1. Introduction

In recent years, the organic light emitting diode (OLED) device technology is aggressively progressing from the rigid to a flexible structure. The flexible OLED (FOLED) device is mainly composed of the driving circuit, the emitting device, and the encapsulation. Since the organic materials used in OLED devices are sensitive to oxygen and moisture, the hermetic encapsulation is vital. In FOLED, the thin film encapsulation (TFE) technique is widely used. In the TFE, a multilayer inorganic thin film structure and/or an inorganic/organic thin film structure are mainly employed. Multilayer inorganic thin film structures include aluminum oxide (Al₂O₃)/ zinc oxide, Al₂O₃/ silicon nitride (SiN_x), Al₂O₃/ zirconia [1–3], etc. While the inorganic/organic thin film structures primarily use inorganic materials such as Al₂O₃, SiN_x, and silicon oxide, and polymers as laminators [4–10].

The top emission OLED (TEOLED) structure is mostly used in FOLED display devices. Features like a high aperture ratio and excellent color purity due to the micro-cavity effects of TEOLED devices made them very attractive for the use in flexible applications. When the TFE is applied to a TEOLED device, the generated light in the emitting layer (EML) has to travel to the top cathode and then exits through the TFE structure. The TFE formed on the top of the resonant structure also has an interfacial reflection and absorption, which affect the optical properties of the device. Furthermore, the TFE formed on the capping layer (CPL) may also affect the micro-cavity effect because it has a dielectric mirror property due to the multilayer thin film structure having different RI values. Such a configuration may negate the advantage of the enhanced micro-cavity effect in the TEOLED device structure. Therefore, it is imperative to control the optical properties of the TEOLED device with TFE. Several studies have been reported on improving the light extraction characteristics of OLED devices, including simulations of methods to enhance their optical efficiency [11,12]. Among these, Jang et al. formed wrinkles directly on the TFE surface without using complex processes such as patterning of a TEOLED device in 2019 [13]. The corrugation on the TFE surface widened the angle of the escape cone, thereby extracting the trapped light into the air mode. In such an arrangement, the current efficiency (CE) was increased by a factor of 1.4. Recently, Cho et al. reported a structure in which the current efficiency could be increased by about 40% using a double-layer nano-structure on the Al₂O₃/ SiN_x TFE [14]. Li et al. formed a micro-lens array on the TFE and achieved the extraction light improvement by 49% [15]. Although several methods for the light extraction at the interface between TFE and air have been reported, there is a need for an innovative technique that maintains the flexibility while minimizing the light loss by the TFE. Most of the reported methods to date are very complex and expensive. Furthermore, the above reported methods are essentially used for extracting the surface trapped light of TFE. Most of the reported techniques for improving the optical efficiency of OLED devices so far have focused on maximizing external light extraction. However, we have concentrated on the extraction of trapped light inside the device. A method for extracting the internal light trapped at the interface due to the difference in refractive index (RI) between CPL and TFE has not been reported yet. There is a pressing need to address this issue of light trapped inside due to the difference in RI between CPL and TFE. Therefore, a facile internal light extraction technology is required.

In this paper, a technique for extracting the internal light trapped between the CPL and TFE of a TEOLED device with TFE through RI matching was studied. Herein, a novel structure incorporating the light extraction concept and the encapsulation of a TEOLED device is presented. By inserting a low RI layer between the CPL and Al₂O₃ of the TFE, the influence of the incident angle is lowered, which reduces the amount of light trapped inside, thereby improving the optical properties. When the low RI layer is applied, an evanescent wave occurs in the low RI layer due to the difference in RI between the CPL/ low RI layer/ Al₂O₃ structure layers. Through the optical simulation, it was confirmed that the optical characteristics of the TEOLED device with TFE using a low RI layer were improved and the waveguide (WG) mode losses could be reduced. A blue TEOLED device was fabricated by applying a low RI layer. Results showed that the CE and blue index (BI) efficiency are increased by about 23% and about 26%, respectively, compared to those of the control device. The improvement in the light extraction is attributed to the evanescent wave and its electric field present in the low RI layer and the modification of the micro-cavity conditions (i.e., minimizing the light loss modes).

2. Experimental

2.1 Optical simulation

SETFOS 5.1 is used as an optical simulator. The RI of the glass substrate, indium tin oxide (ITO), and organic layers are taken as ∼1.5, 1.8∼2.0, and ∼1.8, respectively. The RI and extinction coefficient of silver (Ag) and magnesium (Mg):Ag (10:1) used as an anode and cathode, respectively, are taken from the reported values. For the EML, the thin film photoluminescence spectrum of 2,12-di-tert-butyl-N,N,5,9-tetrakis(4-(tert-butyl)phenyl)-5,9-dihydro-5,9-diaza-13b-boranaphtho[3,2,1-de]anthracen-7-amine (DABNA-NP-TB) for the blue TEOLED device, bis[2-(2-pyridinyl-N)phenyl-C](acetylacetonato)iridium(III) for the green TEOLED device, and (bis(4-methyl-2-(3,5-dimethylphenyl)quinoline))Ir(III)(tetramethylheptadionate) for the red TEOLED device are used.

2.2 TFE fabrication

Al₂O₃, deposited using atomic layer deposition (ALD), is used as the inorganic layer in the TFE structure. Tri-methylaluminum (TMA) is used as a precursor for Al₂O₃ deposition, and deionized water (H₂O) is used as a reactant. Nitrogen (N₂) is used as an inert gas to purge the precursor and reactant. Al₂O₃ thin film with a thickness of about 100 nm was obtained through 800 cycles with a TMA injection time of 0.5 seconds, a TMA purge time of 20.0 seconds, a H₂O injection time of 1.0 seconds and a H₂O purge time of 40.0 seconds. The growth rate per cycle of this Al₂O₃ thin film is about 1.25 Å/cycle. In TFE, the organic layer of polyvinyl butyral (PVB) is formed by the spin coating process. A PVB solution with a concentration of 60 mg/ml was prepared in a glove box at room temperature and in an N₂ atmosphere. The spin coating was performed at 2,000 rpm for 20 seconds in a glove box at room temperature and in an N₂ atmosphere, and the annealing was performed at 80 °C for 30 minutes. Through this process, a thin film with a thickness of about 1,000 nm was obtained.

2.3 Low RI thin film formation

In the ALD process, gaseous precursors and reactants are alternately injected and adsorbed on the surface to form the thin film. After a certain amount of atoms are adsorbed on the surface, a self-limiting reaction occurs in which the adsorption does not proceed any further, thereby forming a uniform thin film. During the precursor injection process, adsorption occurs not only on the substrate surface but also between the precursors. The precursor adsorbed on the substrate is strongly bound by chemical adsorption. However, the adsorption between the precursors is a physical adsorption. As a result, atoms can be easily removed because the binding force between precursor atoms is weak. Unwanted adsorption between precursors can be purged with inert gases such as N₂ and argon. Reactant is also adsorbed on the precursor, and the remaining reactant is purged by injecting the inert gas. If the purge time of the unadsorbed precursor and reactant is too short, some unwanted atoms may remain inside. This causes a vapor phase reaction similar to the chemical vapor deposition (CVD) process, and this parasitic CVD growth leads to a rapid increase in the growth rate of the thin film, resulting in a lower RI [16].

In this study, low RI thin films were fabricated using TMA and H₂O as precursors and reactants for the ALD, respectively. Since H₂O is used as a reactant, it strongly adheres to the chamber wall. The time required for desorption of H₂O takes a considerable amount of time due to its polar nature, and thus the purge time is longer than that of TMA. If the purge time is not sufficient, as described above, a parasitic CVD reaction occurs. As a result, carbon was present in the thin film [17,18], resulting in a decrease in density and a decrease in the RI of the deposited thin film. Taking advantage of this characteristic, we fabricated thin films containing carbon with different refractive indices by controlling the H₂O purge time. The thin film deposition was performed in 100 cycles at the same temperature of 80 °C. TMA injection time, TMA purge time, and H₂O injection time are the same as 0.5, 5.0, and 1.0 seconds, respectively. H₂O purge times are fixed at 5.0, 10.0, 15.0, 20.0, and 40.0 seconds, respectively. The RI and thickness of the five thin films fabricated by controlling the H₂O purge time were measured using an ellipsometer. Fig. S1(a) shows the RI of the thin film according to the purge time. The thin film, according to the H₂O purge times of 5.0, 10.0, 15.0, 20.0, and 40.0 seconds, has refractive indices of 1.22, 1.28, 1.48, 1.60, and 1.63, respectively, measured at a wavelength of 458 nm. As the H₂O purge time increases, the RI of the thin film increases. On the other hand, the extinction coefficient is almost the same, indicating that there is no significant change in the light absorption by the deposited thin films with the variation of RI (see Fig. S1(b)). As a consequence, it was confirmed that the RI of the thin film can be controlled simply through the H₂O purge time. In brief, we can fabricate a low RI thin film. Results of RI and extinction coefficient according to H₂O purge time are summarized in Table S1 (Supplement 1). Additionally, even if the carbon is present in a thin film, the surface roughness does not differ significantly. Therefore, it is argued that there will be a little impact on the light properties of the deposited film [19].

2.4 Blue TEOLED device fabrication

The blue TEOLED devices were fabricated using thermal vacuum evaporation equipment containing the following materials. Anode substrates of Ag (150 nm)/ ITO (10 nm) with a size of 35 mm x 35 mm were fabricated by sputtering deposition. 1,4,5,8,9,11-Hexaazatriphenylenehexacarbonitrile (HATCN) is used as the hole injection layer (HIL). N-(1,1′ -biphenyl-4-yl)-N-[4-(9-phenyl-9Hcarbazol-3-yl)phenyl1]-9,9-dimethyl-9H-fluoren-2-amine (PCBBiF) is applied as the hole transport layer (HTL) as well as the CPL layer. In the EML, 9-(4-(10-phenylan-thracene-9-yl) phenyl)-9H-carbazole (PhPC) is doped with 5% DABNA-NP-TB. 1,3-bis(9-phenyl-1,10-phenanthrolin-2-yl)benzene (BPPB) is used for the electron transport layer (ETL), and 8-hydroxyquinolinolato-lithium (Liq) is used as an electron injection layer. For the semi-transparent cathode, Mg:Ag was used in a ratio of 10:1. The transmittance of the fabricated cathode was about 30-40%.

3. Results and discussion

3.1 Design of thin film encapsulation

Since the TFE process is indispensable for the flexible TEOLED devices, the TFE layer must have acceptable characteristics such as a high transmittance and a hermetic water barrier effect. Therefore, the TFE is designed by taking into consideration the transmittance and barrier effects. To ensure the barrier properties of TFE, the Al₂O₃ thin film was deposited using ALD with excellent step coverage to minimize the pinholes in inorganic layers. As reported, the thickness of Al₂O₃ was set to 100 nm with 1.5 dyad TFEs for the appropriate barrier effect [2,20]. The deposited Al₂O₃ thin film layer by ALD has a high density and a RI of about 1.62 in the blue wavelength region (see Fig. S2(a)). The transmittance is about 83%, which is about 8% lower compared with the normal glass type encapsulation (over 90%). But this value of transmittance is typically observed in amorphous Al₂O₃ films, and it hardly interferes with the light emission (see Fig. S2(b)). As a polymer layer, PVB is widely used as an optically clear adhesive layer since it can improve the optical properties and durability of the TFE structure. The PVB thickness of 1,000 nm is used to increase the water vapor diffusion path in the TFE structure, which sufficiently reduces the diffusivity by chemically trapping the water vapor [21,22]. In the blue wavelength region, the RI of PVB is about 1.53, similar to that of glass. The transmittance is also very high, at about 97% (see Fig. S2(a), (b)). It is well known that the water vapor transmission rate (WVTR) decreases as the dyad of TFE increases. But here, it is intended to use the minimum dyad of the TFE to consider the process time. Earlier, Han et al. and Kwon et al. reported 1.5 dyad TFEs with WVTRs of 1.6 × 10⁻⁴g/m²/day and 2.0 × 10⁻⁶g/m²/day, respectively [23,24]. To verify the water barrier performance of our 1.5 dyad TFEs, the TFE property was determined by the calcium (Ca) test. The 144 Ca dots on a 50 mm x 50 mm glass substrate were deposited with Al₂O₃/ PVB/ Al₂O₃ (APA) structure. The Ca dot samples were stored in an accelerated environment of 85 °C and 85% relative humidity, and the oxidation of Ca was observed. After 300 hours, only one dot was oxidized, and a total of two dots were oxidized at 500 hours. Even after 1,000 hours, there were no more oxidized dots (as shown in Fig. S3), indicating that our 1.5 dyad TFE is good enough for the encapsulation. Therefore, for further optical study, this APA 1.5 dyad structure (Fig. S2(c)) was selected and used. Our measured transmittance of APA TFE is significantly high, at about 83% over the visible region (see Fig. S2(b)). Since the difference in RI between Al₂O₃ (RI = 1.61∼1.59) and PVB (RI = 1.53∼1.51) over the visible range is small (∼0.8) (see Fig. S2(a)), the additional light loss at the interface due to the interference effects was minimized. The small RI difference between Al₂O₃ and PVB results in a critical angle as large as 72.99° when the light is incident from Al₂O₃ to PVB. Indeed, this indicates that our current APA structure is most suitable for use as a TFE for the TEOLED device because the probability of generating total internal reflection (TIR) is almost minimized. As a result, the TFE of the APA 1.5 dyads structure was applied to the TEOLED device.

3.2 Design of device for light extraction

Generally, the high RI material is used as a CPL on the semi-transparent cathode for the light extraction in TEOLED devices. In the present study, the PCBBiF with a refractive index of about 2.03, was used as a CPL. When the TEOLED device using CPL is subjected to the APA TFE, the TIR occurs due to the difference in RI when light is incident on the Al₂O₃ film from the CPL; Al₂O₃ has a relatively low RI compared with the CPL. As a consequence, the incident light is not emitted out completely due to the TIR and is trapped inside, resulting in a light loss (see Fig. 1(a)). Thus, in order to extract the light trapped inside by the CPL/ Al₂O₃ interface, we propose to insert a low RI layer at the interface between the CPL and Al₂O₃ layers of TFE. The CPL/ low RI layer/ Al₂O₃ configuration has a high RI (2.03)/ low RI (≈ 1.28)/ high RI (1.61) value order. With the presence of the low RI layer, an evanescent field is formed in the low RI layer, thereby changing the light direction [25,26]. The incident light appears in the high RI (1.61) Al₂O₃ layer, and the internal reflected light at the CPL/low RI layer interface is reduced considerably. As a consequence, the output light intensity can be improved significantly. Therefore, the concept of a low RI layer to improve the light extraction is applied to the beneath of the general TFE structure (see Fig. 1(b)).

 

Fig. 1. The concept of light amplification by low RI layer. (a) Without low RI layer; (b) with low RI layer

Download Full Size | PDF

3.3 Light extraction through evanescent wave

For the excellent optical properties of the TEOLED devices, the light generated from the EML must pass through the organic layer and be transmitted to the air with a low RI (n = 1) [27,28]. In this process, if the incident angle ${\theta _i}$ is greater than the critical angle ${\theta _c}$, the TIR occurs [29]. TIR light that escapes to the edge of the device or reflects to the inside of the device cannot be emitted out, resulting in deterioration of its optical properties. Generally, when the light is completely internally reflected, no light is propagated in the + z-direction beyond the interface of the high RI (${n_1}$)-low RI layer (${n_3}$). However, the evanescent waves are formed in the low RI layer, propagating in the + x-direction. The source of evanescent waves is the boundary between the ${n_1}$ and ${n_3}$. Further, a finite exponential decaying electric field across the interface penetrates for some distance into the optically low-dense medium (in ${n_3}$) in the + z-direction which is known as the evanescent electric field (see Fig. 2(a)). The amplitudes ${\vec{E}_{0i}}$, ${\vec{E}_{0r}}$ and ${\vec{E}_{0t}}$ of the incident wave, reflected wave, and transmitted wave can be calculated using the Fresnel equation [30,31] when the plane wave travels from ${n_1}$ to ${n_3}$. Depending on the polarization state, the wave has different reflection coefficients and transmission coefficients. If the electric field is perpendicular to the plane of incident, it is stipulated as s-polarization, and the polarization present in the plane consisting of the plane of incident is defined as p-polarization. First, the electric field is separated into each component. Subsequently, the field of the reflected wave and the transmitted wave are obtained using the Fresnel equation, and later the entire field is obtained by adding them all.

In the case of TIR, since $\sin {\theta _c} = {n_{ti}}$, the value $n_{ti}^2 - si{n^2}{\theta _i}$ becomes less than 0, and the reflection coefficient becomes complex number. However, since it becomes ${r_ \bot }r_ \bot ^\ast{=} {r_\parallel }r_\parallel ^\ast{=} 1$, all energy generated is reflected. The transmitted electric field over the interface of high RI and low RI layers appeared. In other words, even when TIR occurs, a finite electromagnetic wave exists in the low RI layer. This can be explained by considering the fact that the electric field cannot be discontinuous at the boundary between the high RI and low RI layers. The incident waves and reflected waves must satisfy the electromagnetic boundary conditions at the interface. Therefore, in this region (in ${n_3}$) the electric and magnetic fields are nonzero and there is an energy transfer from the high RI to low RI layers. Indeed, there is a wave in the low RI layer that satisfies the continuity of the tangent component of the electric field $\vec{E}$ and has the same frequency as the incident wave, called the evanescent wave. The electric field behavior in ${n_3}$ (i.e., in the transmitted region) is given by [32]:

Under the TIR condition $\sin {\theta _t} > 1$, $\cos {\theta _t}$ = $\sqrt {1 - \textrm{si}{\textrm{n}^2}{\theta _t}} $ = imaginary number. Since ${k_t}$ is a real number, ${k_t}\cos {\theta _t}$= imaginary number = Ai. Therefore, ET becomes.

 

Fig. 2. (a) Light propagation from high RI layer (n₁) to low RI layer (n₃). (b) Exponentially decaying evanescent field on interface between n₁ and n₃.

Download Full Size | PDF

For ${\theta _i} > {\theta _c}$, penetration depth is a real number, and for ${\theta _i} = {\theta _c}$, infinity depth can be penetrated without attenuation of amplitude. As the angle of incidence gradually increases, the penetration depth decreases, and as the angle of incidence approaches 90°, δ follows the following equation.

Thus, the δ of the evanescent wave ranges from infinity to about 1/4 wavelength depending on the angle of incidence. Furthermore, in Eq. (7), the evanescent wave proceeds only in the + x direction along the interface. As explained above, when light travels from a high RI layer to a low RI layer (i.e., at the interface where TIR occurs), waves do not propagate beyond the interface. However, if a high RI layer (i.e., ${n_2}$) exists above the low RI layer, the TIR phenomenon is frustrated due to the evanescent wave coupling effect, and the light appears in the upper high RI layer. This phenomenon is defined as frustrated TIR (FTIR) (see Fig. 1(b)) [34]. FTIR is also designated as an optical tunneling effect, which is similar to the quantum mechanical tunneling effect.

Also, the thickness of the low RI layer is much lower than the wavelength of the incident light. The evanescent wave electric field intensity has not become zero at the interface between the ${n_2}$ and ${n_3}$ [35]. The gap is so small that the evanescent electric field continues in the ${n_2}$ region. Under this condition, the evanescent wave appears in the ${n_2}$ region as a real wave, called frustrated total internal reflection. When the ${n_1}$ ≈ ${n_2}$ > ${n_3}$ condition is met, TIR in ${n_1}$ medium is almost negligible, and a lot of the incident light appears in the ${n_2}$ region. That is, the thinner the thickness and the lower the RI of the low RI layer, the greater the light extraction effect. Consequently, the evanescent waves are efficiently transformed into light that propagates in air, and this principle is used in the light extraction process in TEOLED devices.

3.4 Optical simulation for TEOLED device with low RI layer

To maximize the light efficiency of a TEOLED device, the thickness condition of the low RI layer in the APA TFE structure in a blue TEOLED device was investigated through optical simulation. The simulated device structure is shown in Fig. S4. The film with the low RI layer beneath the APA 1.5 dyads encapsulation is shown in Fig. S5(a). We aim to enhance the optical efficiency by placing this film on top of the CPL, as depicted in Fig. S5(b). Since the low RI layer is introduced to reduce the loss of light passing from the semi-transparent cathode to come out of the encapsulation structure, the thickness of the organic materials between the anode and the cathode is fixed at the same as in a normal blue TEOLED device. The optimal conditions for HIL + HTL, ETL, and EML were obtained at 150 nm, 30 nm, and 25 nm in the blue TEOLED device, respectively (see Fig. S6(a)). The thickness of the cathode was also fixed at 20 nm, like a normal blue TEOLED device, in order to adjust the OLED device internal cavity conditions. In the absence of the low RI layer, the optimal thickness of the CPL for high light out-coupling (OC) was calculated to be 55 nm, as shown in Fig. S7. After fixing the thicknesses of all layers between anode and cathode, the optimal thickness of the low RI layer and CPL was simulated, as shown in Fig. 3(a). The thicknesses of the CPL and low RI layers were 48 nm and 15 nm, respectively. The low RI layer is very thin. This is one of the important conditions in order to have a nonzero evanescent electric field at the interface of the low RI layer and the APA TFE. With this condition, the real light wave can be observed in the APA structure even when the incident angle is greater than the critical angle in the CPL (in ${n_1}$) layer.

 

Fig. 3. (a) Simulation results for CPL and low RI layer thickness and (b) current and blue index efficiency simulation results of blue devices with the variation of RI of low RI layer.

Download Full Size | PDF

The optimum thickness conditions of the CPL and low RI layer of the blue TEOLED device obtained by the optical simulation of a strong micro-cavity device were used for the further analysis. With this 15 nm thick low RI layer, CE and BI efficiency were examined with respect to the RI variation. Figure 3(b) shows the calculated result by varying the RI from 1.0 to 2.0 in 0.1 steps. As the RI of the low RI layer increases, the CE gradually increases till the RI value reaches to 1.3 and then starts decreasing again above the RI value of 1.3. To remove the effect of luminosity factor, the blue index value (BI, which is the ratio of current efficiency to the value of the CIE y coordinate) was also validated with the variation of RI (see Fig. 3(b)). The BI value shows the same trend as the CE. Without the low RI layer, simulated CE and BI values were 3.19 cd/A, and 76.34 cd/A/y, respectively. With this condition, the simulated electroluminescence (EL) emission peak (λ_max) was observed at 458 nm, and the full width at half maximum (FWHM) value was 18 nm. When a low RI layer is inserted, the CE, BI, λ_max, and FWHM were observed at 3.88 cd/A, 95.19 cd/A/y, 459 nm, and 15 nm, respectively. The CE and BI values are improved by 22% and 25% with the Low RI layer, respectively. The reflection due to the difference in RI was enhanced, and an additional micro-cavity effect was generated. Thus, the output light is amplified due to the augmented high light OC arrangement. The most noticeable result is the narrowing of the FWHM from 18 to 15 nm with the insertion of the low RI layer in the TFE structure. The optical simulation results with and without a low RI layer were summarized in Table 1.

Table 1. Calculation and measurement values of device characteristics according to the presence or absence of low RI layer.

View Table

As well, the simulations of green and red devices were also carried out to confirm the effectiveness of this low RI layer. Except the material used for the EML, green and red devices were simulated by considering the same structure as a blue TEOLED device. Simulated device structures and their optimum thickness were provided in Fig. S4. The optimal HIL + HTL and ETL thicknesses for the green TEOLED device were 175 nm and 45 nm, respectively (see Fig. S6(b)). Further, the optimal thicknesses of the CPL with and without the low RI layer were 55 nm (Fig. S9(a)) and 57 nm (Fig. S8(a)), respectively. The best low RI layer for the green TEOLED device was confirmed to be at 1.3 (see Fig. S8(b)). When a low RI layer of 17 nm was inserted, the CE before and after using the low RI layer was improved by about 29% to 94.73 cd/A and 122.56 cd/A, respectively (Fig. S8(a), (b)). The FWHM narrows from 21 nm to 16 nm (Fig. S8(c)).

In the case of a red TEOLED device, the optimal HIL + HTL and ETL thicknesses were 220 nm and 60 nm, respectively. The optimal CPL thicknesses with and without the low RI layer were 70 nm (Fig. S8(a)) and 60 nm (Fig. S9(b)), respectively. The optimal RI value and the optimal thickness were confirmed to be 1.4 (Fig. S8(b)) and 20 nm in this TEOLED device, respectively. The simulated CE of a red TEOLED device was found to be 53.85 cd/A and 71.45 cd/A with and without a low RI, respectively. The efficiency value increased by about 33%. The FWHM value of the EL spectrum was changed from 24 nm to 21 nm (see Fig. S8(d)). From these results of the simulation, we believe that the predicted green and red devices will also have the effect of improving light efficiency with the insertion of a low RI layer. All simulation results on green and red TEOLED devices were summarized in Table S2 (Supplement 1).

3.5 Blue TEOLED device characteristics and analysis

In order to understand the precise effectiveness of the low RI layer, blue TEOLED devices were fabricated with and without the low RI layer. The optimal conditions obtained by the optical simulation for the low RI layer were considered and applied for the actual device fabrication.

The structure of fabricated devices is shown in Fig. S4. The measured CE of a blue TEOLED device without and with a low RI layer, as shown in Fig. 4(a), were 3.19 cd/A and 3.92 cd/A, respectively. The CE of with a low RI layer device was increased by 23%. Measured BI values without and with a low RI layer were 74.98 cd/A/y and 94.63 cd/A/y, respectively; that increases by about 26% with a low RI layer. The efficiency increase rate results in the fabricated blue TEOLED are in good agreement with the simulated result. Certainly, these results validate that the simulation and measurement results are consistent. The current density and luminance with respect to the voltage of the blue device are shown in Fig. S10. The current densities of both devices with and without a low RI layer show almost similar behavior. However, the luminance is increased when the low RI layer is used due to its better efficiency. With the low RI layer, the λ_max of the EL spectrum was slightly red shifted by about 1 nm, see Fig. 5(c), (d). The FWHM was narrowed from 19 nm to 16 nm, as expected. Further, measured device spectra were slightly shifted to a longer wavelength by about 1 nm than the simulation result, but both devices had almost the same values. From these results of the blue TEOLED device, it has been confirmed that the OC efficiency of the blue device, and the micro-cavity effect were enhanced by the addition of the low RI layer. The device characteristics are summarized in Table 1. Having acquired adequate knowledge of the importance of a low RI layer in improving the efficiency values, the cause for the light loss in the device was examined by the mode analysis optical simulation technique.

 

Fig. 4. (a) CE, (b) BI, (c) EL spectrum, and (d) normalized EL spectrum of without and with low RI layer TEOLED device.

Download Full Size | PDF

 

Fig. 5. (a) Light loss portion (out-coupling (OC), absorption loss (AL), waveguide (WG), and evanescent-coupling (EC)) and (b) polar plot of without and with low RI layer.

Download Full Size | PDF

For the mode analysis optical simulation scheme, the default optics setting was 458 nm for the blue color. In general, when a direct current is applied to an OLED device, the light is emitted from the emitting layer at various angles and exists in various modes. When the low RI layer is present, the OC remarkably increases from 0.18 to 0.21 and the WG mode decreases from 0.17 to 0.14, as shown in Fig. 5(a). The TIR decreases, the WG mode greatly decreases, and the OC portion increases with the low RI layer. Usually, when the light is passing from the CPL to the low RI layer, the light leaking beyond the critical angle is totally reflected in the high RI phase and then reflected back again. Hence, the absorption-loss (AL) mode, caused mainly by the cathode, is slightly increased from 0.14 to 0.15 through more multiple light pathways. However, the unrelated evanescent-coupling (EC) mode is similar and varies slightly from 0.51 to 0.50. Additionally, the intensity of the emitted light from the front was confirmed through a polar plot. Figure 5(b) shows that the intensity has increased significantly when using the low RI layer. These results confirm that the micro-cavity effect was amplified by the difference in refractive index, thereby amplifying the light and causing more light to be emitted out. In addition, the FWHM of the EL spectrum was reduced, and the purity was improved. Further, we checked whether similar observations are observed in green and red TEOLED devices. For the green and red devices, the OC mode increases and the WG mode decreases significantly (see Fig. S10). Changes in the light loss portion according to optical simulation results for the green and red TEOLED devices are summarized in Table S3 (Supplement 1). These results for green and red TEOLED devices are in good agreement with the blue TEOLED device.

4. Conclusion

We suggest that the improvement in light efficiency of TEOLED devices caused by the insertion of a low RI layer is due to the presence of evanescent waves in the low RI layer and the improved micro-cavity conditions. The difference in RI between CPL and Al₂O₃ in APA TFE causes light trapping at the interface, where light loss occurs. However, inserting a low RI layer between the CPL and Al₂O₃ of the APA TFE interface can change the direction of the light reflected by the evanescent wave, allowing more light to be extracted. Blue TEOLED devices were simulated and fabricated with and without the insertion of a low RI layer of about 1.28 at the interface between CPL and Al₂O₃. With the use of a low RI structure, the CE and BI values of the blue TEOLED device were improved by 23% and 26%, respectively. Hence, the ratio of WG mode was decreased by a low RI layer, and the trapped light could pass through to the outside, which improved the efficiencies of devices. In conclusion, we suggest that the results demonstrated here are useful for the efficient light extraction of the TFE structure for flexible organic display applications.