Efficient red electrophosphorescence from a fluorene-based bipolar host material

(1)

Efﬁcient red electrophosphorescence from a ﬂuorene-based

bipolar host material

Chen-Han Chien

a,b

, Fang-Ming Hsu

a,b

, Ching-Fong Shu

a,*

, Yun Chi

b,* a

Department of Applied Chemistry, National Chiao Tung University, 300 Hsinchu, Taiwan

b

Department of Chemistry, National Tsing Hua University, 300 Hsinchu, Taiwan

a r t i c l e

i n f o

Article history:

Received 25 February 2009 Received in revised form 1 April 2009 Accepted 23 April 2009

Available online 5 May 2009

PACS: 71.20.Rv 72.80.Le 73.61.Ph 78.60.Fi Keywords: OLED Electrophosphorescence Bipolar host material Red-emitting White-emitting

a b s t r a c t

We have prepared efﬁcient red organic light-emitting diodes (OLEDs) incorporating 2,7-bis(diphenylphosphoryl)-9-[4-(N,N-diphenylamino)phenyl]-9-phenylﬂuorene (POAPF) as the host material doped with the osmium phosphor Os(fptz)2(PPh2Me)2(fptz =

3-triﬂuoro-methyl-5-pyridyl-1,2,4-triazole). POAPF, which possesses bipolar functionalities, can facil-itate both hole- and electron-injection from the charge transport layers to provide a balanced charge ﬂux within the emission layer. The peak electroluminescence performance of the device reached as high as 19.9% and 34.5 lm/W – the highest values reported to date for a red phosphorescent OLED. In addition, we fabricated a POAPF-based white light OLED – containing red-[doped with Os(fptz)2(PPh2Me)2] and blue-emitting {doped with

irid-ium(III) bis[(4,6-diﬂuorophenyl)pyridinato-N,C20

] picolinate, FIrpic} layers – that also exhibited satisfactory efﬁciencies (18.4% and 43.9 lm/W).

1. Introduction

Phosphorescent organic light-emitting diodes (OLEDs) attract a great deal of attention because they can theoreti-cally approach 100% internal quantum efficiency by har-nessing both singlet and triplet excitons [1–6]. In these phosphorescent devices, the organometallic phosphors are commonly doped into appropriate host materials to prevent concentration quenching and achieve high effi-ciency[1,2]. The past few years have witnessed the devel-opment of blue and green phosphorescent OLEDs exhibiting high external quantum efficiencies (EQEs) and

power efﬁciencies (PEs)[4,7–12]. There are, however, far fewer reports describing high-performance red phospho-rescent OLEDs. Being one of the primary colors, red emis-sion from OLEDs is important for applications in full-color displays and solid state lighting[13–19]. For future device applications, it will be essential for OLEDs to exhibit efﬁcient, saturated red emissions.

Since Baldo et al. developed the concept of phosphores-cent OLEDs[1], the carbazole derivative 4,40_-N,N0

-dicarba-zolebiphenyl (CBP) has been used widely as a host material for red phosphorescent OLEDs[3,20–27]because of its suitable triplet energy (ET) and good hole-transporting

ability. Because CBP has a wide band gap (Eg), injecting

car-riers are often required to overcome the large energy barri-ers between the charge transport laybarri-ers and its highest occupied molecular orbital (HOMO) and/or lowest unoccu-pied molecular orbital (LUMO). In addition, the poor match

*Corresponding authors. Tel.: +886 3 5712121x56544; fax: +886 3 5723764 (C.-F. Shu).

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

Contents lists available atScienceDirect

Organic Electronics

(2)

between the energy levels of CBP and the red triplet dopants usually results in direct charge trapping within the emitting layer (EML). Therefore, red phosphorescent devices based on CBP often require high operation voltages and provide unsatisfactory PEs. To break free from these constraints, sev-eral attempts have been made recently to use host materials possessing narrower band gaps[28–30]or bipolar charac-teristics[31–33]to improve the charge injection and elec-troluminescence (EL) performance of red phosphorescent OLEDs. Nevertheless, the PEs of the red electrophosphores-cence remain far below those reported for blue- or green-emitting devices[3,20–27].

In this paper, we report the fabrication of efficient red phosphorescent OLEDs incorporating the bipolar host 2,7-bis(diphenylphosphoryl)-9-[4-(N,N-diphenylamino) phenyl]-9-phenylfluorene (POAPF) doped with the effi-cient red-emitting phosphor Os(fptz)2(PPh2Me)2 (fptz =

3-triﬂuoromethyl-5-pyridyl-1,2,4-triazole);Fig. 1presents their chemical structures. This charge–neutral osmium complex has been used previously as a phosphor in red-emitting devices displaying high EQEs[22,34]. The rel-atively low oxidation potential of Os(fptz)2(PPh2Me)2,

however, results in severe charge trapping in devices using CBP as the host, resulting in high driving voltages and unsatisfactory PEs. POAPF, which contains an electron-rich triphenylamine group and an electron-deficient bis(diphenylphosphoryl)fluorene unit [35–37], has re-cently been developed as a host material exhibiting bipolar characteristics[38]that not only provide suitable frontier orbital energies for facile hole- and electron-injection but also improve the balance of charge flux in the emission layer. Moreover, the triplet energy (ET) of POAPF is

esti-mated to be 2.72 eV – a value that is sufﬁciently high for POAPF to act as a host material for red, green, and even blue phosphorescent emitters. Indeed, blue phosphores-cent OLEDs based on POAPF doped with FIrpic can exhibit high values of EQE and PE[38]. We expected that POAPF could also be utilized as a host for the red-emitting Os(fptz)2(PPh2Me)2, thereby improving charge injection

and enhancing charge balance to result in highly efﬁcient red phosphorescent devices.

2. Experimental

The bipolar host molecule POAPF [38] and the red phosphorescent emitter Os(fptz)2(PPh2Me)2[22]were

pre-pared using previously reported procedures. The

hole-transport material 4,40

-bis(3-methylphenylphenylami-no)biphenyl (TPD), the conventional host material CBP, and the electron-transport material 4,7-diphenyl-1,10-phenanthroline (BPhen) were all purchased from LumTec Corp. and used without further puriﬁcation.

The oxidation and reduction potentials were measured, respectively, in anhydrous CH2Cl2 and anhydrous DMF,

containing 0.1 M TBAPF6as the supporting electrolyte, at

a scan rate of 50 mV/s against a Ag/Ag+_{(0.01 M AgNO} 3)

ref-erence electrode, with ferrocene as the internal standard. The onset potentials were determined from the intersec-tion of two tangents drawn at the rising current and back-ground current of the cyclic voltammogram.

The EL devices were fabricated through vacuum deposi-tion (106_{torr) of the materials onto ITO glass (sheet}

resis-tance: 25X/square). All of the organic layers were deposited at a rate of 1.0 Å/s. The cathode was completed through thermal deposition of LiF (15 Å; deposition rate: 0.1 Å/s) and then capping with Al metal (100 nm) through thermal evaporation (deposition rate: 4.0 Å/s). The cur-rent–voltage–luminance relationships of the devices were measured using a Keithley 2400 source meter and a New-port 1835C optical meter equipped with an 818ST silicon photodiode. The EL spectrum was obtained using a Hitachi F4500 spectroﬂuorimeter.

3. Results and discussion

Fig. 2presents the current density–voltage (I–V) charac-teristics of the hole-only devices having the conﬁguration indium tin oxide (ITO)/TPD (30 nm)/host material (30 nm)/TPD (30 nm)/Al (100 nm) and the electron-only devices having the conﬁguration ITO/BPhen (30 nm)/host material (30 nm)/BPhen (30 nm)/LiF (15 Å)/Al (100 nm). Here, ‘‘host material” refers to either the bipolar host POAPF or the conventional host CBP. Because of the high LUMO energy level (2.20 eV) of TPD in the hole-only de-vices and the low HOMO energy level (6.40 eV) of BPhen in the electron-only devices, injections of electrons and holes were prohibited in the hole- and electron-only de-vices, respectively; accordingly, the measured I–V charac-teristics were dominated by holes and electrons, respectively. Fig. 2reveals that the POAPF-based devices exhibited lower turn-on voltages and higher current densi-ties than did the CBP-based devices under the same bias.

Fig. 1. Chemical structures of POAPF and Os(fptz)2(PPh2Me)2.

Fig. 2. Current density–voltage (I–V) curves of (a) hole- and (b) electron-only devices.

(3)

The reduced driving voltages in the POAPF-based devices may be due to the decrease in barrier heights for hole-and electron-injection from the charge transport layers (CTLs) to POAPF.

To further understand the improved injection of holes and electrons in the POAPF-based devices, we employed cyclic voltammetry (CV) to estimate the HOMO and LUMO energy levels of POAPF and CBP (Fig. 3a), using a three-electrode cell with ferrocene as the internal standard. Be-cause of its triphenylamine donor unit, POAPF exhibits a lower onset of its oxidation potential (0.44 V) relative to that of CBP (0.74 V). On the other hand, the onset of the reduction potential of POAPF (2.32 V), originating from its bis(diphenylphosphoryl)ﬂuorene unit, is signiﬁcantly less negative than that of CBP (2.72 V). On the basis of these onset values for oxidation and reduction, we esti-mated the HOMO and LUMO energy levels of POAPF (5.24 and 2.48 eV, respectively) and CBP (5.54 and 2.08 eV, respectively) relative to ferrocene (4.80 eV be-low the vacuum) as a reference[39]. According to the en-ergy level diagram in Fig. 3b, the hole-injection barrier from TPD and the electron-injection barrier from BPhen to the POAPF layer were both considerably lower than those to the CBP layer. As a result, the presence of POAPF facilitates hole- and electron-injection, as observed in the hole- and electron-only measurements (vide supra).

Using POAPF as the host material, we fabricated red OLEDs in the conﬁguration ITO/TPD (30 nm)/R-EML (30 nm)/BPhen (30 nm)/LiF (15 Å)/Al (100 nm), where

TPD and BPhen were employed as hole- and electron-transporting layers, respectively, and R-EML refers to the POAPF layer doped with various concentrations of Os(fptz)2(PPh2Me)2. For comparison, we also fabricated

reference devices incorporating CBP as the host.Fig. 4 de-picts the I–V characteristics of these red-emitting OLEDs. As expected, the POAPF-based devices exhibited signiﬁ-cantly higher current densities than did the CBP-based de-vices under the same bias. The I–V characteristics of the POAPF-based devices underwent no apparent changes when we increased the doping concentration of Os(fptz)2(PPh2Me)2 from 0 to 10 wt%. This phenomenon

can be explained by considering the energy level diagram of the devices (Fig. 3b). Because the HOMO energy level of Os(fptz)2(PPh2Me)2is merely 0.34 eV higher than that

of POAPF, it behaves as a less-effective trapping sites for holes; meanwhile, because the LUMO energy level of Os(fptz)2(PPh2Me)2is 0.28 eV above that of POAPF,

elec-trons injected from the BPhen layer into the POAPF layer are mostly transported through the POAPF layer without being trapped in the osmium phosphor. Thus, varying the concentration of Os(fptz)2(PPh2Me)2 in the EML should

not signiﬁcantly alter the degree of charge injection in POAPF-based devices. On the other hand, the current den-sity of the CBP-based devices exhibited a strong depen-dence on the doping concentration. Because the HOMO energy level of Os(fptz)2(PPh2Me)2is 0.64 eV higher than

that of CBP, holes can potentially be trapped at

Fig. 3. (a) Cyclic voltammograms of POAPF and CBP, recorded at a scan rate of 50 mV/s. (b) Energy level diagram of the red phosphorescent devices.

Fig. 4. I–V characteristics of devices incorporating (a) POAPF and (b) CBP as hosts at various doping concentrations.

(4)

Os(fptz)2(PPh2Me)2 at low doping concentrations. When

the concentration of the osmium phosphor was increased, we attribute the resulting increased hole-injection efﬁ-ciency to direct charge injection from the hole-transport-ing layer (HTL) into Os(fptz)2(PPh2Me)2. Accordingly,

employing POAPF as the host for Os(fptz)2(PPh2Me)2

sig-niﬁcantly reduced the degree of charge trapping within the EML.

The optimal performances of both the POAPF- and CBP-based devices occurred at a doping concentration of 7 wt%.

Table 1lists the key characteristics of these red electroph-osphorescent devices. The POAPF-based red-emitting de-vice had a relatively low turn-on voltage of 2.3 V (corresponding to 1 cd/m2_{) and, at a practical brightness}

of 1000 cd/m2_{, its driving voltage was merely 3.7 V. These}

values are much lower than those of the CBP-based device (3.5 and 5.1 V, respectively). We ascribe the reduced driv-ing voltage of the POAPF-based device to facile hole/elec-tron-injections and less-effective charge trapping within this bipolar host.Fig. 5presents the EQE and PE proﬁles plotted with respect to the current density. The POAPF-based device exhibited a maximum EQE of 19.9% (at a luminance of 96 cd/m2_{) – 1.5 times greater than that of}

its CBP-based counterpart (13.2%). These values reveal that the presence of the bipolar host material resulted in bal-anced charge fluxes within the EML and, therefore, a high EQE. Notably, the PE of the POAPF-based device (34.5 lm/ W) was 2.5 times higher than that of its CBP-based coun-terpart (13.4 lm/W). We attribute this large enhancement in device performance to the substantially lower driving voltage and improved charge balance within the POAPF-based device. In addition to its highly efficient device per-formance, the POAPF-based OLED displayed a less-prob-lematic efficiency roll-off at higher luminance. The efficiencies remained fairly high (18.6% and 26.1 lm/W) at a practical luminance of 1000 cd/m2_{. Our POAPF-based}

red-emitting device possesses superior EL performance, particularly in terms of its PE (Table 2), relative to those of previously reported red electrophosphorescent OLEDs

[29,31,32,34]. Both the POAPF- and CBP-based devices

emitted saturated red light (each main peak at 616 nm) and Commission Internationale de L’Eclairage (CIE) coordi-nates of (0.64, 0.36) at 5 V (inset toFig. 5).

To further utilize the superior host properties of POAPF, we fabricated white light-emitting devices – featuring red emission from Os(fptz)2(PPh2Me)2 and complementary

blue/green emission from FIrpic – having the conﬁguration ITO/TPD (30 nm)/R-EML (x nm)/B-EML (30x nm)/BPhen (30 nm)/LiF (15 Å)/Al (100 nm). Here, we used POAPF as the bipolar host material for both the red EML [R-EML; doped with 7 wt% Os(fptz)2(PPh2Me)2] and the blue EML

(B-EML; doped with 7 wt% FIrpic); the total thickness of these dual emission layers was ﬁxed at 30 nm in each de-vice. To realize white light emission from such composite emission systems, we adjusted the thickness ratio at vari-ous compositions to achieve appropriate fractions of exci-tons for both triplet emitters.Fig. 6reveals an increase in the blue emission intensity relative to red emission upon decreasing the thickness of the R-EML. We obtained white emission when the thicknesses of the R-EML and B-EML were 1 and 29 nm, respectively. This white-emitting device also exhibited a low turn-on voltage (2.2 V) and an applica-ble luminance of 1000 cd/m2_{at merely 3.0 V (}_{Fig. 7}_{a); its}

maximum EQE reached 18.4% at a luminance of 90 cd/m2. As a result of the low driving voltage, the PE reached as high as 43.9 lm/W (Fig. 7b). Even when the luminance

Table 1

Device performance.

Device Red White

Host POAPF CBP POAPF Turn-on voltage [V]a 2.3 3.5 2.2 Max. EQE [%] 19.9 13.2 18.4 Max. LE [cd/A] 32.8 21.4 34.5 Max. PE [lm/W] 34.5 13.4 43.9 Voltage [V]b _3.7 _5.1 _3.0 EQE [%]b 18.6 13.1 14.9 LE [cd/A]b 30.6 21.1 27.9 PE [lm/W]b 26.1 13.1 29.5 EL kmax[nm]c 616 616 472, 498, 612 CIE, x and yc (0.64, 0.36) (0.64, 0.36) (0.38, 0.34) a _{Recorded at 1 cd/m}2_. b Recorded at 1000 cd/m2 . c At 5 V.

Fig. 5. EQE and PE of 7 wt% Os(fptz)2(PPh2Me)2-doped POAPF- and

CBP-based devices plotted with respect to the current density. Inset: EL spectra (applied voltage: 5 V) of the POAPF- and CBP-based devices.

Table 2

Performance of red electrophosphorescent devices.

Host Dopant Max. EQE [%] Max. PE [lm/W] CIE (x, y) Ref. POAPF Os(fptz)2(PPh2Me)2 19.9 34.5 (0.64, 0.36) This study

(ppy)2Ir(acac) Ir(piq)3 9.2 11.0 (0.65, 0.35) [29]

D2ACN Mpq2Iracac 10.8 13.0 (0.66, 0.34) [31]

o-CzOXD Ir(piq)2acac 18.5 11.5 (0.68, 0.32) [32]

TFTPA Os(fptz)2(PPh2Me)2 18.0 25.2 (0.64, 0.36) [34]

(5)

was 1000 cd/m2_{, the EL efﬁciencies remained high (14.9%}

and 29.5 lm/W). The EL efﬁciency of this white electroph-osphorescence is among the highest values ever reported for white light-emitting diodes[12,40,41]. The EL spectra of this device revealed a virtually stable white color emis-sion (inset toFig. 7b). The CIE coordinates shifted slightly from (0.40, 0.34) at a luminance of 1057 cd/m2 _{to (0.38,}

0.34) at 15,170 cd/m2; nevertheless, the color coordinates remained in the white light region.

4. Conclusion

We have realized highly efﬁcient red electrophospho-rescence in OLEDs incorporating the bipolar host POAPF

doped with the red-emitting osmium phosphor

Os(fptz)2(PPh2Me)2. The bipolar characteristics of POAPF

resulted in the red-emitting devices exhibiting very low driving voltages and achieving EL efficiencies as high as 19.9% and 34.5 lm/W – the highest performance of any red phosphorescent OLED reported to date – with a satu-rated red emission located at CIE coordinates of (0.64, 0.36). In addition, we used POAPF to fabricate a white light-emitting device possessing a dual EML configuration; it also exhibited satisfying efficiencies (18.4% and 43.9 lm/ W). The white emission remained stable, with the CIE coor-dinates shifting slightly from (0.40, 0.34) at a luminance of 1057 cd/m2 to (0.38, 0.34) at 15,170 cd/m2. The perfor-mance of these phosphorescent devices makes them very attractive materials for potential commercial applications. Acknowledgment

We thank the National Science Council for ﬁnancial support.

References

[1] M.A. Baldo, D.F. O’Brien, Y. You, A. Shoustikov, S. Sibley, M.E. Thompson, S.R. Forrest, Nature 395 (1998) 151.

[2] M.A. Baldo, S. Lamansky, P.E. Burrows, M.E. Thompson, S.R. Forrest, Appl. Phys. Lett. 75 (1999) 4.

[3] S. Lamansky, P. Djurovich, D. Murphy, F. Abdel-Razzaq, H.E. Lee, C. Adachi, P.E. Burrows, S.R. Forrest, M.E. Thompson, J. Am. Chem. Soc. 123 (2001) 4304.

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

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

[6] S. Watanabe, N. Ide, J. Kido, Jpn. J. Appl. Phys. 46 (2007) 1186. [7] M. Ikai, S. Tokito, Y. Sakamoto, T. Suzuki, Y. Taga, Appl. Phys. Lett. 79

(2001) 156.

[8] S.J. Yeh, M.F. Wu, C.T. Chen, Y.H. Song, Y. Chi, M.H. Ho, S.F. Hsu, C.H. Chen, Adv. Mater. 17 (2005) 285.

[9] M.H. Tasi, H.W. Lin, H.C. Su, T.H. Ke, C.C. Wu, F.C. Fang, Y.L. Liao, K.T. Wong, C.I. Wu, Adv. Mater. 18 (2006) 1216.

[10] D. Tanaka, H. Sasabe, Y.J. Li, S.J. Su, T. Takeda, J. Kido, Jpn. J. Appl. Phys. 46 (2007) L10.

[11] S.J. Su, T. Chiba, T. Takeda, J. Kido, Adv. Mater. 20 (2008) 2125. [12] S.J. Su, E. Gonmori, H. Sasabe, J. Kido, Adv. Mater. 20 (2008) 4189. [13] C.W. Tang, S.A. VanSlyke, Appl. Phys. Lett. 51 (1987) 913. [14] C.W. Tang, S.A. VanSlyke, C.H. Chen, J. Appl. Phys. 65 (1989) 3610. [15] M.A. Baldo, M.E. Thompson, S.R. Forrest, Nature 403 (2000) 750. [16] T. Fuhrmann, J. Salbeck, MRS Bull. 28 (2003) 354.

[17] B.W. D’Andrade, R.J. Holmes, S.R. Forrest, Adv. Mater. 16 (2004) 624. [18] B.W. D’Andrade, S.R. Forrest, Adv. Mater. 16 (2004) 1585. [19] H. Kanno, R.J. Holmes, Y. Sun, S. Kena-Cohen, S.R. Forrest, Adv. Mater.

18 (2006) 339.

[20] A. Tsuboyama, H. Iwawaki, M. Furugori, T. Mukaide, J. Kamatani, S. Igawa, T. Moriyama, S. Miura, T. Takiguchi, S. Okada, M. Hoshino, K. Ueno, J. Am. Chem. Soc. 125 (2003) 12971.

[21] C.H. Yang, C.C. Tai, I.W. Sun, J. Mater. Chem. 14 (2004) 947. [22] Y.L. Tung, S.W. Lee, Y. Chi, Y.T. Tao, C.H. Chien, Y.M. Cheng, P.T. Chou,

S.M. Peng, C.S. Liu, J. Mater. Chem. 15 (2005) 460.

[23] C.L. Li, Y.J. Su, Y.T. Tao, P.T. Chou, C.H. Chien, C.C. Cheng, R.S. Liu, Adv. Funct. Mater. 15 (2005) 387.

[24] D.K. Rayabarapu, B.M.J.S. Paulose, J.P. Duan, C.H. Cheng, Adv. Mater. 17 (2005) 349.

[25] Y.L. Tung, S.W. Lee, Y. Chi, L.S. Chen, C.F. Shu, F.I. Wu, A.J. Carty, P.T. Chou, S.M. Peng, G.H. Lee, Adv. Mater. 17 (2005) 1059.

[26] S. Takizawa, Y. Sasaki, M. Akhtaruzzaman, H. Echizen, J. Nishida, T. Iwata, S. Tokito, Y. Yamashita, J. Mater. Chem. 17 (2007) 841. Fig. 6. EL spectra of dual-emission devices (applied voltage: 5 V)

incor-porating various thicknesses of their R-EML and B-EML.

Fig. 7. Plots of (a) current density and luminance with respect to the voltage and (b) EQE and PE with respect to the current density for the white light-emitting device. Inset: EL spectra recorded at various luminances.

(6)

[27] C.L. Ho, W.Y. Wong, Z.Q. Gao, C.H. Chen, K.W. Cheah, B. Yao, Z. Xie, Q. Wang, D. Ma, L. Wang, X.M. Yu, H.S. Kwok, Z. Lin, Adv. Funct. Mater. 18 (2008) 319.

[28] T. Tsuzuki, Y. Nakayama, J. Nakamura, T. Iwata, S. Tokito, Appl. Phys. Lett. 88 (2006) 243511.

[29] T. Tsuzuki, S. Tokito, Adv. Mater. 19 (2007) 276.

[30] T.J. Park, W.S. Jeon, J.J. Park, S.Y. Kim, Y.K. Lee, J. Jang, J.H. Kwon, R. Pode, Appl. Phys. Lett. 92 (2008) 113308.

[31] W.Y. Hung, T.C. Tsai, S.Y. Ku, L.C. Chi, K.T. Wong, Phys. Chem. Chem. Phys. 10 (2008) 5822.

[32] Y. Tao, Q. Wang, C. Yang, Q. Wang, Z. Zhang, T. Zou, J. Qin, D. Ma, Angew. Chem. Int. Ed. 47 (2008) 8104.

[33] Y. Tao, Q. Wang, Y. Shang, C. Yang, L. Ao, J. Qin, D. Ma, Z. Shuai, Chem. Commun. (2009) 77.

[34] C.H. Wu, P.I. Shih, C.F. Shu, Y. Chi, Appl. Phys. Lett. 92 (2008) 233303. [35] A.B. Padmaperuma, L.S. Sapochak, P.E. Burrows, Chem. Mater. 18

(2006) 2389.

[36] S.O. Jeon, K.S. Yook, C.W. Joo, J.Y. Lee, Appl. Phys. Lett. 94 (2009) 013301.

[37] F.M. Hsu, C.H. Chien, P.I. Shih, C.F. Shu, Chem. Mater. 21 (2009) 1017. [38] F.M. Hsu, C.H. Chien, C.-F. Shu, C.H. Lai, C.C. Hsieh, K.W. Wang, P.T.

Chou, submitted for publication.

[39] J. Pommerehne, H. Vestweber, W. Guss, R.F. Mahrt, H. Bässler, M. Porsch, J. Daub, Adv. Mater. 7 (1995) 551.

[40] Y. Sun, S.R. Forrest, Appl. Phys. Lett. 91 (2007) 263503.

[41] G. Schwartz, M. Pfeiffer, S. Reineke, K. Walzer, K. Leo, Adv. Mater. 19 (2007) 3672.