Tunable microcavities in organic light-emitting diodes by way of low-refractive-index polymer doping

(1)

Tunable microcavities in organic light-emitting diodes by way

of low-refractive-index polymer doping

An-Kai Ling

a,b

, Chun-Hao Lin

a,c

, Hsun Liang

a,c

, Fang-Chung Chen

a,b,⇑

a

Department of Photonics, National Chiao Tung University, Hsinchu 30010, Taiwan

b

Institute of Display, National Chiao Tung University, Hsinchu 30010, Taiwan

c

Institute of Electro-optical Engineering, National Chiao Tung University, Hsinchu 30013, Taiwan

a r t i c l e

i n f o

Article history:

Received 9 September 2014

Received in revised form 6 October 2014 Accepted 10 October 2014

Available online 28 October 2014 Keywords: Organic Light-emitting Microcavity Refractive index Resonant

a b s t r a c t

A method for enhancing the light out-coupling efficiency of organic light-emitting devices (OLEDs) has been demonstrated by blending a low-refractive-index polymer, poly(2,2,3,3,3-pentafluoropropyl methacrylate) (PPFPMA), into the emission layer. The resonant wavelength of the weak microcavity devices blueshifted accompanied with a decrease in refractive indices of the light-emitting layers after the addition of PPFPMA. Stronger directed emission toward the surface normal was obtained when the resonant wavelength became closer to the peak wavelength of intrinsic emission spectrum of the organic emitters. The luminous efficiency of the devices was enhanced by more than 20%. The results suggest that the microcavity properties of the OLEDs can be tunable through blending low-refractive-index materials.

1. Introduction

Organic light-emitting diodes (OLEDs) are playing an increasingly conspicuous role in the next generation display and solid-state lighting technologies because of their high efficiencies, low cost, and their potential to act as large-area and flexible devices[1–3]. Meanwhile, nearly 100% internal quantum efficiencies have been realized by doping phosphorescent molecules to harvest triplet exci-tons[4,5]. Nevertheless, because of the mismatch of the refractive indices, a large proportion of the generated pho-tons is trapped in the devices and the light out-coupling efficiency (

g

out) of conventional OLEDs is limited at ca.

20%[6]. Therefore, substantial enhancement in the light extraction efﬁciencies will be necessary if we are to further improve the device efﬁciencies.

Many approaches, including using high-refractive-index substrates [3,7], microlense arrays [8,9], photonic crystals [10], and plasmonic nanostructures [11], have been studied to enhance

g

out. Meanwhile, the conventional

structure of OLEDs naturally behaves as a weak microcav-ity because the dimension of their functional layers is sim-ilar to the wavelength of the emitting photons. While both electrodes exhibit high reﬂectivities, strong microcavity effects can be observed. Therefore, the microcavity effects redistribute the optical modes in the devices, and the

g

out

can be enhanced along the normal emission direction under certain conditions[12–18]. For instance, better light extraction can be achieved by setting the resonant wave-length of the cavity near the peak wavewave-length (kem) of

the intrinsic emission spectrum [S(k)] of the organic emitters[13,14]. To the best of our knowledge, however, most studies adjust the optical cavity length (L) of the microcavities through altering the thickness (d) of the organic layers [17,18]. No report tunes the values of L, which are equal to the products of the refractive indices [n(k)] and d, by altering the refractive indices of the

http://dx.doi.org/10.1016/j.orgel.2014.10.013

⇑ Corresponding author at: Department of Photonics, National Chiao Tung University, Hsinchu 30010, Taiwan. Tel.: +886 3 5131484; fax: +886 3 5735601.

E-mail address:[email protected](F.-C. Chen).

Contents lists available atScienceDirect

Organic Electronics

(2)

emission layers. Adjusting the values of L through varying the thicknesses of the active layers, however, inevitably affects other electrical properties, such as the internal electrical ﬁeld strength and the width and location of the recombination zone, resulting in a rather complicated situation for structure designs[19]. An alternative way is to change the refractive indices of the emission layers. In this work, we demonstrated a simple, effective method for enhancing

g

out of polymer light-emitting diodes

(PLEDs) by blending a low-refractive-index polymer, poly(2,2,3,3,3-pentaﬂuoropropyl methacrylate) (PPFPMA), into the emission layer. The microcavity effects of the devices can be optimizing by tuning the n(k) of the emis-sion layer with the amount of PPFPMA. As a result, the luminous efﬁciency incorporating PPFPMA was enhanced by approximately 20%.

2. Experimental

Fig. 1shows the device structure and the chemical struc-tures of the materials used in this work. For device fabrica-tion, indium tin oxide (ITO)-coated glass substrates were cleaned with detergent, followed by sequential sonication in deionized water, acetone, and isopropanol. They were then dried in an oven overnight. Prior to use, the substrates were further treated with UV-ozone. Poly(3,4-ethylenedioxythio-phene)–poly(styrenesulfonate) (PEDOT:PSS) was spin coated onto the ITO substrates and then baked at 120 °C for 1 h. The emission layer comprised poly(vinylcarbazole) (PVK), 2-(4-biphenyl)-5-(4-tert-butyl-phenyl)-1,3,4-oxadiazole (PBD), and tris[2-(4-tolyl)phenylpyridine]iridium [Ir(mppy)3] at a

weight ratio of PVK:PBD:Ir(mppy)3= 70:29:1[20–22]. The

emission layer was deposited on top of the PEDOT:PSS layer from a solution of 1,2-dichlorobenzene inside a N2-ﬁlled

glovebox. The polymer ﬁlm was then annealed at 80 °C for 30 min. To complete the devices, a cathode comprising 30 nm Ca and 100 nm Al was thermally deposited under a vacuum of ca. 5 106_{Torr. The control devices were}

deﬁned as those prepared without PPFPMA. For the devices containing PPFPMA, the low-refractive-index polymer was blended into the emission layer at various weight of 0.1– 0.5% (w/w). For example, 0.1% PPFPMA indicated that the weight ratio of the components in the emissive layer was PVK:PBD:Ir(mppy)3:PPFPMA = 70:29:1:0.1. The refractive

index of PPFPMA was 1.140, which was measured by using an ellipsometer (Sopra GES-5). The thicknesses of the

emission layer prepared with and without the addition of PPFPMA were precisely controlled and all of them were almost identical (ca. 70 nm); the thicknesses were obtained by using an atomic force microscope (AFM). The electrical characteristics of the PLEDs were recorded by a Keithley 2400 source measure unit. The brightness and the electrolu-minescent (EL) spectra were measured by using a PR650 SpectraScan colorimeter. The photoluminescent (PL) spectra were obtained using an Ocean Optics HR4000CG-UV-NIR spectrometer. A N2-laser was used as the light source. The

reﬂectivity spectra were obtained using a Perkin Elmer UV– Vis Lambda 650 spectrometer. The viewing angle depen-dence of luminance was determined using a ConoScope™ (Autronic-Melchers, GmbH), where a measuring cone of 80° over an azimuthal angle range of 360° could be obtained. The thin ﬁlm morphology was analyzed using a DI 3100 series AFM.

3. Results and discussion

Fig. 2(a) shows the current density–brightness–voltage (J–B–V) curves of the PLEDs containing various amount of PPFPMA. The turn-on voltage (VT), deﬁned for a luminance

of 0.1 cd/m2_{, was 4.6 V for the control device. The devices}

prepared with 0.1% and 0.2% PPFPMA exhibited almost identical electrical characteristics, suggesting that the addi-tion of PPFPMA did not affect the electrical properties of the devices signiﬁcantly. However, as the concentration of PPFPMA was increased beyond 0.2%, we could see that the current density started to decrease obviously. The values of VTbecame 4.8 V for the devices containing 0.3% PPFPMA,

respectively. The increase in VTcould be attributed to the

insulating nature of PPFPMA. The electrical properties of the devices fabricated under different conditions are sum-marized inTable 1.Fig. 2(b) and (c) display the luminance efﬁciency–current density (

g

–J) and power efﬁciency–cur-rent density (LP–J) curves of the PLEDs prepared with and

without PPFPMA. The maximum

g

and LPof the control

device were 20.5 cd/A and 9.3 lm/W, respectively. After adding PPFPMA into the emission layer, substantial enhancement in the luminance efﬁciency and power efﬁ-ciency could be obtained. Compared with the control device, the peak

g

of the device containing 0.2% PPFPMA increased approximately by 22% to the 25.1 cd/A. Mean-while, the LPwas also enhanced to 11.4 lm/W while the

driving voltage was not signiﬁcantly affected by adding

(3)

PPFPMA.Fig. 2(d) shows the EL spectra for the PLEDs fabri-cated with and without PPFPMA. The peak wavelength of the devices was 516 nm and stayed unchanged after the addition of PPFPMA, suggesting that the device enhance-ment was not due to the spectral change.

Because the optimal thickness of the photoactive films is in the order of a wavelength, a PLED can be considered as a one-dimensional microcavity. A cavity effect is governed mainly by wide-angle interference within the device; how-ever, Fabry–Perot type multiple-beam resonance effect can also be present due to the low but finite reflection at glass/ ITO interfaces[23]. In such structures, directly outgoing beams of the emission interfere with the beams reflected from the top metal cathode, thereby influencing the out-coupling intensity. From the optical point of view, a PLED can be treated as a planar Fabry–Perot microcavity. The res-onant modes of a cavity should satisfy the condition that the phase change during one round trip is a multiple of 2

p

[24]. In other words, the following Eq. holds:

X

i

4

p

cos b

k diniðkÞ

u

topðb; kÞ

u

botðb; kÞ ¼ 2m

p

ð1Þ

where k is the resonance wavelength; m is an integer (0, 1, 2,. . .) that deﬁnes the mode number;

u

top(k),

u

bot(k) are the

wavelength-dependent phase changes upon reﬂection from the top and bottom mirrors, respectively. The sum-mation is performed over constituent organic layers inside the microcavity with thicknesses diand refractive indices

ni(k). The viewing angle is also incorporated, where b is

the internal angle inside the cavity and is related to the external viewing angle

a

through the Snell’s law: nair

sin(

a

) = ncavsin(b). From Eq.(1), it is obvious that the

res-onance wavelengths of a microcavity can be tuned by altering the optical path length, i.e., the refractive indices and thicknesses of the organic layers, including those of the PEDOT:PSS buffer and the emission layers. In this work, the thicknesses of PEDOT:PSS and active polymer layers were kept constant during our investigation. Therefore, we inferred that the resonance wavelengths of the cavity were changed after blending the low-refractive-index polymers into the emissive layer. As a result, the

g

outwas

enhanced along the normal emission direction after the nature of the weak microcavity was modiﬁed.

Fig. 2. (a) J–L–V; (b) LE–J; (c) LP–J characteristics of the devices fabricated with various amount of PPFPMA. (d) Normalized EL spectra of the devices

prepared with and without the addition of PPFPMA; the inset shows the enlarged spectra in the wavelength range from 500 to 600 nm.

Table 1

Device characteristics of the PLEDs prepared with various PPFPMA concentrations. Concentration (%) (w/w)a

Turn-on voltage (V)b

Max. luminance (cd/m2

) Max. luminous efﬁciency (cd/A) Max. power efﬁciency (lm/W)

0 4.5 1464 20.5 9.3

0.1 4.6 1573 24.0 10.8

0.2 4.6 1689 25.1 11.4

0.3 4.8 1339 23.4 9.6

a

PVK:PBD:Ir(mppy)3:PPFPMA = 70:29:1: weight ratio of PPFPMA. b

(4)

In order to characterize the properties of the microcavi-ty, the reflectivity spectroscopy was used [19,24]. Fig. 3 shows the reflectivity spectra in normal direction for the PLED devices prepared with different amount of PPFPMA. We observed two dips around 450 and 560 nm for the control device. Note that the absorption of ITO resulted in a dramatic decrease of the reflectivity at 350 nm. In princi-ple, the presence of a dip in the reflectivity spectrum can be contributed to the interference effect and its location

indicates the resonance wavelength. Further, because the weak microcavity of the PLEDs acts as a ﬁlter, the photons whose wavelength range falls outside of the dips are trapped in the cavity[19]. In our cases, the normal direction resonance wavelength kR(0°) was 590 nm for the control

device and shifted to 570 and 560 nm for the device containing 0.1% and 0.2% PPFPMA, respectively (Fig. 3). Moreover, the PL spectrum of the light-emitting layer inves-tigated in this work, which is the intrinsic spectral distribu-tion of the organic emitters, is also shown inFig. 3. The kR

(0°) of the 0.2% PPFPMA device shifted by 30 nm relative to that of the control device and became more closed to the intrinsic emitting wavelength of the organic emitters. Because of the interference effect, substantial enhancement in forward luminance was achieved, resulting in a higher light out-coupling efﬁciency. Further, the EL spectra as dis-played inFig. 2(d) suggested that the portion of photons at 516 nm relatively increased after the use of PPFPMA. The intensity of the PL shoulder, which is centered at ca. 540 nm, relatively decreased because of the change of reso-nance conditions in the weak microcavities. These spectral results also support that the addition of the PPFPMA modi-ﬁed the microcavity properties of the devices.

The luminance of the PLEDs measured on Conoscope™ as a function of various viewing angles was shown in Fig. 4.Fig. 4(a)–(c) reveal the luminance contour plots of the devices prepared with various amount of PPFPMA.

Fig. 3. Reﬂectivity spectra for the devices prepared with various amount of PPFPMA. The PL spectrum of the light-emitting polymer is also displayed.

Fig. 4. Viewing angle dependences of the luminance of the PLEDs prepared (a) without PPFPMA, (b) with 0.1% PPFPMA, and (c) with 0.2% PPFPMA at constant current density of 0.2 mA/cm2

. The inclination and azimuthal angles were measured from 0° to 80° and 0° to 360°, respectively. The corresponding luminance (normalized to the maximum intensity) at 0° and 180° azimuthal angle and at different inclination angles was also shown in (d). Note that the luminance distribution was noncircular in symmetry due to the rectangular shape of the PLEDs. In other words, a Lambertian proﬁle is corresponding to a ﬂat line for all angles.

(5)

Inclination and azimuthal angles were measured from 0° to 80°and 0° to 360°, respectively. The luminance distribu-tion was noncircular in symmetry because of the rectangu-lar shape of the PLEDs; the active area was 0.1 cm2 (0.2 cm 0.5 cm)[9]. The corresponding luminance at 0° to 180° azimuthal angle and at different inclination angles was also revealed inFig. 4(d); each curve was normalized to the maximum intensity of itself. Interestingly, the emis-sion pattern exhibited the highest intensity at an off-axis angle of 40–50° and a relatively weak intensity in the forward direction. Similar patterns have been reported previously and attributed to the microcavity effect[19]. The changes of the angle profiles were due to the shift of the reflectivity dip with the viewing angle both in wave-length and in magnitude corresponding to their overlaps with the emission spectrum of the organic emitters[19]. In addition, from the reflectivity spectra, we could also realized that the unusual pattern was due to the large mis-match between the resonance conditions and the intrinsic light-emitting wavelength of the polymers. According to Eq.(1), we could also see that the location of the dip in reflectivity varies with the internal angle inside the cavity, which is closely related to the viewing angle. For the con-trol device, the wavelength minimum at 590 nm blue-shifted to 545 nm while the viewing angle changed from 0° to 40°. Therefore, the overlap of the dip in reflectivity with the emission spectrum, in which the peak wavelength was 510 nm, became less at 0° viewing angle (Fig. 3), resulting in a lower luminance. On the other hand, the dip shifted closer to the intrinsic emission wavelength at higher angles, resulting in a peak luminance at viewing

angles of 40–50°. As displayed in Fig. 4(d), the device prepared with 0.2% PPFPMA displayed stronger directed emission toward the surface normal and showed emission patterns more close to the Lambertian distribution. In short, the angle dependence of the device emission further conﬁrmed the importance of the microcavity effects.

The interface between the polymer and the cathode plays an important role in determining the electrical characteristics of devices [2]. In order to investigate the influence of PPFPMA on the polymer film, the surface mor-phology of the composite films was analyzed by AFM. The height images of the active films before and after the addi-tion of PPFPMA are displayed inFig. 5. We observed that the surfaces of these films were very smooth. The surface roughness of the polymer film containing 0.1% PPFPMA was even decreased slightly. The almost identical morphol-ogy suggested no serious phase separation. Therefore, the addition of PPFPMA had little influence on the morphology of thee the films, revealing that the device improvements should be largely due to the optical effects.

Recently, phosphorescent dyes have been reported to have preferred horizontal dipoles, which could result in higher coupling efficiencies [25,26]. In our present work, we cannot completely rule out the possibility that the addition of PPFPMA might change the orientation of the dipoles. However, we consider the chance should be very limited because the amount of PPFPMA in the active films was low (0.1–0.3% in weight). We could not find any signa-ture of dipole changes from our preliminary results, such as the absorption spectra. Further studies, such as angle-dependent PL measurements, are ongoing.

Fig. 5. AFM images of light-emitting polymer ﬁlms prepared (a) without PPFPMA, (b) with 0.1% PPFPMA, and (c) with 0.2% PPFPMA. The image dimension is 5 5lm.

(6)

4. Conclusions

We have demonstrated that the out-coupling efficiency of PLEDs can be enhanced by blending low-refractive-index polymers into the emission layer. Stronger directed emission toward the surface normal was obtained, leading to an efficiency enhancement of about 22%. The device improvement originated from the optimization of optical microcavity effects by adjusting the refractive indices of the emission layers after the addition of PPFPMA. The device prepared with 0.2% PPFPMA exhibited almost iden-tical electrical characteristics and showed emission pat-terns more close to the Lambertian distribution. While the tuning of resonance conditions in OLEDs was achieved largely through changing the thickness of the functional layers, this work suggests that it can be also accomplished by altering the refractive indices of the emission layers. This work provides a very convenient way to tune the opti-cal properties of OLEDs without significant affecting the other electrical properties, such as the internal electrical field and the width and location of the recombination zone. We anticipate the results reported herein might open a new avenue toward higher device performance.

Acknowledgements

We thank the Ministry of Science and Technology of Taiwan (Grant Nos.: NSC 102-ET-E-009-005, NSC 103-2923-E-009-001-MY3, and NSC 101-2628-E-009-008-MY3) and the Ministry of Education of Taiwan (through the ATU program) for ﬁnancial support.

References

[1]B.W. D’Andrade, S.R. Forrest, Adv. Mater. 16 (2004) 1585–1595.

[2]H.B. Wu, J.H. Zou, F. Liu, L. Wang, A. Mikhailovsky, G.C. Bazan, W. Yang, Y. Cao, Adv. Mater. 20 (2008) 696–702.

[3]S. Reineke, F. Lindner, G. Schwartz, N. Seidler, K. Walzer, B. Lüssem, K. Leo, Nature 459 (2009) 234–238.

[4]C. Adachi, M.A. Baldo, M.E. Thompson, S.R. Forrest, J. Appl. Phys. 90 (2001) 5048–5051.

[5]E.L. Williams, K. Haavisto, J. Li, G.E. Jabbour, Adv. Mater. 19 (2007) 197–202.

[6]N.C. Greenham, R.H. Friend, D.C. Bradley, Adv. Mater. 6 (1994) 491– 494.

[7]T. Nakamura, N. Tsutsumi, Noriyuki, H. Fujii, J. Appl. Phys. 97 (2005) 054505.

[8]C.F. Madigan, M.-H. Lu, J.C. Sturm, Appl. Phys. Lett. 76 (2000) 1650– 1652.

[9]W.K. Huang, W.S. Wang, H.C. Kan, F.C. Chen, Jap. J. Appl. Phys. 45 (2006) L1100–L1102.

[10]L. Petti, M. Rippa, R. Capasso, G. Nenna, A.D.G.D. Mauro, M.G. Maglione, C. Minarini, Nanotechnology 24 (2013) 215206. [11]J. Feng, T. Okamoto, Opt. Lett. 30 (2005) 2302–2304.

[12]V. Bulovic´, V.B. Khalﬁn, G. Gu, P.E. Burrows, D.Z. Garbuzov, S.R. Forrest, Phys. Rev. B 58 (1998) 3730–3740.

[13]J. Lee, N. Chopra, F. So, Appl. Phys. Lett. 92 (2008) 033303. [14]R.H. Jordan, A. Dodabalapur, R.E. Slusher, Appl. Phys. Lett. 69 (1996)

1997–1999.

[15]J. Lee, N. Chopra, D. Bera, S. Maslov, S.H. Eom, Y. Zheng, P. Holloway, J. Xue, F. So, Adv. Energy Mater. 1 (2011) 174–178.

[16]C.L. Lin, H.C. Chang, K.C. Tien, C.C. Wu, Appl. Phys. Lett. 90 (2007) 071111.

[17]A. Dodabalapur, L.J. Rothberg, R.H. Jordan, T.M. Miller, R.E. Slusher, J.M. Philips, J. Appl. Phys. 80 (1996) 6954–6964.

[18]S.K. So, W.K. Choi, L.M. Leung, K. Neyts, Appl. Phys. Lett. 74 (1999) 1939–1941.

[19]H. Rooms, D. Hermes, S. Harkema, C. Tanase, T.V. Mol, P.W.M. Blom, J. Wilson, J. Photon. Energy 1 (2011) 011024.

[20]F.C. Chen, S.C. Chang, G. He, S. Pyo, Y. Yang, M. Kurotaki, J. Kido, J. Polym. Sci. Part B: Polym. Phys. 41 (2003) 2681–2690.

[21]F.C. Chen, S.C. Chien, Y.S. Chen, Appl. Phys. Lett. 94 (2009) 043306. [22]F.C. Chen, S.C. Chien, S.W. Lee, Electrochem. Solid-State Lett. 11

(2008) J50–J53.

[23]K.A. Neyts, J. Opt. Soc. Am. A 15 (1998) 962–971.

[24]S.A. Ahkmanov, S.Y. Nikitin (Eds.), Physical Optics, Oxford

University, 1997.

[25]S.Y. Kim, W.I. Jeong, C. Mayr, Y.S. Park, K.H. Kim, J.H. Lee, C.K. Moon, W. Brütting, J.J. Kim, Adv. Funct. Mater. 23 (2013) 3896–3900. [26]K.H. Kim, C.K. Moon, J.H. Lee, S.Y. Kim, J.J. Kim, Adv. Mater. 26 (2014)