eV higher than that of NPB (−5.4 eV). Nevertheless, the lowest unoccupied molecular orbital (LUMO) of C60 (−3.88 eV) is 0.98 eV lower than that of Bphen (−2.9 eV), which means the injection and transportation of holes are better than those of electrons. According to previous reports, the presence of an ultrathin LiF-modifying layer can decrease the energy barrier at organic/organic interfaces via tunneling.[26,27] To confirm this possible mechanism, UPS measurements were further performed. Figure b shows the spectra acquired for C60 and C60 with an ultrathin LiF (1 nm). The Ecutoff was defined by the crosspoint of the linear exploration of the high-binding-energy cutoffs and the straight background line in the spectrum,[28] and the Eonset is referred to as the binding energy onset.[29] From Figure b, the Ecutoff of C60 and C60/LiF are 17.63 and 17.69 eV, respectively, and the corresponding Eonset values are 2.78 and 2.62 eV. The HOMO energies are obtained from the following equation[30]where hv is 21.22 eV. Consequently, the calculated HOMO energies are −6.37 eV (bare C60) and −6.15 eV (C60/LiF), respectively. According to the optical gaps (2.49 eV for C60 and 2.56 eV for C60/LiF) obtained from the UV–vis absorption spectra (Figure c), the LUMO energy levels are estimated to be −3.88 eV for pristine C60 and −3.59 eV for C60/LiF. The optical band gap of C60 and C60/LiF was estimated according to the equation[31,32]where α is the extinction coefficient of C60 and a is a constant. The variation in LUMO energy levels between C60 with and without LiF indicates that LiF can fine-tune the LUMO of C60, but there is still a charge barrier of 0.69 eV between C60/LiF and Bphen.

Table 1

Layer Structure of Charge-Generation Devices (B-1–B-3) in Series B

device	structure
B-1	ITO/MoO3 (3 nm)/Bphen (100 nm)/NPB (100 nm)/LiF (1 nm)/Al (150 nm)
B-2	ITO/MoO3 (3 nm)/Bphen (100 nm)/C60 (5 nm)/CuPc (5 nm)/NPB (100 nm)/LiF (1 nm)/Al (150 nm)
B-3	ITO/MoO3 (3 nm)/Bphen (100 nm)/LiF (1 nm)/C60 (5 nm)/CuPc (5 nm)/MoO3 (3 nm)/NPB (100 nm)/LiF (1 nm)/Al (150 nm)

Figure 2

(a) Current density–voltage characteristics of the charge-generation-only devices (B-1–B-3) and the energy-level diagram of the charge-generation-only device. (b) UPS images of the inelastic cutoff region (left) and the HOMO region (right) of C60 and C60/LiF. (c) Absorption coefficient of C60 and C60/LiF film.

Performance Analysis in Tandem OLEDs with IMU

Consequently, two tandem OLED devices (A-1 and A-2) based on different sequence of emitting units were fabricated, as shown in Table . In these devices, C60/CuPc IMU was employed as the charge-generation layer between two emitting units (blue and red). These two separated emitting units with complementary colors might lead to white OLED.[12,13] As shown in Figure a, when the red-emitting unit is laid adjacent to the anode ITO and the blue-emitting unit was close to the cathode Al, a superior color stability can be obtained in device A-1. The EL spectra of the present devices turn out to be exceedingly stable within the investigated voltage ranges (12–22 V). More specifically, peak emission at 496 nm with a shoulder at 476 nm is from FIrpic emission,[33] whereas another peak emission at 632 nm with a shoulder at 672 nm is from Ir(piq)2acac.[34] Driven by the investigated voltage range (12–22 V), the CIE coordinates of device A-1 kept at (0.31, 0.27), which are close to the standard white CIE coordinates of (0.33, 0.33). Nevertheless, when the red- and blue-emitting units switched position, the red-emitting unit is adjacent to the cathode Al. Although blue-emitting unit was close to the anode ITO, color-tunable property was observed in device A-2. Similar to device A-1, a peak emission at 456 nm with a shoulder at 476 nm is from FIrpic emission and another peak emission at 628 nm with a shoulder at 668 nm is from Ir(piq)2acac. As shown in Figure b, device A-2 initially shows a predominantly red emission spectrum at (0.63, 0.31) with 10.3 cd/m2 at 10 V, when then turns orange at (0.48, 0.29) with 1102 cd/m2 at 16 V. Finally, it becomes white at (0.34, 0.27) with 15 774 cd/m2 at 22 V.

Table 2

Layer Structure of OLEDs in Series A

device	structure
A-1	ITO/MoO3 (3 nm)/red-emitting unit/LiF (1 nm)/C60 (5 nm)/CuPc (5 nm)/MoO3 (3 nm)/blue-emitting unit/LiF (1 nm)/Al (150 nm)
A-2	ITO/MoO3 (3 nm)/blue-emitting unit/LiF (1 nm)/C60 (5 nm)/CuPc (5 nm)/MoO3 (3 nm)/red-emitting unit/LiF (1 nm)/Al (150 nm)
red-emitting unit	NPB (40 nm)/CBP:Ir(piq)2acac (30 nm, 8 vol %)/Bphen (30 nm)
blue-emitting unit	NPB (40 nm)/CBP:FIrpic (30 nm, 8 vol %)/Bphen (30 nm)

Figure 3

Normalized EL spectra of (a) device A-1 at 12–22 V and (b) device A-2 at 10–22 V. (c) CIE 1931 coordinate of devices A-1 and A-2 at 10–22 V.

Typically, TOLEDs have a stacked architecture of multiple emitting units with intermediate layers between adjacent emitting units, which means that each emitting unit can work independently. As described above, combining the blue-emitting unit with red-emitting unit can achieve a color-stable white OLED in device A-1, whereas the emission spectrum of device A-2 can be tuned to be any linear superposition of red and white light emission by varying the driving voltage. As both devices A-1 and A-2 are fabricated by identical materials, the reason of this color-stable and color-tunable variation can be attributed to the different relative positions of the two emitting units. As motioned above, the charge barrier between C60 and Bphen is larger than that between CuPc and NPB, which means that the injection and transportation of holes are superior to those of electrons. Therefore, the emitting unit adjacent to the CuPc side may get more carriers than the emitting unit adjacent to C60. When two emitting units are biased under the same voltage, the emitting unit adjacent to the CuPc side can realize a better balance of hole/electron carriers than the emitting unit adjacent to C60. Another important factor that may make some influences is the effects of fundamental EL mechanisms on host–guest system: the host–guest energy transfer and direct exciton formation. On the one hand, different triplet energy level (ET) of doping and host material can affect the energy transfer. The ET of FIrpic (2.62 eV) is slightly higher than that of CBP host (2.56 eV),[35] whereas the ET of Ir(piq)2acac (2.0 eV) is lower than that of CBP host,[36] leading to a more efficient energy transfer from CBP to Ir(piq)2acac than to FIrpic.[37] On the other hand, according to the J–V curves of carrier-only devices with and without dopants in Figure S1, the red-emitting unit can directly trap more carriers to form excitons, leading to a higher EL intensity than blue unit. Furthermore, by fabricating two single-emitting unit devices and comparing the EL intensity at the same voltage (Figure S2), we find that the efficiency of exciton utilization of Ir(piq)2acac is higher than that of FIrpic. On the basis of the above-mentioned reasons of device structure and organic material, when the red-emitting unit with Ir(piq)2acac is adjacent to C60 (device A-1), the charge barrier between C60 and Bphen will limit the number of injected electrons from IMU to red-emitting unit at low voltage. This difference in injection ability will lead to less excitons in the red-emitting unit. Although the blue-emitting unit can obtain more excitons, it is limited by the inefficient energy transfer from CBP host to FIrpic. In contrast, a more efficient energy transfer from CBP to Ir(piq)2acac leads to better utilization of triplet energy excitons, contributing to a sufficient strength and complementing with blue emission to get white emission. So, with increasing voltage, both blue and red emission intensities remained almost identical. As a consequence, device A-1 can attain a color-stable white emission.

When blue-emitting unit with FIrpic is adjacent to C60 (device A-2), the working mode of color-tunable device can be divided into three steps: (1) At low voltage (from 10 to 12 V), not only the charge-generation barrier between C60 and Bphen, but also the inefficiency energy transfer from CBP host to FIrpic will limit the intensity of blue emission. Therefore, the intensity of red emission is much higher than that of the blue emission, and the resultant of mixing emission exhibits a red color with the CIE coordinates (0.63, 0.31). (2) As the driving voltage continues to increase (from 12 to 18 V), more and more electrons can overcome the charge barrier, which leads to a great increasing in intensity of blue emission. For example, when applying a voltage of 16 V, the blue emission is enhanced, which shifts the CIE coordinates from near (0.63, 0.31) to (0.48, 0.29), as shown in Figure b. (3) At the highest voltage (from 18 to 22 V), the device shows white emission arising from comparable blue and red emission intensities, as the blue intensity is gradually enhanced with increased voltage.

The EL performances of devices A-1 and A-2 are shown in Figure and summarized in Table . The turn-on voltage of device A-1 is 8.4 V, which is higher than that of device A-2 (7.7 V). Therefore, the structure of device A-2 has a higher charge-transfer ability than that of device A-1. At a brightness of 10 000 cd/m2, the operating voltage of device A-2 (20.9 V) is lower than that of device A-1 (21.7 V). C60/CuPc IMU possesses similar conductivity in both devices, indicating that the order of emitting units can achieve a better charge injection in A-2. The current density–voltage–brightness characteristics are displayed in Figure a. Device A-1 has operational current densities of 1.65, 14.98, and 194 mA/cm2 at luminances of 100, 1000, and 10 000 cd/m2, respectively. To get the same luminance, device A-2 with C60/CuPc as photovoltaic IMU exhibits lower current densities of 0.80 mA/cm2 (100 cd/m2), 4.91 mA/cm2 (1000 cd/m2), and 68.4 mA/cm2 (10 000 cd/m2), respectively. Figure b,c shows the luminance–current efficiency–external quantum efficiency (L–CE–EQE) and luminance–power efficiency (L–PE) characteristics of OLEDs. From Figure b and Table , we can see that the current efficiency is enhanced from 15.2/12.9/8.8 (cd/A) in device A-2 to 20.3/17.7/13.8 (cd/A) in device A-1 at 100/1000/10 000 (cd/m2), respectively. According to the definition of current efficiency, the brightness of device A-2 is higher than that of device A-1 for an identical current density. Moreover, the same enhancement of EQE is also accomplished as shown in Figure b. Furthermore, the maximum efficiencies are 4.3 lm/W and 15.5 cd/A for device A-1, and 6.2 lm/W and 20.4 cd/A for device A-2, respectively. As the power efficiency of an OLED is influenced by current efficiency and driving voltage of OLED, the enhancement of power efficiency in device A-2 is attributed to the improvement of current efficiency and the reduction of driving voltage. Actually, the efficiency of whole device is closely related to the efficiency of the blue- and red-emitting units, especially the efficiency of the red-emitting unit. In color-stable OLED, the charge barrier between C60 and Bphen leads to a difficulty of electronic injection and transportation. Therefore, the decrease of efficiency of the red-emitting unit results in a decrease in the whole device efficiency. By comparing the changes in device performance during the increase of voltage and the intensity of non-normalized EL spectral (Figure S3), it can be found that a more recombination could synchronously happen in both red- and blue-emitting units with increasing voltage. As a result, the EL spectra of color-stable device in Figure do not increase with voltage through normalizing the data of EL intensity.

Figure 4

Comparison of devices A-1 and A-2: (a) current density–luminance–voltage characteristics; (b) brightness vs current efficiency and external quantum efficiency; and (c) brightness vs power efficiency.

Table 3

EL Performance of OLEDs; J, Lmax, Von, CE-max and PE-max, Present Current Density, Maximum Luminance, Turn-On Voltage, Maximum Current Efficiency, and Maximum Power Efficiency

device	Von (V)	Lmax (cd/m2)	J at 100 cd/m2 (mA/cm2)	J at 1000 cd/m2 (mA/cm2)	J at 10 000 cd/m2 (mA/cm2)	CE-max (cd/A)	PE-max (lm/W)
A-1	8.4	12 062	1.65	14.98	194	15.5	4.3
A-2	7.7	16 174	0.80	4.91	68.2	20.4	6.2

Optical Modeling of Devices

To further analyze the color-tunable and color-stable behaviors of TOLEDs, the optical modeling calculations of near-field light propagation inside the devices were conducted by employing the FDTD method. In the optical simulation, one plane wave source with vertical orientations at two emission wavelengths is located in different positions (1st and 2nd emitting units, Figure ) to study the light propagation in the tandem structure. For accuracy, the shape features of the contact surface of each functional layer are determined from the AFM images. As shown in Figure , ITO, 1st emitting unit, C60, and CuPc are deposited sequentially, and the surface and section profile morphologies of each functional were measured by AFM. The pristine ITO layer has a root mean square (RMS) of 2.12 nm, when 100 nm emitting unit is deposited on ITO, which can be smoothed to 1.09 nm. Then, 5 nm C60 is deposited on the emitting unit, and only part of sublayer can be covered to form some peaks. Finally, 5 nm CuPc is deposited on C60, which has an RMS of 2.02 nm. We take this practical morphology feature as the simulation parameter to get more accurate simulation results. Figure shows the simulated cross-sectional views of field intensities induced by a plane wave source with vertical orientation (normal to the substrate) at wavelengths of 625 nm (Figure a,b) and 476 nm (Figure c,d). We note that the simulation results qualitatively show the influence of the location of the red- and blue-emitting units on the light path. Specifically, the red regions represent the areas with the greatest field intensities.

Figure 6

Normalized cross section of simulated energy flux densities in TOLEDs with vertical orientation (normal to the substrate) at wavelengths of 625 nm (a, b) and 476 nm (c, d). The source in the second emitting unit (a, d) and the first emitting unit (b, c).

Figure 5

AFM images and line scan profiles of the contact surface of each functional layer: (a) ITO, (b) first emitting unit, (c) C60, and (d) CuPc.

As shown in Figure a,b, wherever the source of 625 nm is, a good light outcoupling efficiency can be obtained. When put the source of 476 nm in the first emitting unit, as shown in Figure c, although a part of internally generated light will be confined in the ITO layer, there is still enough light intensity distribution in glass. However, the calculated field distributions in Figure d clearly show that a large portion of internally generated light at 476 nm will be confined in the emitting unit. By combining the simulation results with the device structure and performance, light intensity distribution in glass will affect the outcoupling efficiency, leading to either color stability or tunability. In color-tunable OLED as device A-2, when blue-emitting unit is adjacent to C60 and red-emitting unit is adjacent to CuPc (Figure a,c), the same light intensity distribution in glass indicates that the outcoupling efficiencies of both blue and red emissions are high. Therefore, the main factor is the relative intensity of the two different colors, which will then determine the emitting color. As previously discussed, color tunability is obtained due to the different charge-injection capacity from IMU to the blue- and red-emitting units. On the contrary, in color-stable OLED as device A-1, when the red-emitting unit is adjacent to C60 and the blue-emitting unit is adjacent to CuPc (Figure b,d), the light outcoupling efficiency of red emission is higher than that of blue emission. Although the intensity of red emission is restricted by the energy barrier between the IMU and the emitting unit, the intensity of the red emission is still higher than that of the blue emission when the driving voltage is relatively low. Consequently, as shown in Figure , it can be inferred that the emitting intensity can be changed by not only adjusting the energy barrier between the IMU and the adjacent emitting units but also manipulating the position of emitting unit on the light path to adjust the light outcoupling efficiency.

Figure 7

Effect of IMU’s charge-carrier migration and outcoupling efficiency on device emission spectral stability: (a) color stability, (b) color tunability.

Conclusions

In summary, color-tunable and color-stable TOLEDs with C60 and CuPc IMU are realized by simply switching the relative positions of the red- and blue-emitting units on the light path. When the blue-emitting unit is adjacent to C60, a maximum current efficiency of 20.4 cd/A is achieved and various colors tuned from red (0.63, 0.31) to white at the CIE coordinates of (0.34, 0.27). Meanwhile, when the red-emitting unit is adjacent to C60, the device demonstrates a stable white (0.31, 0.27) and a maximum current efficiency of 15.5 cd/A. By analyzing the energy level and utilizing optical simulation, it is found that not only the energy barrier between C60/Bphen and different ET of emitting materials, but also the outcoupling efficiency will affect the color stability and tunability. This work demonstrates a promising way to fabricate color-stable and color-tunable OLEDs by using a simple structure for both display panels and lighting technology.