Journal of Materials Science: Materials in Electronics

, Volume 28, Issue 17, pp 12761–12767 | Cite as

Improving the performance of organic light-emitting devices by incorporating non-doped TCNQ as electron buffer layer



The performance of organic light-emitting devices (OLEDs) is improved by inserting non-doped tetracyanoquinodimethane (TCNQ) electron buffer layer (EBL) between 4,7-diphnenyl-1,10-phe-nanthroline (Bphen) electron transport layer (ETL) and LiF/Al cathode. By optimizing the thickness of TCNQ layer, we find that the device with 6 nm TCNQ buffer layer can achieve the best performance. The maximum luminance, current efficiency, power efficiency and half-lifetime of the optimal device are increased by 27.32, 51.70, 127.55, and 73.89%, respectively, compared with those of the control device without TCNQ buffer layer. This improvement can be attributed to that the insertion of non-doped TCNQ buffer layer which is a simple approach can enhance the electron injection and operational stability of the devices. Moreover, we have carried out the tests of the atomic force microscope (AFM), scanning electron microscopy (SEM) and Kelvin probe to explore the effect of insering TCNQ. These tests results further verify that TCNQ layer not only smooth the surface of the films but also improve the electron injection and transport characteristics. As a result, the performances of the OLEDs can be effectively improved.

1 Introduction

With the preceding work of a double-layer organic light-emitting device (OLED) reported by Tang [1], a lot of researchers have paid their attention to OLEDs due to their potential applications in flat-panel displays and solid-state lighting [2, 3, 4]. The efficient and stable OLEDs are important for their commercial applications. Balanced carrier injection and transport are crucial for efficient OLEDs [5]. However, it is more difficult for electrons to be injected into organic layers from electrodes in comparison with holes [6]. In general, for organic semiconductor materials, electron mobility of electron transport materials is lower than hole mobility of hole transport materials. Therefore, enhancing electron injection and transport is an effective strategy to improve the efficiency of devices [7]. In addition, long operational lifetime is essential to achieve commercial success for OLEDs [8]. There were many reports showing that improving the stability of film could enhance the stability of OLEDs [9]. Tetracyanoquinodimethane (TCNQ) is an n-type material, it has high electron mobility and good thermal stability [10, 11]. Furthermore, due to high highest occupied molecule orbital (HOMO) (7.6 eV) level of TCNQ, it can also be used as hole blocking material in organic semiconductor based devices [12]. TCNQ has been extensively applied in organic field-effect transistor and solar cells [13, 14, 15, 16], but there are few reports about applications of TCNQ in OLEDs. Recently, Srivastava et al. has reported that Bphen doped with TCNQ is used to be a novel electron injection layer (EIL) [17]. However, for this kind of thin-doped film, the doping process is relatively complicated based on the co-evaporation method, and so far there were no reports about non-doped TCNQ used as electron buffer layer (EBL) [18, 19]. In this work, we systematically investigate mechanism of TCNQ serving as EBL in a group of blue emitting devices based on 4,7-diphnenyl-1,10-phe-nanthroline (Bphen) used as ETL. The results indicate that TCNQ buffer layer with the optimized thickness can effectively enhance electron injection and operational stability of OLEDs.

2 Experiment

We conducted a series of experiments to clarify the role of TCNQ in OLEDs. The configurations of devices are listed as follows:

  1. (A)

    ITO/NPB (40 nm)/DNCA (30 nm)/Bphen (20 nm)/LiF (0.8 nm)/Al (120 nm)

  2. (B)

    ITO/NPB (40 nm)/DNCA (30 nm)/Bphen (20 nm)/TCNQ (3 nm)/LiF (0.8 nm)/Al (120 nm)

  3. (C)

    ITO/NPB (40 nm)/DNCA (30 nm)/Bphen (20 nm)/TCNQ (6 nm)/LiF (0.8 nm)/Al (120 nm)

  4. (D)

    ITO/NPB (40 nm)/DNCA (30 nm)/Bphen (20 nm)/TCNQ (9 nm)/LiF (0.8 nm)/Al (120 nm)

  5. (E)

    ITO/NPB (40 nm)/DNCA (30 nm)/Bphen (20 nm)/TCNQ (12 nm)/LiF (0.8 nm) /Al (120 nm)


Figure 1 shows the principle scheme of these OLEDs. In these devices, N,N′-Bis-(1-naphthalenyl)-N,N′-bis-phenyl-(1,1′-biphenyl)-4,4′-diamine (NPB), a novel blue fluorescent material derivation of N6,N6,N12,N12-tetrap-tolylchlchrysene-6,12-diamine (DNCA), 4,7-diphnenyl-1,10-phenanthroline (Bphen) and tetracyanoquinodimethane (TCNQ) are used as hole transport layer (HTL), emitting layer (EML), electron transport layer and electron buffer layer, respectively. All devices were fabricated on glass substrates which were coated with indium tin oxide (ITO) with a sheet resistance of 20 Ω/sq. Prior to deposition, the ITO substrates were cleaned successively in detergent, deionized water, acetone, and isopropanol. All the organic layers, as well as LiF and Al, were deposited on ITO glass substrates by using a thermal evaporation of Edwards Auto-500 thermal evaporation system in an M. Braun 20 G glove box at a pressure of about 7 × 10−5 Pa. The deposition rates and thicknesses of the different layers were monitored using a quartz crystal oscillator. The deposition rates were 0.1, 0.05, and 1 nm/s for the organic layers, the LiF layer, and the Al layer, respectively. The current density–voltage–luminance (JVL) characteristics of the devices were simultaneously measured by a PR-650 spectra scan spectrometer with a Keithley model 2400 programmable voltage-current source. Field-emission scanning electron microscopy (SEM, Hitachi S-8010) was used to characterize the cross-sectional images of the films. The atomic force microscopy (AFM, BRUKER INNOVA) using an n-doped silicon cantilever in tapping mode was utilized to analyze the surface morphologies of the films. The Kelvin probe (KP-020) was used to measure the work functions of the films. All the measurements were carried out at room temperature under ambient condition (except for lifetime tests).

Fig. 1

(Color online) Principle scheme of these OLEDs (x = 0, 3, 6, 9, 12)

3 Results and discussion

Figure 2a shows JVL characteristics of devices with TCNQ layer of different thickness. It is obvious that when a TCNQ buffer layer is inserted between Bphen and LiF layer, the current density and luminance of the devices change obviously. When the thickness of TCNQ layer increases from 0 to 6 nm, the current density and luminance of devices A–C increase accordingly at a certain bias. However, with the increase of the thickness of TCNQ layer from 6 to 12 nm, the current density and luminance decrease for devices C–E at the same driving voltage. It can be found that when the thickness of TCNQ layer is 6 nm, the device C achieves the highest current density and luminance at the same voltage among all the devices. From the JVL characteristics of devices, we found that TCNQ layer of the appropriate thickness between Bphen and LiF can enhance electron injection from the cathode to ETL.

Fig. 2

(Color online) EL performances of devices AE. a Current density–voltage–luminance (JVL) characteristics, b current efficiency–current density–power efficiency characteristics, c current efficiency–luminance characteristics

From the energy level of the device (inset of Fig. 7), though energy level difference between the cathode and Bphen remains the same value for the devices with and without TCNQ layer, electrons are more easily injected from the cathode to TCNQ than from the cathode to Bphen. This can be attributed to that the electron mobility of TCNQ and Bphen are about 0.65 and 5 × 10−4 cm2/V s, respectively [10, 20]. The (lowest unoccupied molecular orbital) LUMO level of TCNQ matches well with the work function of the cathode (4.3 eV), so electrons can be smoothly injected from the cathode to TCNQ layer without facing substantial barrier. Furthermore, TCNQ reacts with LiF to form charge transfer complex on the interface to facilitate the electron injection [21]. Besides, a large number of electrons will accumulate at the interface between Bphen and TCNQ layer, so there is a greater chance of electrons being injected into the EML. As shown in Fig. 2b, except for device E, the other devices with TCNQ buffer layer exhibit higher current efficiency than that of the device A. Despite the devices with TCNQ buffer layer presenting almost identical current efficiency, device C exhibits the highest maximum current efficiency of 9.83 cd/A at 7.34 mA/cm2. In addition, we can find that device C presents the highest power efficiency of 7.72 lm/W at the same current density among all the devices. From EL performance of the devices, it is obvious that device C with 6 nm TCNQ layer is optimal.

Figure 2c shows the current efficiency–luminance characteristics of the blue devices. The maximum current efficiency of devices A–E is 6.48, 8.49, 9.83, 7.61, and 8.12 cd/A, respectively. From the luminance at maximum current efficiency to 10,000 cd/m2, the efficiency roll-off of devices A–E is about 31.79, 31.44, 24.72, 37.50, and 54.92%, respectively. We attributed this to the formation of a broader carrier recombination zone and charge-balancing in device C. This will eliminate the carrier accumulation at the interface and prevent excitons quenching.

To clarify the capability of electron injection caused by TCNQ buffer layer, a group of electron-only devices with the structure of ITO/Bphen (50 nm)/TCNQ (x nm)/LiF (0.8 nm)/Al (120 nm) was fabricated, in which x is 0, 3, 6, 9, and 12, respectively. As shown in Fig. 3, the current density of the device with 6 nm TCNQ layer is higher than that of any other electron-only devices at the same voltage. When the voltage reaches to 4 V, the current density of electron-only devices with 0, 3, 6, 9, and 12 nm TCNQ buffer layer is 284.24, 412.93, 520.06, 308.68, and 257.21 mA/cm2, respectively. Thus the insertion of 6 nm TCNQ buffer layer between Bphen and the cathode can effectively enhance the electron injection from the cathode into Bphen and the current density of the device is improved accordingly. Therefore, the incorporation of the 6 nm TCNQ buffer layer could also facilitate the high electron injection for OLED, which matches well with the EL performance of OLED.

Fig. 3

(Color online) Current density–voltage characteristics of the electron-only devices. Inset the configuration of the electron-only devices (x = 0, 3, 6, 9, 12)

The preliminary test of the operational stability of devices was carried out. Figure 4 shows the half-lifetime of devices A and C without encapsulation. The lifetime tests for two devices were measured in glove box at initial luminance (L0) of 5113 and 5295 cd/m2, respectively. It is obvious that device C exhibits a slower rate of luminance degradation than that of device A. The half-lifetime of devices A and C is about 200 and 330 min, respectively. Thus it can be seen that insertion of TCNQ layer between Bphen and LiF layer can enhance the operational stability of the OLEDs. The extension of the device’s lifetime results from the fact that depositing 6 nm TCNQ on Bphen can effectively optimize the surface morphology of Bphen to enhance the stability of the OLEDs.

Fig. 4

(Color online) Half-lifetime of devices A and C. The notation of L/L0 denotes normalized luminance (luminance/initial luminance). Inset the photograph of the optimal OLED at 5 V

In addition, we have found that Bphen is an unstable electron-transporting material with a low glass transition temperature (62 °C) and thus has very low thermal tolerance, which is consistent with that reported by Naito et al. [22]. In the other words, its morphological instability can lead to device degradation due to the re-crystallization of the BPhen films [12, 23]. To verify that 6 nm TCNQ layer can effectively optimize the surface morphology of Bphen, we conducted a comparative analysis on surface morphologies of pure Bphen layer and Bphen layer with 6 nm TCNQ layer on ITO glass substrate. Surface morphologies and cross-sectional SEM images of Bphen and Bphen/TCNQ films are shown in Figs. 5 and 6.

Fig. 5

(Color online) Surface morphologies of Bphen and Bphen with 6 nm TCNQ on it by AFM. a Bphen, b Bphen with 6 nm TCNQ on it

Fig. 6

(Color online) Cross-sectional SEM images of Bphen and Bphen with 6 nm TCNQ on it by SEM. a Bphen, b Bphen with 6 nm TCNQ on it

Figure 5 shows surface morphologies of Bphen and Bphen/TCNQ films by the AFM. The AFM image in Fig. 5 was taken with scan size of 10 × 10 μm. The surface roughness decreases from 7.83 to 2.89 nm when 6 nm TCNQ was deposited on Bphen. Obviously, for the pure Bphen and Bphen with 6 nm TCNQ on it, the difference of RMS roughness of the latter film is smaller than that of the former one, and the surface of Bphen/TCNQ thin film is relatively uniform. Therefore, depositing 6 nm TCNQ on Bphen can contribute to optimize the surface morphology of Bphen which it changes in morphology of the organic interface [24, 25, 26].

Figure 6 shows cross-sectional SEM images of pure Bphen and Bphen with 6 nm TCNQ layer. In addition, in order to make the measurement results more intuitive, the thickness of Bphen and TCNQ layers would be tripled compared with the pristine layers of Bphen and TCNQ. That is to say, the thickness of Bphen is 60 nm and the thickness of TCNQ is 18 nm. It can be found that the surface roughness is receded considerably by depositing TCNQ on Bphen. We expect that the film stability of Bphen/TCNQ can be improved by physical cross-linking that takes place due to the interaction of p–p bond between Bphen and TCNQ, causing inter-molecular interactions in the films [27, 28]. Bphen with 6 nm TCNQ is expected to shorten the inter-molecular distance, so that it will promote efficient transport of charge through adjacent molecules and the smooth surface of the film will provide easy contact between the layers resulting in increasing of the current density in the device [29].

Based on the incorporating non-doped TCNQ as electron buffer layer, the surface morphology quality of Bphen/TCNQ film is improved. To understand the role of 6 nm TCNQ in improving the electron injection and transport characteristics, we further investigate the work function \(q{\varphi _m}\) of the ITO/Bphen film and the ITO/Bphen/TCNQ film which were measured by Kelvin probe (in Fig. 7). The work function values were recorded by scanning 100 points for each sample test. The inset of Fig. 7 shows the schematic illustration of energy level diagram of the device. The work function \(q{\varphi _m}\) (in units of eV) is defined as the minimum energy required for electrons to escape from the Fermi level (Ef) to the vacuum. By depositing 6 nm TCNQ on Bphen, the work function of the deposited film shifted from −6.1943 to −5.8205 eV. In other words, when the TCNQ is introduced, the relative Ef of TCNQ would up-shift about 0.3 eV, which results in the increase of the LUMO level of TCNQ by 0.3 eV. Therefore the electron injection barrier between EBL and ETL was reduced from 1.4 to 1.1 eV. So the incorporation of the 6 nm TCNQ is beneficial for electrons effectively injected from the cathode and transported into the EML, which improves the charge carrier balance in OLEDs, thus reduces the turn-on voltage and enhances the device efficiency. From Fig. 2a, it can be found that the turn-on voltage of device C is lower than that of device A. The turn-on voltage of these OLEDs can be obviously decreased from 3.0 to 2.3 V. This also means the energy levels matching can be optimized by inserting TCNQ layer. On the other hand, the HOMO level of TCNQ is about 7.6 eV, which is 2.3 eV lower than that of the DNCA. The relatively lower HOMO level of TCNQ can block the diffusion of excitons from EML to cathode and reduce the excitons quenching.

Fig. 7

(Color online) The work function of Bphen and Bphen with 6 nm TCNQ are measured by Kelvin probe. Inset schematic illustration of energy level diagram of the device

In addition, we also conducted some experiments about thermal annealing for TCNQ layer. We found that the thermal annealing can improve the quality of surface morphology and enhance the work function when the TCNO film was annealed at 50 °C for 1 h. The specific annealing conditions will be further optimized in the follow-up experiments, by which it is expected to further improve the performances of OLEDs.

4 Conclusions

In summary, the EL performance of the OLEDs is effectively improved by inserting a non-doped TCNQ buffer layer between Bphen layer and LiF/Al cathode. When the thickness of TCNQ layer is optimized to 6 nm, the device can achieve maximum luminance of 26,123 cd/m2, maximum current efficiency of 9.83 cd/A, maximum power efficiency of 7.72 lm/W and half-lifetime of 320 min which increased by 27.32, 57.70, 127.55, and 73.89% respectively, in comparison with the control device without TCNQ layer. The improvement of the performance of the devices can be attributed to the fact that insertion of TCNQ buffer layer can improve the surface morphology quality of the films and enhance the electron injection and transport from the cathode into the organic layer.



The Project was supported by the National Natural Science Foundation of China (Grant No. 60906022), the Natural Science Foundation of Tianjin, China (Grant No. 10JCYBJC01100), the Scientific Developing Foundation of Tianjin Education Commission, China (Grant No. 2011ZD02), the Key Science and Technology Support Program of Tianjin, China (Grant No. 14ZCZDGX00006), and the National High Technology Research and Development Program of China (Grant No. 2013AA014201).


  1. 1.
    C.W. Tang, S.A. VanSlyke, Appl. Phys. Lett. 51(12), 913–915 (1987)CrossRefGoogle Scholar
  2. 2.
    N. Thejokalyani, S.J. Dhoble, Renew. Sust. Energ. Rev 32(5), 448–467 (2014)CrossRefGoogle Scholar
  3. 3.
    A.D. Almeida, B. Santos, B. Paolo, M. Quicheron, Renew. Sust. Energ. Rev 34(34), 30–48 (2014)CrossRefGoogle Scholar
  4. 4.
    T. Higuchi, H. Nakanotani, C. Adachi, Adv. Mater. 27(12), 2019–2023 (2015)CrossRefGoogle Scholar
  5. 5.
    H. Moval, Optik 126(24), 5237–5240 (2015)CrossRefGoogle Scholar
  6. 6.
    D.S.M. Lian, K. Xie, Q.S.M. Yong, J. Soc. Inf. Display. 19(6),453–461 (2011)CrossRefGoogle Scholar
  7. 7.
    C.N. Li, B.W. Xiao, S.Y. Liu, Chin. Phys. Lett. 18(1), 120–122 (2001)CrossRefGoogle Scholar
  8. 8.
    Y. Zhang, J. Lee, S.R. Forrest, Nat. Commun. 5(5), 5008–5008 (2014)CrossRefGoogle Scholar
  9. 9.
    T. Mori, T. Mitsuoka, M. Ishii, H. Fujikawa, Y. Taga, Appl. Phys. Lett. 80(21), 3895–3897 (2002)CrossRefGoogle Scholar
  10. 10.
    L.B. Schein, P.J. Nigrey, Phys. Rev. B 18(6), 2929–2930 (1978)CrossRefGoogle Scholar
  11. 11.
    A.R. Brown, D.W.D. Leeuw, E.J. Lous, E.E. Havinga, Synthetic Met. 66(3), 257–261 (1994)CrossRefGoogle Scholar
  12. 12.
    R. Grover, R. Srivastava, M.N. Kamalasanan, D.S. Mehta, Org. Electron. 13(12), 3074–3078 (2012)CrossRefGoogle Scholar
  13. 13.
    D.W. Chou, K.L. Chen, C.J. Huang, Y.J. Tsao, W.R. Chen, T.H. Meen, Thin Solid Films 536(5), 235–239 (2013)CrossRefGoogle Scholar
  14. 14.
    Y. Jiang, H. Liu, X. Gong, C. Li, R. Qin, H. Ma, J. Power Sources 331, 240–246 (2016)CrossRefGoogle Scholar
  15. 15.
    T. Xiao, F. Fungura, M. Cai, J.W. Anderegg, J. Shinar, R. Shinar, Org. Electron. 14(10), 2555–2563 (2013)CrossRefGoogle Scholar
  16. 16.
    I. Naik, R. Bhajantri, Mater. Today. 3(10), 3608–3613 (2016)CrossRefGoogle Scholar
  17. 17.
    R. Grover, R. Srivastava, M.N. Kamalasanan, D.S. Mehta, J. Lumin. 146(1), 53–56 (2014)CrossRefGoogle Scholar
  18. 18.
    Z.Y. Lü, Z. Lü, J. Xiao, Vacuum 128, 240–243 (2016)CrossRefGoogle Scholar
  19. 19.
    K.S. Lee, D.H. Kim, D.U. Lee, T.W. Kim, Thin Solid Films 521(3), 193–196 (2012)CrossRefGoogle Scholar
  20. 20.
    S. Naka, H. Okada, H. Onnagawa, T. Tsutsui, Appl. Phys. Lett. 76(2), 197–199 (2000)CrossRefGoogle Scholar
  21. 21.
    C. Wäckerlin, C. Iacovita, D. Chylarecka, P. Fesser, T.A. Jung, N. Ballav, Chem. Commun. 47(32), 9146–9148 (2011)CrossRefGoogle Scholar
  22. 22.
    K. Naito, A. Miura, J. Phys. Chem. 97(23), 6240–6248 (1993)CrossRefGoogle Scholar
  23. 23.
    B.W. Dandrade, S.R. Forrest, A.B. Chwang, Appl. Phys. Lett. 83(19), 3858–3860 (2003)CrossRefGoogle Scholar
  24. 24.
    N. Juhari, W.H.A. Majid, A.I. Zainol, Proc. Eng. 53(9), 354–361 (2013)CrossRefGoogle Scholar
  25. 25.
    M. Zubair, M. Mustafa, A. Ali, Y.H. Doh, K.H. Choi, J. Mater. Sci. Mater. Electron. 26(5), 3344–3351 (2015)CrossRefGoogle Scholar
  26. 26.
    S.J. Hwang, M.C. Tseng, K.C. Hwang, H.H. Yu, J. Disp. Technol. 3(3), 253–258 (2007)CrossRefGoogle Scholar
  27. 27.
    C. Ganzorig, K.J. Kwak, K. Yagi, M. Fujihira, Appl. Phys. Lett. 79(2), 272–274 (2001)CrossRefGoogle Scholar
  28. 28.
    N.S. Sariciftci, L. Smilowitz, A.J. Heeger, F. Wudl, Science 258(5087), 1474–1476 (1992)CrossRefGoogle Scholar
  29. 29.
    H.S. Kang, K.N. Park, Y.R. Cho, D.W. Park, Y. Choe, J. Ind. Eng. Chem. 15(5), 752–757 (2009)CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media New York 2017

Authors and Affiliations

  • Juan Zhang
    • 1
    • 2
    • 3
  • Liwen Xin
    • 1
    • 2
    • 3
  • Jian Gao
    • 1
    • 2
    • 3
  • Yang Liu
    • 1
    • 2
    • 3
  • Hongsong Rui
    • 1
    • 2
    • 3
  • Xin Lin
    • 1
    • 2
    • 3
  • Yulin Hua
    • 1
    • 2
    • 3
  • Xiaoming Wu
    • 1
    • 2
    • 3
  • Shougen Yin
    • 1
    • 2
    • 3
  1. 1.School of Materials Science and EngineeringTianjin University of TechnologyTianjinChina
  2. 2.Key Laboratory of Display Materials and Photoelectric DevicesMinistry of EducationTianjinChina
  3. 3.Tianjin Key Laboratory of Photoelectric Materials and DevicesTianjinChina

Personalised recommendations