Improving the light extraction efficiency of polymeric light emitting diodes using two-dimensional photonic crystals

Organic Electronics 7 (2006) 222–228 www.elsevier.com/locate/orgel

Improving the light extraction efficiency of polymeric light emitting diodes using two-dimensional photonic crystals A.M. Adawi a, R. Kullock a, J.L. Turner a, C. Vasilev a, D.G. Lidzey A. Tahraoui b, P.W. Fry b, D. Gibson c, E. Smith d, C. Foden d, M. Roberts d, F. Qureshi d, N. Athanassopoulou d

a,*

,

a

d

Department of Physics and Astronomy, The University of Sheffield, Hicks Building, Hounsfield Road, Sheffield S3 7RH, United Kingdom b Department of Electronic and Electrical Engineering, The University of Sheffield, Mappin Street, Sheffield S1 3JD, United Kingdom c Applied Multilayers Ltd., West Stone House, West Stone, Berry Hill Industrial Estate, Droitwich Worcestershire WR9 9AS, United Kingdom Cambridge Display Technology, Greenwich House, Madingley Road, Madingley Rise, Cambridge CB3 0TX, United Kingdom Received 23 December 2005; received in revised form 17 February 2006; accepted 21 February 2006 Available online 22 March 2006

Abstract We have fabricated light emitting diodes based on a conjugated polymer, in which a planarized two-dimensional photonic crystal (PC) was inserted between the glass substrate and the ITO anode. Planarized PCs were fabricated into a highindex layer via interference lithography, followed by dry-etching and the spin-casting of a low-index glass. We characterize the electroluminescence (EL) emission from devices containing a PC, and compare this with photoluminescence (PL) generated from within the same structure. We show that LEDs incorporating the PC have an increased EL external quantum efficiency of (2.3 ± 1.0) times compared to a standard non-patterned control. This efficiency increase is in excellent agreement with PL measurements on similar structures, which also demonstrate relative increases in external emission intensity of 2.3 times.  2006 Elsevier B.V. All rights reserved. Keywords: Polymeric light emitting diodes; Photonic crystal

Over the last decade polymeric light emitting diodes (PLEDs) have received significant attention from both academic and industrial research groups

*

Corresponding author. E-mail addresses: a.adawi@sheffield.ac.uk (A.M. Adawi), d.g.lidzey@sheffield.ac.uk (D.G. Lidzey).

as systems for use in flat panel display technologies due to their low fabrication cost, low power consumption, wide viewing angle and fast switching times [1–5]. Improving the external efficiency of organic LEDs is a critical issue in addressing their commercial uptake, as this plays a significant role in determining their useful operational lifetime. Standard organic LEDs are composed of a glass

1566-1199/$ - see front matter  2006 Elsevier B.V. All rights reserved. doi:10.1016/j.orgel.2006.02.002

A.M. Adawi et al. / Organic Electronics 7 (2006) 222–228

substrate coated with a transparent anode layer such as indium tin oxide (ITO), one or more of organic layers and a low work-function metallic cathode. In such a configuration only 20% [6] of the light generated within the device can usefully escape to an external observer, with the remainder of the light being trapped in optical modes in the substrate, or in the ITO/organic layers [6]. Several methods have been explored to extract more light from organic LEDs, including the introduction of scattering centres within the device to overcome the critical angle condition [7,8] the incorporation of microcavities [9], the use of microlenses [10], the insertion of Bragg diffraction gratings [11– 13], or the use photonic crystals (PCs) within the device [14–22]. There are several positions in which such a grating or a photonic crystal can be inserted into an LED; these include placing it between the metallic cathode and the active organic [22], between the active region and the ITO layer [14– 17] or between the ITO layer and the glass substrate [18–21]. The latter approach is probably the most promising for practical applications, as it does not result in a modification in the electronic properties of the device. In this letter, we demonstrate the effect of inserting a deep photonic crystal (PC) between the glass substrate and the ITO layer. To create a flat surface on which the device is deposited, we use a simple spin-casting technique to planarize the surface of the grating on which an ITO anode is deposited. This approach permits us to improve the external luminous EL efficiency of red-emitting PLEDs by a factor of (2.3 ± 1.0) times (defined within an external emission cone angle of 35). We show that the electrical properties of the patterned

223

PLEDs are identical to un-patterned control devices, and use photoluminescence (PL) measurements to confirm that the improvement in external EL efficiency results from purely optical effects. Furthermore, we demonstrate that there are no unwanted colour shifts associated with utilizing such a PC structure. We argue that this approach to improving PLED efficiency could also be applied over large areas by using low-cost embossing techniques to create suitably patterned substrates. A PC is a structure in which a periodic variation in refractive index occurs at the scale of the wavelength of light in one or more direction [23]. If the refractive index contrast of the PC is sufficiently large, it can result in the formation of a photonic bandgap (a range of frequencies in which the propagation of light is forbidden) [23]. In principle, there are two methods to use PCs to improve the external efficiency of organic LEDs. One approach is to match the trapped waveguide modes within the LED to the bandgap of the photonic crystal. The waveguided light thus lies within the bandgap of the PC, blocking its propagation in lateral directions within the structure, leaving only the external emission channel for light to exit the device [24]. Such an approach is however difficult to realize in practice, as there are significant material processing problems associated in creating a planarized structure having a sufficiently large refractive index contrast to open a full optical-bandgap. A second approach (which we use here) is to utilize the refractive index periodicity of a PC to diffract waveguide modes above a certain cut-off frequency into externally propagating modes (as schematically illustrated in Fig. 1(c)), which thus improves the

Fig. 1. A schematic of the conjugated polymer LEDs explored in this work. The structure shown in part (a) is a standard device, whilst that shown in part (b) is deposited upon a photonic grating. Part (c) shows how a periodic grating can be used to couple waveguide modes * in an LED into propagating modes and therefore improve optical extraction efficiency. Here, k wg is the wave-vector of a trapped *

*

of the photonic crystal and k k is the wave-vector of the diffracted light. The diffracted waveguide mode, k pc is the reciprocal*lattice vector * light propagates with a wave-vector k , where jk k j ¼ jkj sin h.

224

A.M. Adawi et al. / Organic Electronics 7 (2006) 222–228

extraction efficiency of the device [24]. In Fig. 1(c), * k wg is the component of the* waveguided light parallel to the device plane and k pc is the reciprocal wave-vector   of  the photonic crystal, given by * 2p ^ ^j, where ax and ay are the periodk pc ¼ ax i þ 2p ay icity of the lattice in the x and y directions, respectively. The effect of the photonic crystal on the waveguided modes is to change their in-plane * * * * wave-vector to k k , where k k ¼ k wg  nk pc , where n is an integer number. Such diffracted modes will * escape from the device provided that k k , is less than the wave-vector above which light is totally internally reflected. Recently Lee et al. [18] have made an experimental and theoretical study of the effect of inserting a PC between the glass substrate and the ITO layer in an organic LED based on vapor deposited low molecular-mass materials. Their results indicated that the extraction efficiency increases with increasing depth of the photonic crystal, and also showed that the external efficiency of the device is maximized when the periodicity of the photonic crystal is larger than the (vacuum) wavelength of the emitted light. This allowed them to demonstrate an improvement in external efficiency by a factor of 1.5 times using a 200 nm deep grating. Whilst this is clearly a promising result, we note that it is less than the enhancement factors that have been demonstrated when PC structures have been incorporated into inorganic LEDs. In particular, a PC etched into the surface of an AlGaN LED [25] improved the power output of the device by a factor of 2.5 times. Here, the depth of the PC was 190 nm, with a periodicity approximately 3 times that of the AlGaN emission wavelength. In this paper, we have explored the use of a deep optical grating to enhance the extraction efficiency of a polymer LED. The grating used had a depth of 450 nm with a periodicity approximately twice that of the polymer emission wavelength. Using this approach we demonstrate improvements in external efficiency of a red-emitting polymer LED by a factor of 2.3 times. This enhancement is close to the largest enhancement observed in inorganic devices, and suggests a promising method to gain significant improvements in the external efficiency of organic LED. The devices that were fabricated and evaluated are shown schematically in Fig. 1(a) and (b). Fig. 1(a) shows a standard PLED device, composed of a glass substrate coated with a 150 nm thick layer

of ITO (deposited using closed field magnetron sputtering), a 150 nm thick layer of the conducting polymer PEDOT:PSS (poly(3,4-ethylendioxythiophene)/polystyrene-sulfonic acid) (Baytron P PE FL supplied by H.C. Starck Ltd.) included to facilitate hole injection, and a 70 nm thick layer of an active red polymer (supplied by Covion Organic Semiconductors GmbH). Here, each polymer layer was deposited by spin coating, with the PEDOT: PSS being annealed at 150 C in air for 15 min. The device was then finished by the thermal evaporation of a cathode, composed of a 20 nm thick film of calcium used to aid electron injection, capped by an optically thick (50 nm) film of silver. The cathode was evaporated through a shadow-mask onto the surface of the conjugated polymer, with each individual LED having an active area of 5 mm2. All metal depositions were made at a base-pressure of 108 mbar in a Kurt J. Leskar Spectros deposition system. Fig. 1(b) shows a schematic of an LED whose structure is identical to that shown in Fig. 1(a), however, it is fabricated on a planarized PC. The PC was created by first depositing a 500 nm thick layer of SixNy (n = 1.95) on a 12 · 12 mm2 glass substrate using plasma-enhanced chemical vapor deposition. Interference lithography [26–28] was then used to write a photonic structure into a 1 lm thick layer of a high sensitivity positive photoresist (SPR350 supplied by Chestech Ltd.) spin-cast onto the SixNy layer. Interference lithography was achieved using a HeCd laser (k = 442 nm), which was split into two beams using a beam splitter. Each beam was expanded using a 10· microscope objective lens, and then recombined onto the substrate surface (with the beams having a mutual angular separation of 20). The diameter of the expanded beams was 6 cm, permitting us to micro-pattern the entire PLED substrate. To do this, the photoresist was first exposed to the laser having a power of 5 mW cm2 for 20 s. The sample was then rotated by 90 and the exposure repeated. This resulted in the creation of a square lattice PC having a pitch of 1.2 lm. The photoresist was then developed for 10 s in a developer (MF26 supplied by Chestech Ltd.). This pattern was then transferred into the SixNy layer using a CHF3 coupled plasma-etching technique. Fig. 2(a) shows an AFM image of a patterned SixNy layer following dry-etching. The pattern is comprised of holes having a diameter of 930 nm, with a periodicity of 1245 nm and depth of 430 nm.

A.M. Adawi et al. / Organic Electronics 7 (2006) 222–228

Fig. 2. Part (a) shows an AFM image of the photonic crystal (PC) structure after dry-etching. Part (b) shows the same grating after filling with a spin-on glass material. Part (c) shows a series of topographic cross-sections of the PC anode surface recorded (using an AFM) at various steps in its fabrication.

To planarize the grating, a low refractive index (n = 1.45) spin-on glass (supplied by Allied Signal Inc.) was coated onto the surface and then thermally annealed in air at 280 C for 1 h. As the thickness of the spin-on glass layer was (after annealing) 150 nm, the coating process was repeated three times to totally infill the holes in the SixNy and thus planarize the grating (see the AFM image presented in Fig. 2(b)). This structure was then coated by a 150 nm thick layer of ITO. The internal structure of the ITO-covered PC can be visualized by plotting a series of AFM cross-sections taken at various

225

stages during its fabrication as shown in Fig. 2(c). It can be seen that the spin-on glass largely planarized the SixNy grating and created a surface having a height modulation of ±25 nm. This planarization process was completed by the ITO deposition, which reduced the amplitude of the surface corrugations to ±2 nm. Current–voltage–luminance measurements of the PC LED and control devices were made in an overpressure nitrogen glove-box, having water and O2 background levels less than 0.1 ppm. To measure luminance, a calibrated photodiode was placed above the sample, subtending a cone angle of 35 to the device. The current generated by the photodiode was converted to an equivalent brightness (in units of cd m2) using a pre-determined correction factor obtained using a calibrated Topcon luminance meter. As we show below, there is very little difference in the EL emission spectra between the control devices and the PC LEDs (both at normal incidence and as a function of external viewing angle). Thus the efficiency of the device in terms of cd A1 can be thought of as being a relative measurement of external quantum efficiency. Emission spectra were also measured from both control and patterned devices using a fibre-coupled CCD-spectrograph. We emphasise that both PC LEDs and control devices emitted uniform EL across their active area and were free from so-called ‘black spots’. We have complemented the electrical measurements on the devices using optical excitation to generate photoluminescence (PL) within the structure. Importantly, this permits us to obtain a second measurement of the effect of the periodic patterning on the device emission, and – as we show below – it demonstrates that the external efficiency enhancements observed are purely optical in nature and do not rely on effects relating to charge injection and transport. For PL measurements, we studied structures identical to those shown in Fig. 1(a) and (b), however, the thin calcium layer was omitted to avoid any problems due to photo-oxidation (as all excitation and emission measurements were performed in air). To generate PL, the sample surface was excited at normal incident using a 4 mW 405 nm diode laser, with emission being measured as a function of viewing angle. In Fig. 3(a) we plot the I–V characteristics of both patterned and control devices. It can be seen that both types of device have (as expected) similar I–V characteristics. Fig. 3(b) plots the L–V characteristics of control and patterned devices. It can be

226

A.M. Adawi et al. / Organic Electronics 7 (2006) 222–228

(a)

1.0

100 10 0

2

4

6

8

10

(b)

cdm-2

100

Patterned LED Control device

0.5

0.0

500

10

0

0.3

2

4

6

8

10

600 700 Wavelength (nm)

800

Fig. 4. Normalised EL spectra recorded at normal incidence from a control device (closed circles), and a laterally patterned LED (solid line).

(c)

0.2

cdA-1

Normalised EL intensity

Am-2

1000

0.1 0.0 4

6

8

10

Voltage (V) Fig. 3. Part (a) shows current–voltage curves recorded for control and laterally patterned devices. The luminance versus voltage for the same devices is shown in part (b), which is then used to calculate external emission efficiency as shown in part (c). Note, that all data shown are an average deduced from a large number of measurements made on individual devices as discussed in the text.

seen that both patterned and control devices emit light at voltages greater than 2.5 V. The luminance of the patterned device is higher than that of the control device by a factor of 2.7 times at all voltages. Fig. 3(c) compares the luminous efficiency (in terms of cd A1) as a function of driving voltage for both patterned device and the control. Here, the data presented in Fig. 3(a)–(c) is an average of measurements made from seven identical patterned devices and 10 control devices. It can be seen that there is a clear enhancement by a factor of (2.3 ± 1.0) times in the external efficiency of the patterned LED, which – as indicated by the I–V data shown in Fig. 3(a) – is not related to changes in the electronic properties of the device. Note that the absolute efficiency of the patterned or control devices presented in Fig. 3 is significantly lower than that of state-of-the-art devices. This was due to the fact that the ITO used here was not optimized for OLED operation. Despite this, our results indicate

that the observed improvements in the external efficiency in Fig. 3(c) originate from inserting PC beneath the ITO. In Fig. 4, we plot the normal incidence EL emission spectrum recorded from the patterned device and the control PLED. It can be seen that the EL emission spectrum of both devices are quite similar, indicating that the adoption of a PC structure does not cause an apparent shift in the emission colour of the LED. We have also characterized the CIE (x, y) emission colour-coordinates from both control and patterned devices as a function of external viewing angle and find them to be largely identical, being (0.66, 0.33) and (0.67, 0.33) at normal incident from the control devices and patterned devices, respectively, and (0.67, 0.33), (0.67, 0.32) from the control devices and the patterned devices at a viewing angle of 40. Such differences in the CIE coordinates are clearly very small, permitting us to conclude that the observed enhancement in the efficiency (in terms of cd A1) from the patterned devices come as a result of increasing the relative external efficiency rather than tuning the EL emission to a part of the spectrum where the eye has an improved sensitivity. This also demonstrates that unlike microcavity-based LEDs, the enhancements in external efficiency observed here do not come at the cost of a severe angular-dependent colour-shift. We can gain further confidence in the effect of the PC on the improved output coupling by comparing the PL emission from a patterned LED with a control device. This is shown in Fig. 5(a), where we plot the PL spectra from a control LED with a patterned device as a function of viewing angle. It can be seen

A.M. Adawi et al. / Organic Electronics 7 (2006) 222–228 Patterned LED

Control Device o

PL intensity

10 o 30 o 50

500

600 700 Wavelength (nm)

Relative PL enhancement

(a)

10 o 30 o 50

800

2.4

2.0

1.6 10

(b)

o

20

30 40 50 Viewing angle (θ)

60

Fig. 5. Part (a) shows the PL spectra from standard device and a laterally patterned structure as a function of external viewing angle. Part (b) shows the enhancement in PL emission intensity (averaged over wavelength), also presented as a function of viewing angle.

that the emission spectra are very similar to those generated via electrical excitation, with the PL intensity from both devices decreasing with increasing viewing angle as expected. As was observed under electrical excitation, the device emission colour is angle-independent. Importantly, the PL emission intensity from the laterally patterned structure is higher than from the control at all external viewing-angles. This is illustrated in Fig. 5(b), where we plot the PL enhancement factor as a function of viewing angle. This enhancement factor was calculated as the ratio of the PL intensity (integrated over 500–800 nm) from the patterned structure to that of the control. It can be seen that close to normal incidence, the PL from the patterned structure was enhanced by approximately 2.3 times – a value in good agreement with the results of the EL enhancement as presented above. This enhancement in relative PL emission efficiency also survives to large off-axis angles, being a factor of 1.6 times at a viewing angle of 60. This observation further confirms that inserting a photonic crystal between the glass substrate and the ITO anode improves the fraction of light that can escape from of the structure.

227

We have therefore investigated the effect of inserting a two-dimensional PC having a pitch of 1.2 lm between the glass substrate and the ITO anode layer of red-emitting polymer LED. Such periodic patterning can be used to generate a significant improvement in device external emission efficiency (in this case by a factor of >2 times). Usefully, the enhancements that we observe are not accompanied by an angular-dependent colour shift. The grating used here was not sufficient to open up a photonic bandgap, however, we argued that the PC effectively diffracted waveguided modes trapped in the ITO layer into useful external emission. It is also possible however, that the close proximity of the PC to the active emissive region of the LED can modify the optical density of states within the device, which may also have an effect on the spontaneous emission process itself. The approach that we used to pattern the LED substrate was based on interference lithography. Such a process would probably not be scalable to a manufacturing environment, however, we note that many other techniques can be used to pattern large areas at this resolution. For example, polymer embossing techniques can be used in a manufacturing environment to pattern relatively large areas with a resolution of around 100 nm [29]. Thus the simple approach prototyped here could be used to significantly improve the efficiency of a number of organic light-emitting devices, ranging from light emitting displays to organic flat panel lighting systems. Acknowledgements We wish to thank the UK EPSRC and DTI for support of this research via grant GR/S05687/01 ‘Light Emission for Active Polymers’ (LEAP). References [1] G. Gu, S.R. Forrest, IEEE J. Sel. Top. Quant. Electron. 4 (1998) 83. [2] K. Ziemelis, Nature 399 (1999) 408. [3] I.D.W. Samuel, A. Beeby, Nature 403 (2002) 710. [4] M. Pfeiffer, S.R. Forrest, K. Leo, M.E. Thompson, Adv. Mater. 14 (2002) 1633. [5] A.B. Chwang, M.A. Rothman, S.Y. Mao, R.H. Hewitt, M.S. Weaver, J.A. Silvernail, K. Rajan, M. Hack, J.J. Brown, X. Chu, L. Moro, T. Krajewski, N. Rutherford, Appl. Phys. Lett. 83 (2003) 413. [6] C.F. Madigan, M.H. Lu, J.C. Sturm, Appl. Phys. Lett. 76 (2000) 1650. [7] T. Yamasaki, K. Sumioka, T. Tsustui, Appl. Phys. Lett. 76 (2000) 1243.

228

A.M. Adawi et al. / Organic Electronics 7 (2006) 222–228

[8] H.J. Peng, Y.L. Ho, X.J. Yu, H.S. Kwok, J. Appl. Phys. 96 (2004) 1649. [9] R.H. Jordan, L.J. Rothberg, A. Dodabablapur, R.E. Slusher, Appl. Phys. Lett. 69 (1996) 1997. [10] S. Moller, S.R. Forrest, J. Appl. Phys. 91 (2002) 3324. [11] J.M. Lupton, B.J. Matterson, I.D.W. Samuel, M.J. Jory, W.L. Barnes, Appl. Phys. Lett. 77 (2000) 3340. [12] B.J. Matterson, J.M. Lupton, A.F. Safonov, M.G. Salt, W.L. Barnes, I.D.W. Samuel, Adv. Mater. 13 (2001) 123. [13] P.A. Hobson, S. Wedge, J.A.E. Wasey, I. Sage, W.L. Barnes, Adv. Mater. 14 (2002) 1393. [14] J.M. Ziebarth, A.K. Saafir, S. Fan, M.D. McGehee, Adv. Funct. Mater. 14 (2004) 451. [15] M. Fujita, T. Ueno, T. Asano, S. Noda, H. Ohhata, T. Tsuji, H. Nakada, N. Shimoji, Electron. Lett. 39 (2003) 1750. [16] M. Fujita, K. Ishihara, T. Ueno, T. Asano, S. Noda, H. Ohata, T. Tsuji, H. Nakada, N. Shimoji, Jpn. J. Appl. Phys. 44 (2005) 3669. [17] G. Fichet, N. Corcoran, P.K.H. Ho, A.C. Arias, J.D. MacKenzie, W.T.S. Huck, R.H. Friend, Adv. Mater. 16 (2004) 1908. [18] Y.-J. Lee, S.-H. Kim, J. Huh, G.-H. Kim, Y.-H. Lee, S.-H. Cho, Y.-C. Kim, Y.R. Do, Appl. Phys. Lett. 82 (2003) 3779.

[19] Y.R. Do, Y.-C. Kim, Y.-W. Song, C.-O. Cho, H. Jeon, Y.-J. Lee, S.-H. Kim, Adv. Mater. 15 (2003) 1214. [20] M. Kitamura, S. Iwamoto, Y. Arakawa, Jpn. J. Appl. Phys. 44 (2005) 2844. [21] Y.R. Do, Y.-C. Kim, Y.-W. Song, Y.-H. Lee, J. Appl. Phys. 96 (2004) 7629. [22] C. Liu, V. Kamaev, Z.V. Vardeny, Appl. Phys. Lett. 86 (2005) 143501-1. [23] J.D. Joannopoulos, R.D. Meade, J.N. Win, Photonic Crystals, Princeton University Press, 1995. [24] S. Fan, P. Villeneuve, J.D. Joannopoulos, Phys. Rev. Lett. 78 (1997) 3294. [25] J. Shaky, K.H. Kim, J.Y. Lin, H.X. Jiang, Appl. Phys. Lett. 85 (2004) 142. [26] J.P. Spallas, A.M. Hawryluk, D.R. Kania, J. Vac. Sci. Technol. B 13 (1995) 1973. [27] S.C. Kitson, W.L. Barnes, J.R. Sambles, IEE Photonics Technol. Lett. 8 (1996) 1041. [28] I. Mikulskas, J. Mickevicius, J. Vaitkus, R. Tomasiunas, V. Grigaliunas, V. Kopusyainskas, S. Meskinins, Appl. Surf. Sci. 186 (2002) 599. [29] L.S. Roman, O. Ingana¨s, T. Granlund, T. Nyberg, M. Svensson, M.R. Andersson, J.C. Hummelen, Adv. Mater. 12 (2000) 189.