Device properties of Alq3-based organic light-emitting diodes studied by displacement current measurement Yutaka Noguchi Takamitsu Tamura Hyung Jun Kim Hisao Ishii

Downloaded From: http://photonicsforenergy.spiedigitallibrary.org/ on 05/17/2015 Terms of Use: http://spiedl.org/terms

Device properties of Alq3 -based organic light-emitting diodes studied by displacement current measurement Yutaka Noguchi,a,b,c Takamitsu Tamura,b Hyung Jun Kim,b and Hisao Ishiia,b a

Chiba University, Center for Frontier Science, 1-33 Yayoi-cho, Inage, Chiba 263-8522, Japan [email protected] b Chiba University, Graduate School of Advanced Integration Science, 1-33 Yayoi-cho, Inage-ku, Chiba 263-8522, Japan c PRESTO, Japan Science and Technology Agency, 4-1-8 Honcho Kawaguchi, Saitama 332-0012, Japan

Abstract. Displacement current measurement (DCM) is a simple but powerful tool for exploring charge carrier dynamics in organic semiconductor devices. In the first section, we review the basic concept of DCM and how it detects the charge injection, extraction, accumulation, and trapping behaviors in organic semiconductor devices within a quasistatic regime. Subsequently, we present applications of this technique to investigate the device properties of tris-(8-hydroxyquinolate) aluminum (Alq3 )-based organic light-emitting diodes. We observed that light irradiation during device fabrication induces additional negative space charges and charge traps in the Alq3 layer. In addition, the device containing the illuminated Alq3 film exhibits a lower luminous efficiency and shorter lifetime compared to the device fabricated in dark conditions, possibly because of the additional hole accumulation in the illuminated Alq3 film. DCM detects the formation of charge traps in the aged devices, decay of the negative space charge, and increase in hole injection voltage with device aging. The origins of these behaviors can be attributed to orientation polarization and charge traps in Alq3 film. © 2012 Society of Photo-Optical

Instrumentation Engineers (SPIE). [DOI: 10.1117/1.JPE.2.021214]

Keywords: organic light-emitting diodes; displacement current measurement; charge accumulation; degradation. Paper 12027SS received Mar. 28, 2012; revised manuscript received May 16, 2012; accepted for publication Jun. 25, 2012; published online Jul. 18, 2012.

1 Introduction Displacement current measurement (DCM) is a powerful tool for exploring charge carrier dynamics in organic semiconductor devices.1–5 DCM is a type of capacitance–voltage (C − V) measurement, but it uses a triangular wave as the applied voltage and measures both quasistatic and transient current responses. Although the quasistatic DCM is similar to the conventional C − V measurement, DCM can detect both injected and extracted charge carriers and can consequently detect the amount of trapped charge in the device; this is because the true current (not the root mean square) is measured as a function of time and applied voltage. DCM was specifically introduced to analyze organic diodes.6,7 Egusa et al. conducted pioneering research on DCM to investigate the interface properties of organic light-emitting diodes (OLEDs).1 Ogawa and Ishii expanded this technique to analyze charge carrier dynamics in organic field effect transistors.2,8,9 DCM has mainly been used to analyze quasistatic carrier behavior; however, the DCM signal contains a current originating from transient carrier behavior, and the transient current increases when the sweep rate of the applied triangular wave is high. From the sweep rate dependence of the DCM curve, we can obtain information on the contact and bulk resistance of the device.4 Moreover, the dynamic carrier behavior can be

0091-3286/2012/$25.00 © 2012 SPIE

Journal of Photonics for Energy

021214-1

Downloaded From: http://photonicsforenergy.spiedigitallibrary.org/ on 05/17/2015 Terms of Use: http://spiedl.org/terms

Vol. 2, 2012

Noguchi et al.: Device properties of Alq3 -based organic light-emitting diodes : : :

directly examined using a combined waveform consisting of a constant voltage following a ramp voltage.5 By applying the DCM technique to bilayer devices, which consist of two stacked organic layers between anode and cathode, we can investigate the charge accumulation properties at organic heterointerfaces.3,10 Charge accumulation is an important process for improving the device performance because many device functions originate from the heterointerfaces.11 In multilayer OLEDs, accumulated charges in the emission layer increase the probability of exciton formation.12,13 On the contrary, the charged molecules that accumulated near the emission region can act as exciton quenchers.14,15 Furthermore, the presence of tris-(8-hydroxyquinolate) aluminum (Alq3 ) and 4,4′-bis[N-(1-naphthyl)-N-phenylamino]-biphenyl (α-NPD) cation species is considered responsible for device degradation.16,17 Therefore, detailed understanding of charge accumulation behaviors is required to develop highly efficient multilayer OLEDs with appropriately constructed interfaces. Interface charge is an important parameter to understand the charge accumulation behavior at the organic heterointerface. Brütting et al. first detected the interface charge at the α-NPD/Alq3 interface in the bilayer OLED.18,19 They found that the hole injection voltage (V inj ) of these devices appear lower than the built-in voltage (V bi ), and it shifts to the negative side with increasing Alq3 film thickness. This behavior is well explained by assuming a constant amount of negative fixed charge at the α-NPD/Alq3 interface (typically −1.1 mC∕m2 ). The interface charge defines the least amount of accumulated charge in the operating device because charge accumulation occurs at biases below V bi and the actual current starts flowing after compensating interface charge by accumulated charge. Moreover, Kondakov et al. have reported that the apparent σ int decreases in proportion to the degradation of luminous efficiency.20 Although the origin of the interface charge remained unclear for several years, we recently revealed that the origin of the interface charge is the polarization charge owing to spontaneous ordering of the permanent dipole moment of molecules in the evaporated film.3,10 The orientation polarization was observed in the films of several polar molecules, which are commonly used in OLEDs and organic photo voltaic cells.10 As many organic semiconductor molecules possess permanent dipole moments, interface charge can exist at common organic semiconductor interfaces. In addition to the negative interface charge, we also observed that light absorption of Alq3 during its film deposition process results in the formation of charge traps and negative space charges in the resultant Alq3 film.21 Because these negative space charges induce additional hole accumulation in an operating device, they can affect device performance and degradation properties; however, they have not yet been studied in detail. In the first section of this paper, we review the basic concept of the DCM technique, and how DCM evaluates charge injection, accumulation, traps, and interface charge in bilayer devices in a quasistatic regime. Subsequently, this technique is applied to examine the device properties of Alq3 -based OLED, particularly the luminous efficiency, its degradation process, and the impacts of light-induced space charges on these properties. DCM detects the displacement current; thus it can be a useful tool to examine charge accumulation behaviors even at voltages lower than V bi where the actual current does not flow. We observed that light irradiation during the film deposition results in lower luminous efficiency and shorter lifetime, possibly because of the additional charge formed in the illuminated Alq3 film. Moreover, DCM detects formation of charge traps in the aged devices, decay of the negative space charge, and apparent interface charge density with device aging.

2 Displacement Current Measurement 2.1 Basic Concept of DCM We assume intrinsic organic semiconductors as a dielectric material in the following analyses; the density of thermally activated carriers is extremely low, and carriers are injected from the electrodes of devices. The principle of DCM is simple: a triangular wave bias is applied to the device, and the current response is measured. The total current density flowing through the system (itotal ) consists of the actual and displacement current densities, iact and idis , respectively. In a device consisting of organic semiconductor layers between two parallel electrodes, where a Journal of Photonics for Energy

021214-2

Downloaded From: http://photonicsforenergy.spiedigitallibrary.org/ on 05/17/2015 Terms of Use: http://spiedl.org/terms

Vol. 2, 2012

Noguchi et al.: Device properties of Alq3 -based organic light-emitting diodes : : :

voltage source and a current meter are connected to the electrodes (Fig. 1), idis corresponds to the time derivative of the induced charge on the electrode. For instance, if one positive charge (q) is located at a distance x from the electrode next to a voltage source in a single-layer device with geometrical capacitance C (independent of applied voltage), applied voltage V, and thickness d, then the induced charge on the current meter side, Qelec is given by x −Qelec ¼ CV þ q: d

(1)

The first term of Eq. (1) denotes the charges supplied from the voltage source, while the second term shows the contribution of the induced charges due to the charge in the organic layer. The equation can be generalized by assuming a charge distribution along the x axis as Z d −σ elec ¼ Cdev V þ AðxÞfqnðx; t; VÞ þ ρgdx; (2) 0

where σ elec is the charge density induced on the current meter side electrode; Cdev is the geometrical capacitance of the device per unit area; AðxÞ is the charge induction efficiency [corresponding to x∕d in Eq. (1)], which depends on the device structure including boundary conditions at heterointerfaces; nðx; t; VÞ is the carrier (mobile charge) density; ρ is the trapped (immobile) charge density; and Vð¼ V ex − V bi Þ is the effective applied voltage. V ex is the external applied voltage. The displacement current is then given by idis ¼

∂σ elec ; ∂t

(3)

where idis includes both quasistatic and transient carrier behavior in the device. On the other hand, iact is qnðd; t; VÞμEjx¼d − Dq∂nðx; t; VÞ∕∂xjx¼d , where μ is the carrier mobility, E is the electric field, and D is the diffusion constant. When the carriers move fast as compared to the change in the applied voltage, the device immediately achieves a steady state at a given voltage. We define this low frequency limit as a quasistatic state. Within the quasistatic regime, Eq. (3) corresponds to idis ¼ Capp dV∕dt, where the apparent capacitance, Capp ¼ dσ elec ∕dV, is used. Capp directly provides us with the information relevant to the charge injection/extraction and accumulation properties of the device. At a higher sweep rate (corresponding to a large dV∕dt), the contribution of the transient current increases and the system is no longer considered to be quasistatic. In this situation, analyzing idis is not easy; however, the dynamic response can be derived from idis by 5 holding the sweep voltage at a certain voltage (dV dt ¼ 0). Therefore, we can obtain transport characteristics of the device from the transient current behavior. In the following sections, we focus on the analyses of DCM curves within the quasistatic regime.

2.2 DCM for Quasistatic States To explain typical DCM characteristics, we assume a bilayer device consisting of an organic semiconductor layer and an insulator between two metal electrodes (Fig. 2). When the device with a negligible iact is in a quasistatic state, the measured current, itotal corresponds to Capp dV∕dt. If reverse bias is applied to the device, no carrier injection occurs and both the organic and insulator layers act as insulators [defined as depletion state, (i) in Fig. 2(d)].

Organic layers

+ A

-

Fig. 1 A typical experimental setup for DCM. Organic layers are sandwiched between two electrodes connected to the voltage source (left side) and current meter (right side). Journal of Photonics for Energy

021214-3

Downloaded From: http://photonicsforenergy.spiedigitallibrary.org/ on 05/17/2015 Terms of Use: http://spiedl.org/terms

Vol. 2, 2012

Noguchi et al.: Device properties of Alq3 -based organic light-emitting diodes : : :

(a)

(c)

(ii)

(i)

(iii)

(d)

Organic Insulator

-

+ + + +

(i) Depletion

(b)

Forward

Backward

-

+ + + +

Forward

Forward

Backward

(ii) Intermediate + + + +

-

(iii) Accumulation

Fig. 2 Schematic illustration of (a) the applied triangular wave voltage, (b) current response, (c) DCM curve of bilayer device. The DCM curve represents the capacitance of the depletion and accumulation states, injection/extraction voltage, and amount and polarity of injected/ extracted charge. In this case, the positive charge carrier behavior is detected because the displacement current increases in the positive bias side. (d) (i)–(iii) Schematic illustration of charge location in each state in (c).

Capp is considered as the serially combined capacitance of the organic and insulator layers, that −1 −1 −1 is, C−1 app ¼ Cdev ¼ Corg þ Cins . Here, Corg and Cins are the geometrical capacitances of the organic and insulator layers, respectively. When the carrier injection to the organic layer occurs at voltages larger than V bi , the organic layer is no longer regarded as a simple capacitor and the injected carriers mainly accumulate at the organic/insulator interface within the quasistatic regime [defined as accumulation state, (iii) in Fig. 2(d)]. Consequently, Capp increases to Cins at this voltage region (defined as accumulation state). The total amount of charges per unit area injected from the electrode to the organic layer (σ inj ) is given by Z σ inj ¼

t

idis dt 0 ¼

Z

tinj

V

Capp dV 0 ;

(4)

V inj

where tinj (V inj ) is the time (voltage) when the carrier injection begins. Similarly, charge extraction processes from the organic layer to the electrode are observed in the backward sweep. Here, we define dV∕dt > 0 as a forward sweep, and dV∕dt < 0 as a backward sweep. The amount of the extracted charge, σ ext and the charge extraction voltage, V ext are also obtained from idis . Figure 2(a)–2(c) shows a schematic illustration of an ideal DCM curve obtained by applying one cycle of triangular wave voltage. If a part of the injected charge is trapped in the device during a sweep cycle of the applied voltage, then the amount of extracted charge is less than the injected charge and the DCM curve appears as an asymmetric curve. The total amount of trapped charge can be determined as σ trap ¼ σ inj − σ ext [Fig. 3(a)]. Although σ trap is easily (a)

(c)

(b) 1st sweep 2nd sweep

Fig. 3 (a) Schematic illustration of typical DCM curve when a part of the injected charge is trapped in the device. The DCM curve becomes asymmetric and V inj shifts to a higher side with the subsequent voltage sweep. σ trap can be estimated from σ inj − σ ext . (b) and (c) Schematic energy diagram at V inj with (b) no charge traps and (c) negative interface charge in the device. Journal of Photonics for Energy

021214-4

Downloaded From: http://photonicsforenergy.spiedigitallibrary.org/ on 05/17/2015 Terms of Use: http://spiedl.org/terms

Vol. 2, 2012

Noguchi et al.: Device properties of Alq3 -based organic light-emitting diodes : : :

obtained, further analyses, such as film thickness dependence of V inj , are required to estimate the trapped charge distribution in the film. Because the trapped charges form an electric field to prevent charge injection, the value of V inj concerning the subsequent voltage sweep shifts to the higher side. V inj can be assumed as the voltage that achieves the flat-band condition of the organic layer.19 If the device is electrically neutral, V inj corresponds to V bi [Fig. 3(b)], but the presence of the interface charge results in V inj ¼ V bi þ

σ int dins ; ϵins

(5)

where ϵins is the dielectric permittivity of the insulator and V inj is proportional to dins and σ int , when the interface charge exists at the interface [Fig. 3(c)]. The negative interface charge results in a value of V inj , which is lower than V bi . Furthermore, the flat-band condition throughout the device is achieved when the accumulated holes at the interface compensate for the negative interface charge. We can estimate the interface charge density from the slope of the V inj − dins relation or integrating Capp from V inj to V bi , though the contribution from the intermediate state should be subtracted.10

2.3 Intermediate State between Depletion and Accumulation In the above discussions, we ignored the intermediate state between depletion and accumulation, although this intermediate state often appears in the actual DCM curves even within a quasistatic regime. The curve in the intermediate state can be attributed to the depletion width in the device, which can be formed by low-density tail states into highest occupied molecular orbital (HOMO)–lowest unoccupied molecular orbital (LUMO) gap owing to a structural fluctuation or intentional/unintentional dopants in the film.11,22–24 For example, the 1∕C2 − V plot has often been used to analyze dopant characteristics and estimate the dopant density from its slope.6,25,26 Here, we consider how Capp of the bilayer device changes from the depletion (Cdev ) to accumulation (Cins ) state as a function of the applied voltages within the quasistatic regime. If a sufficient reverse bias is applied to the device, mobile charge carriers in the device are completely extracted and all thermally active dopants are ionized. This is the depletion state and Capp at this voltage range corresponds to Cdev . Then, by increasing the applied bias, charge injection occurs from the electrode to the device, and the injected charges penetrate into the organic layer by filling the inter-gap states with density of nint . The carrier penetration reduces the thickness of the depletion region, and an intermediate capacitance is observed between Cdev and Cins . With assuming that Capp is formed between the carrier front and counter electrode (insulator side), the following equation can be obtained. Z V Z x σ inj ¼ Capp dV 0 ¼ −qnint dx 0 : (6) V inj

0

This equation corresponds to −qnint ¼ Capp

dV ; dx

with a boundary condition of V ¼ V inj at x ¼ 0. On the other hand,
−1 ϵorg −1 þ C−1 ¼ C ; app ins dorg − x

(7)

(8)

where x is the penetration depth of the carrier from the electrode. The equation can be rewritten as −1 x ¼ ϵorg ðC−1 dev − Capp Þ:

(9)

Here, Cdev and Capp ½¼ idis ∕ðdV∕dtÞ
are measurable parameters; thus, if ϵorg is known, x can be obtained at a given voltage V. dV∕dx is numerically obtained from the list of ðx; VÞ, and nint is given by Eq. (7). Meanwhile, by assuming a constant nint , Capp can be derived as Journal of Photonics for Energy

021214-5

Downloaded From: http://photonicsforenergy.spiedigitallibrary.org/ on 05/17/2015 Terms of Use: http://spiedl.org/terms

Vol. 2, 2012

Noguchi et al.: Device properties of Alq3 -based organic light-emitting diodes : : :

 Capp ¼

C−2 dev þ

2ðV − V inj Þ ϵorg qnint

− 1 2

:

(10)

Therefore, the 1∕C2app versusV plot is a straight line with a slope of 2∕ϵorg qnint , which is often used to analyze the dopant density. However, it is available only for the case in which nint in the film is assumed constant. The proposed method enables us to obtain the nint profile from the measured DCM curves within the quasistatic regime.

3 DCM Study on Device Properties of Alq3 -Based OLED 3.1 Typical DCM Curve of Alq 3 -Based OLED at Quasistatic State

100

(a)

ITO 1019 (b)

10 V/s 100 V/s 1 kV/s

80 40

Vacc

20

Vth

Vinj

0

α-NPD

Alq3

1018

60

nint [cm-3]

Current density / sweep rate [nA/cm2/(V/s)]

Figure 4(a) shows typical DCM curves of an Alq3 -based OLED, which has an indium-tin oxide (ITO)/α-NPD (70 nm)/Alq3 (50 nm)/Ca/Al structure at a sweep rate in the range of 10 V∕s − 1 kV∕s. The vertical axis represents the current density normalized by the sweep rate in order to specify the shape of the curve. The DCM curve at 10 V∕s is almost identical to that at 100 V∕s at biases lower than V th ; thus, one can assume that the device was in a quasistatic state at sweep rates in the range of 10 to 100 V∕s. The depletion, intermediate, and accumulation states are clearly seen in these curves. At high reverse voltages (< − 1.4 V ), Capp corresponds to Cdev , indicating that the device is in the depletion state. Subsequently, the current increases at a voltage of −1.4 V (V inj ) owing to hole injection from the ITO substrate. The current increases until the biasing voltage reaches −0.6 V (V acc ). The current again shows constant intensity at voltages up to 2.0 V (V th ). Capp in this bias region corresponds to the capacitance of the Alq3 layer (CAlq ), indicating that hole accumulation occurs at the α-NPD/Alq3 interface. V th corresponds to the threshold voltage of the actual current and it agrees well with V bi of the device (typically ∼1.9 V). This indicates that when the accumulated holes compensate for interface charge, the flat-band condition is achieved throughout the device and the actual current flows through the device. Considering the intermediate state between V inj and V acc , σ int is approximately estimated to be −1.2 mC∕m2 by integrating Capp from V acc to V th . Further details of carrier dynamics in this type of bilayer device have been described elsewhere.3,10,21,27 According to the discussions in Subsec. 2.3, x and nint are obtained from the DCM curve [Fig. 4(b)], where the dielectric constant of 3.3 was used for the α-NPD film. The estimated value of nint increases near the interfaces, namely, ITO/α-NPD and α-NPD/Alq3 , but the tail of nint at the α-NPD/Alq3 interface is longer than that at the ITO/α-NPD interface. Furthermore, significantly higher trap density is observed at the α-NPD/Alq3 interface. A possible origin of this higher trap density can be attributed to the negative interface charge originating from the orientation polarization of Alq3 film.3,10 In the range 15 to 40 nm, the value of nint appears constant

1017

-20 1016

-40

-3

-2

-1 0 1 Voltage [V]

2

3

-10 0 10 20 30 40 50 60 70 80 Distance from electrode [nm]

Fig. 4 (a) Typical DCM curves of the device (ITO/α-NPD ½70 nm
∕Alq3 ½50 nm
∕Ca∕Al structure). The measured sweep rate was 10 V∕s − 1 kV∕s. (b) Depth profile of nint in the α-NPD layer derived from the DCM curve. Journal of Photonics for Energy

021214-6

Downloaded From: http://photonicsforenergy.spiedigitallibrary.org/ on 05/17/2015 Terms of Use: http://spiedl.org/terms

Vol. 2, 2012

Noguchi et al.: Device properties of Alq3 -based organic light-emitting diodes : : :

(∼2 × 1016 cm−3 ). The typical molecular density of the evaporated films of organic semiconductors is ∼1021 cm−3 ; thus, the estimated value of nint is on the order of 10 ppm. This value agrees with the previously reported value for nondoped organic semiconductor films.28,29 These results imply that low-density tail states exist in the α-NPD film, although the exact origin of the tail states is not clear at the moment. For example, although O2 is often considered to be a dopant, recent ultrahigh-sensitivity UPS measurements revealed that structural fluctuation induces the gap states.22,23 We believe that small amounts of electronic states in the energy gap (e.g., HOMO–LUMO gap) play important roles in the charge injection and transport characteristics. We are also conducting high-resolution photoelectron yield spectroscopy experiments to understand the device physics in organic semiconductor devices based on direct observations of electronic structures in the gap states.

3.2 Impacts of Light Irradiation During the Film Formation of Alq 3 The structure of the device being studied is ITO/α-NPD ð50 nmÞ∕Alq3 ð40 nmÞ∕Ca∕Al, with a device active area of 4 mm2 . The organic layers and the Al electrode were successively formed on an ITO-coated glass substrate by a conventional vacuum sublimation technique at a typical base pressure of ∼10−4 Pa. Two types of devices were fabricated: dark and illuminated devices. The dark device was fabricated with preventing ambient light throughout the film deposition processes, whereas the Alq3 layer of the illuminated device was formed under 400 nm-light irradiation with an intensity of ∼30 μW∕cm2 . We have examined the film growth of Alq3 under light irradiation with wavelengths of 400, 525, and 630 nm.21 Compared to the others, 400 nm-light-induced significant changes in the charge accumulation properties of the resultant Alq3 film, because the absorption peak of Alq3 exists around 400 nm. Thus we used 400 nmlight to examine the impacts on the device properties. All films were deposited without iongauge operation, because the emissions from the filament can change the device properties.21 The devices were then encapsulated in a glove box without being exposed to air. Current density-voltage-luminance (J-V-L) measurements, DCM, and electrical device aging were performed in the dark at room temperature. The J-V-L curves were measured with a source-measure unit (Keithley 237) and a luminance meter (Topcon, Bm-9). For DCM, we used a measurement system for ferroelectrics (Toyo Corp., FCE-1). The DCM curves were measured over five sequential cycles of the triangular bias sweep. Both dark and illuminated devices were aged under a constant current density of 47.2 mA∕cm2 . DCM curves during the device aging process were measured while temporally stopping the aging current. Figure 5 shows typical luminous efficiency vs. current density curves of the pristine dark and pristine illuminated devices. The curves clearly indicate that the luminous efficiency of the illuminated device is lower than that of the dark device. A typical luminous efficiency of the illuminated device was less than half of the dark device. The origins of the lower device performance in illuminated devices are still under investigation; however, the space charges and charge traps formed in the illuminated Alq3 film can be responsible.21

Luminous efficiency [cd/A]

2.5 2 1.5 1 0.5 Dark Illuminated

0 0

10 20 30 40 50 60 Current density [mA/cm2]

70

Fig. 5 Typical luminous efficiency vs. current density curves of pristine-dark and illuminated devices. The sample structure was ITO/α-NPD (50 nm)/Alq3 (40 nm)/Ca/Al. Journal of Photonics for Energy

021214-7

Downloaded From: http://photonicsforenergy.spiedigitallibrary.org/ on 05/17/2015 Terms of Use: http://spiedl.org/terms

Vol. 2, 2012

Noguchi et al.: Device properties of Alq3 -based organic light-emitting diodes : : :

Figure 6(a) and 6(b) shows the DCM curves of the pristine devices fabricated under dark conditions and light irradiation, respectively. There are some specific features in the DCM curves of the illuminated device; a peak structure at 2 V (V p ), lower V inj and V th , and a slight shift of V inj in the second cycle. The voltage shift suggests that a part of the injected holes during the first voltage sweep was trapped in the device as previously described. Note that Capp appeared at the biases between V th , and V p was larger than the estimated capacitance of the Alq3 layer (CAlq ∼ 70 nF∕cm2 ). This indicates that the dielectric thickness of the Alq3 layer was reduced, that is, the injected holes from the ITO electrode proceeded into the Alq3 layer across the α-NPD/ Alq3 interface and accumulated in the bulk of the Alq3 film. V p corresponds to the V th of the dark device and the typical V bi of a ITO/α-NPD/Alq3 /Ca/Al device. In addition, V inj of the illuminated device appears more at negative side compared to that of the dark device. The results suggest the presence of additional negative charges in the illuminated Alq3 layer. The negative space charge results in an additional hole accumulation at biases between V th and V p , and thus Capp , which has a value larger than CAlq , was observed.21 Figure 7(a) and 7(b) shows the DCM curves of dark and illuminated devices at various aging times, respectively. For the dark device, V inj shifts to the positive side as a function of aging time, though V th remains almost constant (∼2.0 V), indicating that the apparent interface charge density at the α-NPD/Alq3 interface decreased in the aged device as previously reported by Kondakov et al.20 Similar behavior is also observed in the illuminated device; however, in addition to the V inj shift, the peak structure in the DCM curves decayed and finally disappeared without changing the position of V p . The results indicate that the device aging diminishes not only the interface charge but also the negative space charge in the illuminated Alq3 film. The distorted shape, which is particularly pronounced at the negative biases in the DCM curves of the aged-dark device, is attributed to a leakage current through the device. Figure 8 shows the decay curves representing luminous efficiency of the aged-dark and illuminated devices, where the luminous efficiency is normalized by that of the pristine devices. After continuing the initial degradation process for a span of 10−1 h, the luminance loss of the illuminated device is accelerated much faster than that of the dark device. At the aging time of 20 h, the normalized efficiency of the illuminated device decreased to 73%, while that of the dark device remains over 90%. These results clearly demonstrate that the light 200

(a) 150 Pristine-dark

1st cycle 2nd cycle

Pristine-illuminated

Vacc

100

0

1st cycle 2nd cycle

Vacc Vth

50 Current density [nA/cm2]

(b)

Vth

Vinj

Vinj

Vp

-50 -100 200

(c) 150 Aged-dark

1st cycle 2nd cycle

(d) Aged-illuminated

1st cycle 2nd cycle

100 50 0 -50 -100

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 Voltage [V]

Fig. 6 Typical DCM curves of pristine devices fabricated under (a) dark conditions and (b) light irradiation (400 nm, 30 μW∕cm2 ). V inj , V acc , V th , and V p are indicated in the figures. The DCM curve of the illuminated device shows a slight change in the second sweep, indicating the presence of charge traps. (c) and (d) The DCM curves of the aged-(c) dark and (d) illuminated devices, respectively. V inj for the both devices shift to the higher voltages and charge traps are observed. All curves were measured at a sweep rate of 1 V∕s. Journal of Photonics for Energy

021214-8

Downloaded From: http://photonicsforenergy.spiedigitallibrary.org/ on 05/17/2015 Terms of Use: http://spiedl.org/terms

Vol. 2, 2012

Noguchi et al.: Device properties of Alq3 -based organic light-emitting diodes : : :

200

(a)

Aging time 0h 25 h 112 h 401 h 784 h

150 100

Current density [nA/cm2]

50 0

Aging

-50 -100 200

(b)

Aging time 0 min 1 min 60 min 3h 11 h 19 h

150 100 50 0

Aging

-50 -100

-2

-1.5

-1

-0.5 0 0.5 Voltage [V]

1

1.5

2

Fig. 7 The DCM curves of the (a) dark and (b) illuminated devices at various aging time. V inj increased with aging time in both devices. All curves were measured at a sweep rate of 1 V∕s.

Normalized luminance

1.1

Dark Illuminated

1 0.9 0.8 0.7 0.6 10-3

10-2

10-1 1 Aging time [h]

10

Fig. 8 The normalized luminance under constant aging current operation. The aging current was 47.2 mA∕cm2 for both dark and illuminated devices. The rapid decay after 10−1 h is seen in the illuminated device.

irradiation during the Alq3 film deposition significantly enhances the degradation under a constant current condition of the resultant device. There are two possible explanations for the decay of apparent interface charge with device degradation. First is generation of hole traps and the second is loss of orientation polarization in the Alq3 film (PAlq ). Kondakov et al. reported that the transition voltage, corresponding to V inj in our DCM curves, increased proportionally to the loss of luminous efficiency.20 They explained this behavior as follows; at first, the electrical aging creates hole traps near the α-NPD/Alq3 interface, and then the trapped holes compensate for the negative interface charge. These states can act as nonradiative recombination centers, and thus the apparent interface charge density can monitor the luminance loss. Here, the cycle dependence of DCM curves actually indicates the presence of the hole traps generated in the aged devices [Fig. 6(c) and 6(d)]. Furthermore, we confirmed that hole traps increased with aging time. On the other hand, the negative interface charge originates from the polarization charge induced by spontaneous ordering of the dipole moment of Alq3 in the evaporated film.3,10 Thus, if the electrical aging has disordered the molecular orientation in the Alq3 film, then the interface charge should be reduced and as a result the Journal of Photonics for Energy

021214-9

Downloaded From: http://photonicsforenergy.spiedigitallibrary.org/ on 05/17/2015 Terms of Use: http://spiedl.org/terms

Vol. 2, 2012

Noguchi et al.: Device properties of Alq3 -based organic light-emitting diodes : : :

DOS may broaden because of the dipolar disorder effect.30,31 The reduction of the interface charge results in fewer accumulated holes at the α-NPD/Alq3 interface in the operating device. It can result in a lower probability of exciton formation in the device and moreover, the DOS broadening can cause a lower conductivity and charge traps. Although experimental evidence for the dipole moment disordering in the aged device is required, either or both scenarios might occur in aged devices. The degradation of the luminous efficiency in the illuminated device was significantly faster than that in the dark device (Fig. 8). This rapid degradation can be attributed to the negative space charge formed in the Alq3 layer. The presence of a negative space charge induces an additional amount of hole accumulation in the bulk of the Alq3 layer, as shown in the DCM curves. As previously reported by Aziz et al., instability of a cationic Alq3 is a root of intrinsic degradation in Alq3 -based OLEDs, and thus the additional hole accumulation can enhance device degradation.16 Here, we consider two possible origins for the negative space charges in the Alq3 film. First is the polarization charge owing to the spatial inhomogeneity of orientation polarization of the Alq3 film (divPAlq ), and second is the real charge owing to deep electron traps in the Alq3 film.21 At the present stage, the dominant origin has not been determined conclusively. In the case of the polarization charge, Alq3 itself is neutral before the hole accumulation. Thus, the cationic Alq3 species are additionally created by the hole accumulation in the illuminated Alq3 layer. On the other hand, different situations should be considered if deep electron traps hold the negative space charge, because the accumulated holes might neutralize the electron traps but no additional Alq3 cation are created. In this case, the negatively charged traps instead of the Alq3 cation can be responsible for the rapid device degradation. We are proceeding with further investigations to clarify the details such as direct observation of dipole moment ordering and the trap states in illuminated Alq3 film.

4 Conclusion In this study, we review the basic concept of the DCM technique and how to analyze a DCM curve of bilayer devices within the quasistatic regime. DCM is a powerful tool to evaluate charge carrier dynamics, particularly injection, extraction, accumulation, and traps in organic semiconductor devices. By applying this technique to Alq3 -based OLED, we examine the device properties of the luminous efficiency and its degradation process as well as the impacts of light-induced space charges in an Alq3 film on these properties. We observed that light irradiation during device fabrication induces additional negative space charges and charge traps in the Alq3 layer. In addition, the device exhibits a lower luminous efficiency and shorter lifetime compared to the device fabricated in dark conditions, possibly because of the additional hole accumulation in the bulk region of the Alq3 film. DCM detects formation of charge traps in the aged devices, decay of the negative space charge, and apparent interface charge density with device aging. The origins of these behaviors can be attributed to orientation polarization and charge traps in Alq3 film. Investigations regarding direct observation of dipole moment ordering and the trap states in the illuminated Alq3 film are in progress.

Acknowledgments We would like to thank Prof. Wolfgang Brütting (Universität Augsburg) for fruitful discussions. We thank Nippon Steel Chemical Co., Ltd. for providing α-NPD and Alq3 molecules. This research is supported by the Japan Society for the Promotion of Science (JSPS) through the “Funding Program for World-Leading Innovative R&D on Science and Technology (FIRST Program)” initiated by the Council for Science and Technology Policy (CSTP), the Global-COE Project of Chiba University (Advanced School for Organic Electronics), and KAKENHI (Grant Nos. 21245042 and 22750167).

References 1. S. Egusa et al., “Carrier injection characteristics of organic electroluminescent devices,” Jpn. J. Appl. Phys. 33, 2741–2745 (1994), http://dx.doi.org/10.1143/JJAP.33.2741. Journal of Photonics for Energy

021214-10

Downloaded From: http://photonicsforenergy.spiedigitallibrary.org/ on 05/17/2015 Terms of Use: http://spiedl.org/terms

Vol. 2, 2012

Noguchi et al.: Device properties of Alq3 -based organic light-emitting diodes : : :

2. S. Ogawa et al., “Carrier injection characteristics in organic field effect transistors studied by displacement current measurement,” Jpn. J. Appl. Phys. 42(2), L1275–L1278 (2003), http:// dx.doi.org/10.1143/JJAP.42.L1275. 3. Y. Noguchi et al., “Threshold voltage shift and formation of charge traps induced by light irradiation during the fabrication of organic light-emitting diodes,” Appl. Phys. Lett. 92(20), 203306 (2008), http://dx.doi.org/10.1063/1.2936084. 4. Y. Noguchi et al., “Higher resistance to hole injection and electric field distribution in organic light-emitting diodes with copper phthalocyanine interlayer,” Jpn. J. Appl. Phys. 49, 01AA01 (2010), http://dx.doi.org/10.1143/JJAP.49.01AA01. 5. Y. Tanaka et al., “Displacement current measurement of a pentacene metal-insulatorsemiconductor device to investigate both quasi-static and dynamic carrier behavior using a combined waveform,” Org. Electron. 12(9), 1560–1565 (2011), http://dx.doi.org/10 .1016/j.orgel.2011.06.002. 6. A. J. Twarowski and A. C. Albrecht, “Depletion layer studies in organic films: low frequency capacitance measurements in polycrystalline tetracene,” J. Chem. Phys. 70(5), 2255–2262 (1979), http://dx.doi.org/10.1063/1.437729. 7. K. Iriyama et al., “Photoelectric effects in organic diodes,” Jpn. J. Appl. Phys. 19(Suppl. 19-2), 173–177 (1980). 8. S. Ogawa et al., “Photoinduced doping effect of pentacene field effect transistor in oxygen atmosphere studied by displacement current measurement,” Appl. Phys. Lett. 86(25), 252104 (2005), http://dx.doi.org/10.1063/1.1949281. 9. S. Ogawa et al., “Displacement current measurement as a tool to characterize organic field effect transistors,” Synth. Met. 153(1–3), 253–256 (2005), http://dx.doi.org/10.1016/j .synthmet.2005.07.267. 10. Y. Noguchi et al., “Charge accumulation at organic semiconductor interfaces due to a permanent dipole moment and its orientational order in bilayer devices,” J. Appl. Phys. 111(11), 114508 (2012). 11. N. Koch, “Organic electronic devices and their functional interfaces,” Chem. Phys. Chem. 8(10), 1438–1455 (2007), http://dx.doi.org/10.1002/(ISSN)1439-7641. 12. G. G. Malliaras and J. C. Scott, “The roles of injection and mobility in organic light emitting diodes,” J. Appl. Phys. 83(10), 5399–5403 (1998), http://dx.doi.org/10.1063/1.367369. 13. B. Ruhstaller et al., “Transient and steady-state behavior of space charges in multilayer organic light-emitting diodes,” J. Appl. Phys. 89(8), 4575–4586 (2001), http://dx.doi .org/10.1063/1.1352027. 14. R. H. Young, C. W. Tang, and A. P. Marchetti, “Current-induced fluorescence quenching in organic light-emitting diodes,” Appl. Phys. Lett. 80(5), 874–876 (2002), http://dx.doi.org/10 .1063/1.1445271. 15. T. Haskins et al., “Charge-induced luminescence quenching in organic light-emitting diodes,” Chem. Mater. 16(23), 4675–4680 (2004), http://dx.doi.org/10.1021/cm049538k. 16. H. Aziz et al., “Degradation mechanism of small molecule-based organic light-emitting devices,” Science 283(5409), 1900–1902 (1999), http://dx.doi.org/10.1126/science.283 .5409.1900. 17. D. Y. Kondakov and R. H. Young, “Variable sensitivity of organic light-emitting diodes to operation-induced chemical degradation: Nature of the antagonistic relationship between lifetime and efficiency,” J. Appl. Phys. 108(7), 074513 (2010), http://dx.doi.org/10 .1063/1.3483251. 18. S. Berleb, W. Brütting, and G. Paasch, “Interfacial charges in organic hetero-layer light emitting diodes probed by capacitance-voltage measurements,” Synth. Met. 122(1), 37–39 (2001), http://dx.doi.org/10.1016/S0379-6779(00)01356-4. 19. W. Brütting, S. Berleb, and A. G. Mückl, “Device physics of organic light-emitting diodes based on molecular materials,” Org. Electron. 2(1), 1–36 (2001), http://dx.doi.org/10.1016/ S1566-1199(01)00009-X. 20. D. Y. Kondakov et al., “Nonradiative recombination centers and electrical aging of organic light-emitting diodes: Direct connection between accumulation of trapped charge and luminance loss,” J. Appl. Phys. 93(2), 1108–1119 (2003), http://dx.doi.org/10.1063/1 .1531231. Journal of Photonics for Energy

021214-11

Downloaded From: http://photonicsforenergy.spiedigitallibrary.org/ on 05/17/2015 Terms of Use: http://spiedl.org/terms

Vol. 2, 2012

Noguchi et al.: Device properties of Alq3 -based organic light-emitting diodes : : :

21. Y. Noguchi et al., “Light- and ion-gauge-induced space charges in tris-(8-hydroxyquinolate) aluminum-based organic light-emitting diodes,” Appl. Phys. Lett. 96(14), 143305 (2010), http://dx.doi.org/10.1063/1.3374405. 22. T. Sueyoshi et al., “Low-density band-gap states in pentacene thin films probed with ultrahigh-sensitivity ultraviolet photoelectron spectroscopy,” Appl. Phys. Lett. 95(18), 183303 (2009), http://dx.doi.org/10.1063/1.3258351. 23. T. Sueyoshi et al., “Band gap states of copper phthalocyanine thin films induced by nitrogen exposure,” Appl. Phys. Lett. 96(9), 093303 (2010), http://dx.doi.org/10.1063/1.3332577. 24. I. Lange et al., “Band bending in conjugated polymer layers,” Phys. Rev. Lett. 106(21), 216402 (2011), http://dx.doi.org/10.1103/PhysRevLett.106.216402. 25. J. H. Burroughes, C. A. Jones, and R. H. Friend, “New semiconductor device physics in polymer diodes and transistors,” Nature 335(6186), 137–141 (1988), http://dx.doi.org/10 .1038/335137a0. 26. F. Ebisawa, T. Kurokawa, and S. Nara, “Electrical properties of polyacetylene/polysiloxane interface,” J. Appl. Phys. 54(6), 3255–3259 (1983), http://dx.doi.org/10.1063/1.332488. 27. N. Sato et al., “Mechanism of hole accumulation at α-NPD/Alq3 interface studied by displacement current measurement,” Proc. SPIE 7051, 70511S (2008). 28. C. K. Chan and A. Kahn, “N-doping of pentacene by decamethylcobaltocene,” Appl. Phys. A 95(1), 7–13 (2009), http://dx.doi.org/10.1007/s00339-008-4997-x. 29. J. Laubender et al., “Influence of oxygen and air on the characteristics of organic lightemitting devices studied by in vacuo measurements,” Synth. Met. 111–112, 373–376 (2000). 30. M. A. Baldo, Z. G. Soos, and S. R. Forrest, “Local order in amorphous organic molecular thin films,” Chem. Phys. Lett. 347(4–6), 297–303 (2001), http://dx.doi.org/10.1016/S00092614(01)01063-6. 31. M. A. Baldo and S. R. Forrest, “Interface-limited injection in amorphous organic semiconductors,” Phys. Rev. B 64(8), 085201 (2001), http://dx.doi.org/10.1103/PhysRevB.64 .085201. Yutaka Noguchi received his doctor degree in engineering from Tokyo Institute of Technology, Japan, in 2004 under the supervision of professor Mitsumasa Iwamoto. From 2004 to 2007, he was a postdoctoral fellow at KARC, National Institute of Information and Communications Technology, Japan. He has joined professor Hisao Ishii’s group as an assistant professor in Center for Frontier Science, Chiba University, Japan since 2007. He has also been a PRESTO researcher of Japan Science and Technology Agency since 2009. His research interests include charge carrier behaviors in organic semiconductor devices, charge transport properties of single-molecular junctions, and development of organic single-electron devices. Takamitsu Tamura received his bachelor degree in engineering from Chiba University, Japan, in 2010. He is a graduate student of Graduate School of Advanced Integration Science, Chiba University. He is currently studying on the degradation mechanisms in Alq3 -based organic lightemitting diode with focusing on the orientation polarization of the organic films.

Hyung Jun Kim studied precision mechanical engineering at Kangwon National University, Korea. From 2004 to 2010, he was employed as customer service engineer for mass production of OLED evaporation equipment and sales engineer for electric materials. Since 2011, he entered a master course of Graduate School of Advanced Integration Science, Chiba University. His research focused on heterojunction interfaces, carrier behaviors, and degradation mechanisms in OLED.

Journal of Photonics for Energy

021214-12

Downloaded From: http://photonicsforenergy.spiedigitallibrary.org/ on 05/17/2015 Terms of Use: http://spiedl.org/terms

Vol. 2, 2012

Noguchi et al.: Device properties of Alq3 -based organic light-emitting diodes : : :

Hisao Ishii received his MS and PhD degrees in chemistry from the University of Tokyo, in 1988 and 1991, respectively. In 1991, he joined the Solid State Chemistry Laboratory, Nagoya University, Nagoya, Japan, where he investigated the electronic structures of various organic/metal interfaces by using photo emission spectroscopy. In 2002, he moved to the research institute of electrical communication, Tohoku University, Sendai, Japan, and started to study organic device physics by using displacement current measurement. Since 2006, as professor, he is investigating both the electronic structure and device properties of organic electronics to understand organic semiconductor physics.

Journal of Photonics for Energy

021214-13

Downloaded From: http://photonicsforenergy.spiedigitallibrary.org/ on 05/17/2015 Terms of Use: http://spiedl.org/terms

Vol. 2, 2012