Organic semiconductor heterojunctions and its application in organic light-emitting diodes

Springer Series in Materials Science 250

Dongge MaYonghua Chen

Organic Semiconductor Heterojunctions and Its Application in Organic Light-Emitting Diodes

Springer Series in Materials Science

Volume 250

Series editors

Robert Hull, Troy, USAChennupati Jagadish, Canberra, AustraliaYoshiyuki Kawazoe, Sendai, JapanRichard M. Osgood, New York, USAJürgen Parisi, Oldenburg, GermanyTae-Yeon Seong, Seoul, Republic of Korea (South Korea)Udo W. Pohl, Berlin, GermanyShin-ichi Uchida, Tokyo, JapanZhiming M. Wang, Chengdu, China

The Springer Series in Materials Science covers the complete spectrum of materialsphysics, including fundamental principles, physical properties, materials theory anddesign. Recognizing the increasing importance of materials science in future devicetechnologies, the book titles in this series reflect the state-of-the-art in understand-ing and controlling the structure and properties of all important classes of materials.

More information about this series at http://www.springer.com/series/856

Dongge Ma • Yonghua Chen

Organic SemiconductorHeterojunctions and ItsApplication in OrganicLight-Emitting Diodes

123

Dongge MaState Key Laboratory of LuminescentMaterials and Devices,

Institute of Polymer OptoelectronicMaterials and Devices

South China University of TechnologyGuangzhouChina

Yonghua ChenKey Laboratory of Flexible Electronics andInstitute of Advanced Materials, JiangsuNational Synergetic Innovation Center forAdvanced Materials

Nanjing Tech UniversityNanjingChina

ISSN 0933-033X ISSN 2196-2812 (electronic)Springer Series in Materials ScienceISBN 978-3-662-53693-3 ISBN 978-3-662-53695-7 (eBook)https://doi.org/10.1007/978-3-662-53695-7

Library of Congress Control Number: 2017952909

© Springer-Verlag GmbH Germany 2017This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or partof the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmissionor information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilarmethodology now known or hereafter developed.The use of general descriptive names, registered names, trademarks, service marks, etc. in thispublication does not imply, even in the absence of a specific statement, that such names are exempt fromthe relevant protective laws and regulations and therefore free for general use.The publisher, the authors and the editors are safe to assume that the advice and information in thisbook are believed to be true and accurate at the date of publication. Neither the publisher nor theauthors or the editors give a warranty, express or implied, with respect to the material contained herein orfor any errors or omissions that may have been made. The publisher remains neutral with regard tojurisdictional claims in published maps and institutional affiliations.

Printed on acid-free paper

This Springer imprint is published by Springer NatureThe registered company is Springer-Verlag GmbH, DEThe registered company address is: Heidelberger Platz 3, 14197 Berlin, Germany

Preface

The developers and engineers who are doing in products in display and lighting oforganic light-emitting diodes (OLEDs) based on organic semiconductors will finduseful information on the design principles of high-performance OLEDs. This bookwill also serve as helpful and valuable support and reference to graduate students tofreshly enter this field from synthetic chemistry, electrical engineering, appliedphysics, and material science.

The semiconductor heterojunctions are the basic for constructinghigh-performance optoelectronic devices. During past decades, more and moreorganic semiconductors are utilized to fabricate the heterojunction devices, espe-cially the OLEDs. This subject has attracted great attention and evoked many newphenomena and interpretations in the field. This book, organic semiconductorheterojunctions and its application in OLEDs, systematically introduces theimportant aspects of organic semiconductor heterojunctions, including the basicconcepts and electrical properties. The application of organic semiconductorheterojunctions in OLEDs, as charge injector and as charge generation layer, iscomprehensively discussed in this work. This important application is based on thelow dielectric constant of organic semiconductors and the weak non-covalentelectronic interactions between organic semiconductors, thus easily forming accu-mulation heterojunction. As we know, the accumulation-type space charge region ishighly conductive, which is an important property for highly efficient chargegeneration in this application as charge injector and charge generation layer inOLEDs. This book can serve as a useful reference for researchers and a textbook forgraduate students focusing on the studies and development of OLED for displayand lighting.

Guangzhou, China Dongge MaNanjing, China Yonghua Chen

v

Contents

1 Physics Basis of Organic Semiconductor Heterojunctions . . . . . . . . . 11.1 Basic Concept of Heterojunctions. . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Theory of Heterojunctions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.2.1 Diffusion Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71.2.2 Emission Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91.2.3 Tunneling Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101.2.4 Emission Recombination Model . . . . . . . . . . . . . . . . . . . . . 111.2.5 Tunneling Recombination Model . . . . . . . . . . . . . . . . . . . . 11

1.3 Energy Band Profiles of Heterojunctions . . . . . . . . . . . . . . . . . . . . 131.3.1 Profiles of Abrupt Anisotype P/N Heterojunctions . . . . . . . 141.3.2 Profiles of Abrupt Anisotype N/P Heterojunctions . . . . . . . 17

1.4 Basic Properties of Organic Heterojunctions. . . . . . . . . . . . . . . . . . 201.5 Brief Description of Organic Heterojunction Application

in Organic Light-Emitting Diodes. . . . . . . . . . . . . . . . . . . . . . . . . . 31References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

2 Electrical Properties of Organic Semiconductor Heterojunctions . . .. . . . 372.1 Current–Voltage Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . 372.2 Capacitance–Voltage Characteristics . . . . . . . . . . . . . . . . . . . . . . . . 462.3 Charge Transport Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 512.4 Charge Generation Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

3 Organic Semiconductor Heterojunctions as Charge Injectorin Organic Light-Emitting Diodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 593.1 Basic Condition as Charge Injector . . . . . . . . . . . . . . . . . . . . . . . . 593.2 As Hole Injector for High-Efficiency Organic

Light-Emitting Diodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 613.3 As Electron Injector for High-Efficiency Organic

Light-Emitting Diodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

vii

3.4 As Hole and Electron Injectors for High-Efficiency OrganicLight-Emitting Diodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

4 Organic Semiconductor Heterojunctions as Charge GenerationLayer in Tandem Organic Light-Emitting Diodes . . . . . . . . . . . . . . . 894.1 Basic Condition as Charge Generation Layer . . . . . . . . . . . . . . . . . 894.2 Doped-N/Doped-P Heterojunction as Charge Generation Layer

for High-Efficiency Tandem Organic Light-Emitting Diodes . . . . . 924.3 N/P Bilayer Heterojunction as Charge Generation Layer for

High-Efficiency Tandem Organic Light-Emitting Diodes . . . . . . . . 974.4 N:P Bulk Heterojunction as Charge Generation Layer

for High-Efficiency Tandem Organic Light-Emitting Diodes . . . . . 1154.5 N/N:P/P Composited Heterojunction as Charge Generation Layer

for High-Efficiency Tandem Organic Light-Emitting Diodes . . . . . 118References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

5 Tandem White Organic Light-Emitting Diodes Basedon Organic Semiconductor Heterojunctions . . . . . . . . . . . . . . . . . . . . 1275.1 Basic Structures of Tandem White Organic Light-Emitting

Diodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1275.2 Fluorescence Tandem White Organic Light-Emitting Diodes . . . . . 1325.3 Phosphorescence Tandem White Organic

Light-Emitting Diodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1365.4 Fluorescence/Phosphorescence Hybrid Tandem White Organic

Light-Emitting Diodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1425.5 Applications of Tandem White Organic Light-Emitting Diodes

in Display and Lighting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150

Subject Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153

viii Contents

Chapter 1Physics Basis of Organic SemiconductorHeterojunctions

1.1 Basic Concept of Heterojunctions

A heterojunction is the interface that occurs between two layers or regions ofdissimilar crystalline semiconductors. These semiconducting materials haveunequal band gaps as opposed to a homojunction. It is often advantageous toengineer the electronic energy bands in many solid state device applicationsincluding light emission diodes (LEDs), semiconductor lasers, solar cells, andtransistors to name a few. The combination of multiple heterojunctions together in adevice is called a heterostructure although the two terms are commonly usedinterchangeably. The requirement that each material be a semiconductor withunequal band gaps is somewhat loose especially on small length scales whereelectronic properties depend on spatial properties. A more modern definition ofheterojunction is the interface between any two solid-state materials, includingcrystalline and amorphous structures of metallic, insulating, fast ion conductor andsemiconducting materials, even widely used organic semiconductors. In 2000, theNobel Prize in physics was awarded jointly to Herbert Kroemer (University ofCalifornia, Santa Barbara, California, USA) and Zhores I. Alferov (Ioffe Institute,Saint Petersburg, Russia) for “developing semiconductor heterostructures used inhigh-speed- and optoelectronics”.

Semiconductors are the foundation of constructing the heterojunctions.A semiconductor is a substance, usually a solid chemical element or compound thatcan conduct electricity under some conditions but not others, making it a goodmedium for the control of electrical current. Its conductance varies depending onthe current or voltage applied to a control electrode.

Semiconductors are divided into two types, P-type and N-type. An N-typesemiconductor carries current mainly in the form of negatively charged electrons, ina manner similar to the conduction of current in a wire. A P-type semiconductorcarries current predominantly as electron deficiencies called holes. A hole has apositive electric charge, equal and opposite to the charge on an electron. In a

© Springer-Verlag GmbH Germany 2017D. Ma and Y. Chen, Organic Semiconductor Heterojunctions and Its Application inOrganic Light-Emitting Diodes, Springer Series in Materials Science 250,https://doi.org/10.1007/978-3-662-53695-7_1

1

semiconductor material, the flow of holes occurs in a direction opposite to the flowof electrons. Broadly speaking, semiconductors include two kinds of inorganic andorganic semiconductors. For inorganic semiconductors, its conductivity type isrealized by doping, whereas the conductance type of organic semiconductors isdetermined by their intrinsic property, which can directly conduct electrons andholes without doping.

In semiconductors, Fermi level is an important parameter, which is defined as theenergy point where the probability of occupancy by an electron is exactly 50%, or0.5. It determines the conductance type and also the electronic properties of asemiconductor. Therefore, a precise understanding of the Fermi level is essential toan understanding of solid-state physics [1].

In an intrinsic semiconductor, n = p. If we use the band-symmetry approxima-tion, which assumes that there are equal number of states in equal-sized energybands at the edges of the conduction and valence bands, n = p implies that there isan equal chance of finding an electron at the conduction band edge as there is offinding a hole at the valence band edge. It can deduce that the Fermi level Ef mustbe in the middle of the band gap for an intrinsic semiconductor. For an N-typesemiconductor, there are more electrons in the conduction band and the holes in thevalence band. This also implies that the probability of finding an electron near theconduction band edge is larger than the probability of finding a hole at the valenceband edge. Therefore, the Fermi level is closer to the conduction band in an N-typesemiconductor. For a P-type semiconductor, there are more holes in the valenceband than the electrons in the conduction band. This also implies that the proba-bility of finding an electron near the conduction band edge is smaller than theprobability of finding a hole at the valence band edge. Therefore, the Fermi level iscloser to the valence band in an P-type semiconductor. For organic semiconductors,the rule is also applicable [2].

By definition, N-type semiconductor has an excess of free electrons compared tothe P-type region, and P-type has an excess of holes compared to the N-type region.Therefore, when N-doped and P-doped pieces of semiconductors are placed toge-ther to form a junction, electrons migrate into the P-side and holes migrate into theN-side. Departure of an electron from the N-side to the P-side leaves a positivedonor ion behind on the N-side, and likewise the hole leaves a negative acceptor ionon the P-side. The two charged regions are called as space charge region. The netresult in this space charge region is that the diffused electrons and holes are gone,leaving behind the charged ions adjacent to the interface in a region with no mobilecarriers; therefore, in this case, the space charge region is also called as depletionregion. The result of positive and negative charges in this region creates an internalelectric field with direction from positive charge region to negative charge regionthat provides a force opposing the continued exchange of charge carriers. When theelectric field is sufficient to arrest further transfer of holes and electrons, thedepletion region has reached its equilibrium dimensions. The electric field acrossthe depletion region at equilibrium is called the built-in voltage. Because thisdepletion region is composed of immobile negative and positive ions, in this case,

2 1 Physics Basis of Organic Semiconductor Heterojunctions

the heterojunction is called a depletion junction. Most heterojunctions based oninorganic semiconductors belong to this case.

There is still another junction that the positive ions are accumulated on theP-side and the negative ions are on the N-side within the space charge region. Thisis generally called as an accumulation heterojunction. This direction of built-inpotential is from P-side to N-side. More importantly, the positive and negative ionsin space charge region are free mobile charges. This means that the space chargeregion is highly conductive, which can be better utilized in organic optoelectronicdevices. The accumulation junctions have well formed in organic semiconductors.

Figure 1.1 shows the case of a P/N heterojunction with depletion junction andaccumulation junction in thermal equilibrium with zero bias voltage.

Fig. 1.1 P/N junction with depletion junction (upper) and accumulation junction (down) inthermal equilibrium with zero bias voltage applied. Electron and hole concentrations are reported,respectively, with dot and dash lines. Two sides of charge space region are charge neutral. Thezone with + is positively charged. The zone with – is negatively charged. The electric field isshown on the bottom, the electrostatic force on electrons and holes and the direction in which thediffusion tends to move electrons and holes

1.1 Basic Concept of Heterojunctions 3

1.2 Theory of Heterojunctions

As early as 1951, Gubanov [3, 4] first theoretically analyzed heterojunction, but theheterojunction research was taken up only after Kroemer [5] proposed that aniso-type heterojunctions might show extremely high injection efficiency compared tohomojunctions. In 1960, Anderson [6, 7] first fabricated isotype and anisotypeheterojunctions and also presented a more detailed model for the arrangement of theenergy bands near the interface between two semiconductors. After then, variousmodels for the different types of heterojunctions were proposed by manyresearchers and were verified by experiments.

According to the physical thickness of the interface, heterojunctions are dividedinto abrupt heterojunctions and graded heterojunctions [8]. Since the abruptheterojunction models are found to be a good approximation for many hetero-junctions, here we primarily confined the discussion to the abrupt anisotypeheterojunctions without interface states.

Figure 1.2 shows the typical energy band profiles of two pieces of P- and N-typesemiconductors before and after the formation of an abrupt heterojunction (a) forisolation state and (b) for equilibrium state. Here, Eg is expressed as the energyband gap of the used semiconductors, U as the work function, and v as the electronaffinity. Ef is the Fermi level. The electron affinity v and work function U of a givensemiconductor are defined, respectively, as that energy which is required to removean electron from the bottom of the conduction band Ec and from the Fermi level (Ef)to a position just outside the semiconductor (i.e., vacuum level or a distance fromthe surface greater than the range of the image forces but small compared with thedimensions of the sample). The top of the valence band is represented by Ev. Thesubscripts 1 and 2 refer to P- and N-type semiconductors. In Fig. 1.2a, Ef2 is higherthan Ef1, and in this case, the electrons will transfer from N-type semiconductor toP-type semiconductor; as a result, a built-in potential is produced, which preventsthe electrons from continuing to transfer, up to the P- and N-type semiconductor tohave the same Ef, and the P-N heterojunction then reaches thermal equilibrium. Thedirection of built-in potential is from N-type semiconductor to P-type semicon-ductor. It is shown in Fig. 1.2b that the N-type semiconductor near the interfaceforms positive space charge region, leading the energy band to bend up, and theP-type semiconductor near the interface forms negative space charge region,resulting in the band bending down. Because the vacuum level is always contin-uous, and the v and Eg as intrinsic parameters of semiconductor are invariable, thismeans that energy band edges should be parallel to the vacuum level, resulting in a“spike” on the conduction band at the interface of N-type semiconductor and a“notch” on the conduction band at the interface of P-type semiconductor. Thesimilar results also appear in valence band edges. The discontinuity in the bandedges is referred to as band offset, which has an important effect on the hetero-junction performance. Here, the conduction band offset and the valence band offsetare expressed as DEc and DEv, respectively. Their relation with v is given by


DEc ¼ v1 � v2 ð1:1Þ

DEv ¼ Eg2 � Eg1� �� v1 � v2ð Þ ð1:2Þ

and

DEc þDEv ¼ Eg2 � Eg1 ð1:3Þ

The bending degree of vacuum level qVD is equal to the difference between workfunctions of two semiconductors, i.e.,

qVD ¼ /1 � /2 ð1:4Þ

Here, VD is the built-in potential of heterojunction. VD1 and VD2 are the built-inpotentials in P-type semiconductor and N-type semiconductor, and their relation is

Fig. 1.2 Diagrams ofequilibrium energy bandbefore (a) and after (b) theformation of an abruptanisotype P/N heterojunction

1.2 Theory of Heterojunctions 5

VD ¼ VD1 þVD2 ð1:5Þ

Because the work function U is related to the Fermi level v, so the built-inpotential VD has the relation as follows

VD ¼ Ef 2 � Ef 1

qð1:6Þ

In Fig. 1.2b, the coordinate of the interface is denoted by x0, then x0 – x1 isexpressed as negative space charge width, and x2 – x0 is expressed as positive spacecharge width; the positive and negative charge quantity should be equal, i.e.,

Q ¼ qNA1 x0 � x1ð Þ ¼ qND2 x2 � x0ð Þ ð1:7Þ

where Q is expressed as space charge per area, q is the electronic charge, NA1 is theconcentration of acceptors in a P-type semiconductor, and ND2 is the concentrationof donors in N-type semiconductor. Then, the following relation can be obtained by

x0 � x1x2 � x0

¼ ND2

NA1ð1:8Þ

The built-in potentials on either side of the interface, obtained by generalizingthe solution of Poisson’s equation under boundary conditions, are given by

VD1 ¼ qNA1 x0 � x1ð Þ2e1

2

ð1:9Þ

VD2 ¼ qND2 x2 � x0ð Þ2e2

2

ð1:10Þ

where e1 and e2 are the dielectric constant of P-type semiconductor and N-typesemiconductor, respectively.

The relation between relative built-in potentials VD1 and VD2 in each of semi-conductors is

VD1

VD2¼ e2ND2

e1NA1ð1:11Þ

From formula (1.11) and (1.5), we obtain

VD1 ¼ e2ND2

e1NA1 þ e2ND2VD ð1:12Þ

VD2 ¼ e1NA1

e1NA1 þ e2ND2VD ð1:13Þ


Then, the space charge widths on either side of the interface for an abruptanisotype P/N heterojunction are given by

x0 � x1ð Þ ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

2e1e2ND2

qNA1 e1NA1 þ e2ND2ð ÞVD

sð1:14Þ

x2 � x0ð Þ ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

2e1e2NA1

qND2 e1NA1 þ e2ND2ð ÞVD

sð1:15Þ

When an external voltage V is applied across such a junction, then the aboveformula can be rewritten by just replacingVD byVD−V andVD1 andVD2 by (VD1 –V1)and (VD2 – V2) as V = V1 + V2.

From formula (1.7)

Q ¼ qNA1ND2 x2 � x0ð Þ � x0 � x1ð Þ½ �NA1 þND2

ð1:16Þ

Then, the capacitance of the space charge region is given by

CPN ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

e1e2qNA1ND2

2 e1NA1 þ e2ND2ð ÞVD

sð1:17Þ

It can be seen from the heterojunction energy band diagrams, in the case ofheterojunctions, the properties of the interface vary greatly from material to materialand largely depend on the method of formation; apparently it is very difficult to useany model proposed by various researchers to explain nearly all the physicalphenomena of such heterojunctions. At present, several models have been pro-posed. In this section, we will discuss the mechanisms for charge carrier transport inabrupt anisotype heterojunctions in the chronological sequence.

1.2.1 Diffusion Model

This model was first developed by Anderson [6, 7]. In this model, Andersonassumes that the diffusion current will consist almost entirely of electrons or holesdue to the discontinuities in the band edges at the interface. As the case of theenergy band diagram shown in Fig. 1.2b, because the holes have to overcomelarger barrier than the electrons, the main current carriers should be electrons.Neglecting the generation and recombination current in space charge region, thenthe predicted current-voltage relation is given by


J ¼ A exp � qVD2

kT

� �� exp

qV2

kT

� �� exp � qV1

kT

� �� ð1:18Þ

where V1 and V2 are the applied voltages on P- and N-semiconductors, k is the

Boltzmann constant, T is the absolute temperature, and A ¼ qX Dn1ND2Ln1

, where X is

the transmission coefficient for electrons across the interface, Dn1 and Ln1 are thediffusion constant and diffusion length, respectively, for electrons in P-typesemiconductor.

It is clearly seen from the above expression that the first term in the bracket isimportant for forward bias and the second term for reverse bias, and the currentincreases exponentially with voltage in forward and reverse directions. This meansthat the relationship between current and voltage is symmetrical (the dotted line inFig. 1.3), which is not consistent with the experimental results in reverse becausethe reverse current is determined by the minority carrier concentration in P-regionwhen the elevation of conduction band bottom in P-region is over the barrier peakin N-region at the interface; therefore, the reverse current should be saturated underlarger reverse voltage.

For the case of the peak barrier in N-region at the interface lower than con-duction band bottom in P-region, the current-voltage relation can be described byShockley [9]

J ¼ qDn1n10Ln1

þ qDp2p20Lp2

� �exp

qVkT

� �� 1

� �ð1:19Þ

where n10 and p20 are the electron concentration and hole concentration, respec-tively, in P-type semiconductor and N-type semiconductor at equilibrium, Dp2 andLp2 are the diffusion constant and diffusion length, respectively, for holes in N-typesemiconductor, V is the external applied voltage.

Fig. 1.3 Current-voltagerelation for an abruptanisotype heterojunction withpositive peak (dotted line) andnegative peak (solid line)


For the case of the electron current is much larger than the hole current, then thecurrent-voltage relation can be expressed finally as

J ¼ qDn1ND2

Ln1exp � qVD � DEc

kT

� �exp

qVkT

� �� 1

� �ð1:20Þ

Obviously, the expressed current-voltage relation in Eq. (1.20) is asymmetric,the slid line shown in Fig. 1.3, indicating that the P/N heterojunction like thispossesses unidirectional conductivity. The reverse current is not related to the biasvoltage, called reverse saturated current.

1.2.2 Emission Model

If the electron current is realized by thermal motion way, not by diffusion way, thecurrent-voltage relation can be predicted by emission model [10]

J ¼ qXND2kT2pm

� �12

exp � qVD2

kT

� �� exp

qV2

kT

� �� exp � qV1

kT

� �� ð1:21Þ

where m is the effective mass of electrons in N-type semiconductor. The emissionmodel also shows that the current increases exponentially with voltage in forward.The reverse current at large voltage should also be saturated, indicating thatEq. (1.21) cannot be used to the case of reverse bias.

Figure 1.4 gives the theoretical current-voltage characteristics of an abruptanisotype P/N heterojunction predicted by Eq. (1.21), in which Schottky emissionresults in a reduced current above the critical voltage VT. However, this model isconfirmed experimentally by only a very small fraction of the investigated aniso-type heterojunctions due to the complexity of heterojunction cases.

Fig. 1.4 Semilogarithmiccurrent-voltage characteristicsfor an abrupt P/Nheterojunction


1.2.3 Tunneling Model

The tunneling mechanism to describe the current-voltage characteristics of anabrupt anisotype P/N heterojunction was first introduced by Rediker et al. [11]. Inthis model, it is predicted that the electron flow is mainly due to tunneling throughthe potential barrier in the P-type semiconductor. The general expression for thecurrent-voltage characteristics under forward bias can be written in the form [8]

J ¼ Js Tð Þ exp VV0

� �ð1:22Þ

where V0 is a constant and JS(T) is a weakly increasing function of temperature.Although Rediker et al. did not provide any expression or interpretation for the

temperature dependence of JS(T), their experimental observations of the voltage andtemperature dependence of forward current and of many other anisotype hetero-junctions, graphically exemplified in Fig. 1.5, indicated that forwardcurrent-voltage characteristics either over the whole range or above certain appliedvoltage can be represented by Eq. (1.22).

However, Newman first observed [12] that JS(T) is empirically found to beproportional to exp(T/T0). Using this variation for JS(T), the expression for J givenby Eq. (1.22) can be rewritten as

J ¼ Js Tð Þ exp TT0

� �exp

VV0

� �ð1:23Þ

where T0 is the characteristic temperature.It can be seen that the above expression clearly indicates that (i) voltage and

temperature appear as separable variables (i.e., d (ln J)/dV is temperature inde-pendent) and (ii) the temperature dependence is exponential in T (i.e., In JaT).

Fig. 1.5 Typicalexperimentally observedcurrent-voltage characteristicsfor an abrupt anisotype P/N heterojunction at threedifferent temperatures


1.2.4 Emission Recombination Model

This model assumes that there exists a thin layer at the interface having a stronglydisturbed lattice and a fast recombination and that the electrons and holes reach theinterface via thermal emission over their respective barriers [13]. The interfacestates within the band gap have great effect on the transport of the charge carriers.The injected electrons and holes by thermal emission process will very fastrecombine in the interface layer, implying that there can be no rectification unlessthe space charge region is wider than this layer. Figure 1.6 shows the equilibriumenergy band diagram for an abrupt P/N heterojunction, which is found to corre-spond to two metal semiconductor contacts in series having the boundary con-centration of the current carriers dependent on the applied voltage. The forwardcurrent-voltage characteristics determined by Shockley diode with higher barrierheight can be written by the simplified form as,

J ¼ B exp � qVD

kT

� �exp

qVbkT

� �� 1

� �ð1:24Þ

where B is only weak temperature dependent. The value b depends on the ratio ofthe densities of imperfects in the two semiconductors. The slope of the linear regionof the logarithmic forward characteristics lies between q/kT and q/2kT, i.e., b variesbetween 1 and 2. Figure 1.7 gives the ln J-V characteristics for an abrupt anisotypeP/N heterojunction at different temperatures. It can be seen that the slope of straightlines is decreased with the temperature increase.

1.2.5 Tunneling Recombination Model

The interface states at the heterojunction interface between two semiconductors arealso as the mediate energy levels of tunneling recombination, leading to the charge

Fig. 1.6 Schematicrepresentation of the emissionrecombination model fora P/N heterojunction


carriers to tunnel into the opposite region by the interface states, and finallyrecombine with the opposite charge carriers. This is generally called as tunnelingrecombination model [14]. For the case of one-step tunneling recombination pro-cesses, the forward current can be written as follows

j ¼ B exp �a VD � Vð Þ½ � ð1:25Þ

where B is a weak function of voltage and temperature, VD is the diffusion voltage,V is the applied voltage, and a depends on the electron effective mass in theforbidden region, the dielectric constant, the equilibrium carrier concentration, andthe exact shape of the barrier. For the approximation of the linear barrier, i.e., fieldis constant with position, then

a ¼ 43�h

e2mn2

ND2

� �1=2

ð1:26Þ

where mn2, e2 and ND2 are values in the N-type material. The agreement betweentheoretical and experimental values of a is generally poor, and Riben and Feucht[15] postulated a multistep tunneling recombination process.

The function behavior of the reverse current as a function of voltage and tem-perature can be described by a Zener tunneling model (see Fig. 1.8) [13]

J ¼ C exp � 43�h

e2mn2

ND2

� �1=2

E3=2g2 VD � Vð Þ�1=2

" #ð1:27Þ

Here again, as in the case of forward bias, a multistep process must be postulatedto obtain an agreement between theory and experiment.

Generally, there exist many current transport mechanisms in heterojunction, asshown in Fig. 1.5, the experimental curves have an evident turning point. It can beseen from the temperature properties that the curve slope is related to the

Fig. 1.7 ln J-Vcharacteristics for an abruptanisotype P/N heterojunctionat different temperatures


temperature below the turning point, corresponding to the emission model, whereasthe slope is independent of the temperature above the turning point; in this case, itbelongs to the tunneling model.

1.3 Energy Band Profiles of Heterojunctions

As we see, an energy band profile near the interface plays an important role inunderstanding the current transport mechanism of a semiconductor heterojunction.In the absence of interface states, the energy band profile of any heterojunction isdependent on electron affinities, energy band gaps, and work functions of the twosemiconductors forming the heterojunction. Among three parameters, the electronaffinity and energy band gap are the intrinsically basic properties of semiconductorsand are independent on doping, while the work function is dependent on doping,besides it is related to semiconductor. Anderson [16] proposed systematically theenergy band profiles of various abrupt heterojunctions by neglecting interfacestates. It is observed that the energy band profiles based on Anderson show a goodapproximation with the experimental results.

In this section, the discussion is mainly focused on the energy band profiles ofabrupt anisotype heterojunctions based on Anderson’s model, and thecurrent-voltage relationship is given to deeply understand them.

Fig. 1.8 Energy banddiagram for an abruptP/N heterojunction underreverse bias, indicatingtunneling mechanism


1.3.1 Profiles of Abrupt Anisotype P/N Heterojunctions

Depending on the electron affinities (v1, v2), energy band gaps (Eg1, Eg2), and workfunctions (/1, /2) of two semiconductors, the band profiles of probable P/Nheterojunctions can be classified into four cases.

1. v1\v2;/1\/2

As v2 [ v1 þEg1, the energy band profile is shown in Fig. 1.9, whereasv2\v1 þEg1, the energy band profile is given in Fig. 1.10. In this case, thecurrent-voltage relation can be written as

J ¼ A exp �DEc � qVD2

kT

� �� exp

qV2

kT

� �� exp � qV1

kT

� �� ð1:28Þ

where A ¼ qND2Dn1Ln1.

2. v1\v2\v1 þEg1;/1 [/2

The energy band profile is shown in Fig. 1.11. In this case, the current-voltagerelation is

J ¼ A exp �DEc � qVD

kT

� �� exp

qVkT

� �� 1

� �ð1:29Þ


3. v1 [ v2;/1 [/2; v1 þEg1\v2 þEg2

Fig. 1.9 Energy band profileof v1\v2;/1\/2,v2 [ v1 þEg1


For the case of qVD1 [DEc, the energy band profile is shown in Fig. 1.12. Thecurrent-voltage relation is written as

J ¼ A exp � qVD � DEc

kT

� �� exp

qVkT

� �� 1

� �ð1:30Þ


As the negative peak barrier is changed to positive peak barrier in energy banddiagram at forward bias, i.e., q VD1 � V1ð Þ\DEc, the related current-voltage rela-tion is changed as

Fig. 1.10 Energy bandprofile of v1\v2;/1\/2,v2\v1 þEg1

Fig. 1.11 Energy bandprofile ofv1\v2\v1 þEg1;/1 [/2

1.3 Energy Band Profiles of Heterojunctions 15

J ¼ A exp � qVD2

kT

� �� exp

qV2

kT

� �� exp � qV1

kT

� �� ð1:31Þ

For the case of qVD1\DEc, the energy band profile is shown in Fig. 1.13. Thecurrent-voltage relation is

J ¼ A exp � qVD2

kT

� �� exp

qV2

kT

� �� exp � qV1

kT

� �� ð1:32Þ


If q VD1 þ V1j jð Þ[DEc, the related current-voltage relation is written as

J ¼ A exp � qVD � DEc

kT

� �� exp

q Vj jkT

� �� 1

� �ð1:33Þ

4. v1 [ v2; v1\v2 þEg2\v1 þEg1

Fig. 1.13 Energy band profile of v1 [ v2;/1 [/2; v1 þEg1\v2 þEg2, qVD1\DEc

Fig. 1.12 Energy band profile of v1 [ v2;/1 [/2; v1 þEg1\v2 þEg2, qVD1 [DEc


Also there are two kinds of cases, qVD1 [DEc, its energy band profile is shownin Fig. 1.14, and the current-voltage relation is the same as Eq. 1.30. ForqVD1\DEc, the energy band profile is shown in Fig. 1.15, and its current-voltagerelation is the same as Eq. (1.32).

1.3.2 Profiles of Abrupt Anisotype N/P Heterojunctions

There are four kinds of cases for the profile of an abrupt anisotype N/P hetero-junction based on Anderson model.

1. v1 [ v2;/1 [/2

Figure 1.16 shows the energy band profile of v1 [ v2 þEg2, and Fig. 1.17shows the energy band profile of v1\v2 þEg2. The current-voltage relation iswritten as

J ¼ A0exp �DEv � qVD2

kT

� �� exp

qV2

kT

� �� exp � qV1

kT

� �� ð1:33Þ

where A0 ¼ qNA2

Dp1

Lp1.

Fig. 1.14 Energy band profile of v1 [ v2; v1\v2 þEg2\v1 þEg1, qVD1 [DEc

Fig. 1.15 Energy band profile of v1 [ v2; v1\v2 þEg2\v1 þEg1, qVD1\DEc


2. v1 [ v2;/1\/2; v1 þEg1 [ v2 þEg2

For this case, the energy band profile is shown in Fig. 1.18. The current-voltagerelation can be written as

J ¼ A0exp �DEv þ qVD

kT

� �� exp

qVkT

� �� 1

� �ð1:34Þ

where A0 ¼ qNA2

Dp1

Lp1.

3. v1 [ v2;/1\/2; v1 þEg1\v2 þEg2

As qVD1 [DEv, the energy band profile is shown in Fig. 1.19, and thecurrent-voltage relation is

J ¼ A0exp � qVD � DEv

kT

� �� exp

qVkT

� �� 1

� �ð1:35Þ

where A0 ¼ qNA2

Dp1

Lp1.

Fig. 1.16 Energy bandprofile of v1 [ v2;/1 [/2,v1 [ v2 þEg2

Fig. 1.17 Energy bandprofile of v1 [ v2;/1 [/2,v1\v2 þEg2


As qVD1\DEv, the energy band profile is shown in Fig. 1.20, and thecurrent-voltage relation is written as

J ¼ A0exp � qVD2

kT

� �� exp

qV2

kT

� �� exp � qV1

kT

� �� ð1:36Þ

Fig. 1.18 Energy band profile of v1 [ v2;/1\/2; v1 þEg1 [ v2 þEg2

Fig. 1.19 Energy band profile of v1 [ v2;/1\/2; v1 þEg1\v2 þEg2, qVD1 [DEv

Fig. 1.20 Energy band profile of v1 [ v2;/1\/2; v1 þEg1\v2 þEg2, qVD1\DEv

where A0 ¼ qNA2

Dp1

Lp1.


4. v1\v2; v1 þEg1 [ v2

Also there are two kinds of cases, qVD1 [DEv, its energy band profile is shownin Fig. 1.21, and the current-voltage relation is the same as Eq. 1.35. ForqVD1\DEv, the energy band profile is shown in Fig. 1.22, and its current-voltagerelation is the same as Eq. (1.36).

If the effect of interface states is taken into account, then the energy band profilesgiven above will get modified, depending on the net charge of the interface states,even the direction of the band bending also occurs changed [8].

1.4 Basic Properties of Organic Heterojunctions

By definition, organic heterojunction is a junction formed two organic semicon-ductors when contact. In contrast to inorganic heterojunctions, organic hetero-junctions generally do not have a significant amount of free charge that redistributes

Fig. 1.21 Energy bandprofile ofv1\v2; v1 þEg1 [ v2,qVD1 [DEv

Fig. 1.22 Energy bandprofile ofv1\v2; v1 þEg1 [ v2,qVD1\DEv


when materials are brought into contact [17]. Furthermore, their energetics arerarely influenced by the crystalline morphology at the interface since most of thesevan der Waals bonded materials do not require lattice matching to form orderedstructures [18]. Due to the intrinsically low carrier concentrations in organicsemiconductors, most organic heterojunctions involve negligible charge transferacross the junctions [19–21]. Such organic heterojunctions can be well described bythe classical Shockley–Mott model and are characterized by flat energy levels andaligned vacuum levels (VLs) across the junctions (Fig. 1.23a). In the past fewyears, it has been observed that when strong electron donors contact with strongelectron acceptors, substantial electrons can be transferred from the donors (typi-cally P-type materials) to the acceptors (typically N-type materials) (Fig. 1.23b)[20, 22]. This leads to energy level bending and accumulation of majority mobilecarriers in the organic heterojunctions in contrast to the depletion zone typicallyobserved in inorganic P-N heterojunctions. For this case, the electrons are accu-mulated on the side of N-type material and the holes are on the side of P-typematerial within the space charge region, and the direction of built-in potential isfrom P-side to N-side. This means that the space charge region is highly conductiveand the charge carriers in the space charge region are free mobile. This is generallycalled an accumulation junction, and the space charge region is called as accu-mulation zone. This kind of junction has been observed in both CuPc/F16CuPc andBP2T/F16CuPc [23, 24].

Fig. 1.23 Energy level diagrams of two types of reported organic heterojunctions. a Scenario I: aflat band structure with a common VL, e.g., NPB/Alq3 interface. b Scenario II: heterojunction withconsiderable carrier transfer leading to energy level bending and accumulation of mobile carriers atthe two sides of the junction, e.g., CuPc/F16CuPc interface. Reprinted from [25]

1.4 Basic Properties of Organic Heterojunctions 21

It can be seen that different from the mechanism of charge carrier diffusion ininorganic heterojunctions, the formation of organic heterojunctions is a chargetransfer process, depending on the Fermi level of organic semiconductors, wherethe electrons prefer to flow from a high Fermi level to a low Fermi level when twoorganic semiconductors contact. Therefore, the Fermi level is an important factor inthe formation of organic heterojunctions. This also leads to the formation of a newkind of heterojunction.

Similar to inorganic heterojunctions, organic heterojunctions can also be dividedinto anisotype heterojunction and isotype heterojunction by the conductivity type ofthe used two organic semiconductors. Moreover, on the basis of difference in Fermilevels, both isotype and anisotype organic heterojunctions can still be furtherclassified into two categories: depletion heterojunction and accumulation hetero-junction in the anisotype case and electron accumulation/depletion heterojunctionand hole accumulation/depletion heterojunction in the isotype case. Hence, theorganic heterojunctions include four cases [2]. The band alignment diagrams andcharge distributions in space charge region are summarized in Fig. 1.24.

(a) Depletion heterojunction

As the Fermi level of N-type organic semiconductor is higher than P-typeorganic semiconductor (/N < /P), the electron and hole depletion layers are formedon the corresponding semiconductor layer near the heterojunction interface, i.e., theelectrons are depleted at the P-type side and the holes are depleted at the N-typeside. In this case, the space charge region is composed of immobile negative andpositive ions, as depicted in Fig. 1.24a. This heterojunction is called as a depletionheterojunction. Most inorganic heterojunctions belong to this case. Recently, thedepletion heterojunction is also observed in organic semiconductor systems [25].

(b) Accumulation heterojunction

If the Fermi level of N-type organic semiconductor is lower than P-type organicsemiconductor (/N > /P), i.e., the electrons and holes are accumulated at the N-typeside and P-type side, respectively, thus the heterojunction with free charge carriersis thus formed on both sides of the space charge region. This heterojunction iscalled as an accumulation heterojunction, as shown in Fig. 1.24b. In this case, thecharges within the space charge region are free mobile, and as a result, the spacecharge region is highly conductive and the direction of the built-in potential is fromP-type to N-type, which is opposite to the inorganic heterojunction. Organicheterojunctions behavior typically as such junction property [20, 23, 24, 26, 27].

(c) Electron accumulation/depletion heterojunction

The heterojunction is formed by two N-type organic semiconductors havingdifferent work functions (Fermi levels) assuming /N1 < /N2. In this case, theorganic semiconductor layer with higher Fermi level /N1 near the heterojunctioninterface is electron depletion region, and other organic semiconductor with work


function /N2 is electron-accumulated region, as shown in Fig. 1.24c. The hetero-junction is called as electron accumulation/depletion heterojunction, which hasbeen observed in F16CuPc/SnCl2Pc isotype organic heterojunction system [28].

(d) Hole accumulation/depletion heterojunction

The heterojunction is formed by two P-type organic semiconductors with dif-ferent Fermi levels assuming /P1 < /P2. In this case, the holes are accumulated at

Fig. 1.24 Different types of organic semiconductor heterojunctions, a depletion heterojunction,b accumulation heterojunction, c electron accumulation/depletion heterojunction, and d holeaccumulation/depletion heterojunction


the side of organic semiconductor with high Fermi levels (low work function /P1)and are depleted at the side of the organic semiconductor with low Fermi level (highwork function /P2), as shown in Fig. 1.24d. This type organic heterojunction iscalled as hole accumulation/depletion heterojunction. VoPc/Ph3 organic hetero-junction exhibits the property of this junction [29].

The Anderson affinity rule is generally used to construct the energy band profileof an ideal semiconductor heterojunction without interface dipole. The distance dbetween Fermi level and energy band is relative to the potential barrier qVD and thespace charge region width W. In this case, DEc and DEv can be expressed as

DEc ¼ Eg1 � ðd1 þ d2Þþ q VD1 þVD2ð Þ ð1:37Þ

DEv ¼ Eg2 � ðd1 þ d2Þþ q VD1 þVD2ð Þ ð1:38Þ

So,

d1 þ d2 ¼ q VD1 þVD2ð Þþ 1=2 Eg1 � Eg2� ��1=2ðDEc þDEvÞ ð1:39Þ

For CuPc and F16CuPc organic semiconductors, the electron affinities are 3.12and 5.16 eV, and the band gaps are 1.7 and 1.5 eV, respectively. Therefore, thediscontinuities DEc and DEv in the conduction band and the valence band are 2.04and 1.84 eV, respectively. Then, the Eq. 1.39 becomes as

d1 þ d2 ¼ q VD1 þVD2ð Þ�0:34 ð1:40Þ

Assuming uniform distribution of charge carriers in space charge region, thecharge carrier density N can be written as

N ¼ 2ee0VD

qW2 ð1:41Þ

Again,

N ¼ N0 expqVD

kT

� �ð1:42Þ

where N0 is the charge carrier density at equilibrium, and

N0 ¼ Neff exp � dkT

� �ð1:43Þ

where Neff is the effective density of states.According to Eqs. (1.41), (1.42), and (1.43), then


d1 � d2 ¼ q VD1 � VD2ð Þþ kT lnNcW2

1 e2VD2

NvW22 e1VD1

� �ð1:44Þ

where Nc and Nv are the effective density of states in the conduction band of CuPcand the valence band of F16CuPc, respectively.

If assuming the maximum of energy band is at the center of Brillouin zone, thenthe effective state density is

Nc ¼ 22pm�

nkT� �3=2

h3 ; Nv ¼ 22pm�

pkTð Þ3=2h3

ð1:45Þ

m�n and m

�p are the effective mass of electrons and holes, respectively, and h is the

Plank constant.Then,

Nc=Nv ¼ m�n=m

�p

3=2ð1:46Þ

Additionally,

l ¼ qs=m� ð1:47Þ

where l is charge carrier mobility, and s is the relaxation time. Assuming sn � sp,then

Nc=Nv ¼ lp=ln� �3=2 ð1:48Þ

Therefore, Eq. (1.44) becomes

d1 � d2 ¼ q VD1 � VD2ð Þþ kT lnW2

1 e2VD2

W22 e1VD1

lpln

� �3=2 !

ð1:49Þ

For the case of CuPc/F16CuPc heterojunction, the Hall mobility of electrons andholes are 1.0 and 2.0 cm2/Vs, respectively. Due to e1 = e2 = 4, W1 = W2 = 15 nm,VD1 = 0.5 V, and VD2 = 0.37 V, then the values of d1 = 0.35 eV and d2 = 0.20 eVare obtained from Eqs. 1.40 and 1.49. Figure 1.25 shows the ideal interfacialelectronic structure of CuPc/F16CuPc heterojunction.

However, the existence of the interfacial dipole between two organics inheterojunction causes the difference of interfacial electronic structure of organicheterojunction from the ideal case. Ultraviolet photoelectron spectroscopy(UPS) can be used to directly determine the interfacial electronic structure oforganic heterojunction. Figure 1.26 shows the real diagram of energy band atCuPc/F16CuPc heterojunction interface [24]. It can be seen that an interfacial dipoleof 0.19 eV is formed, resulting in d1 = 0.58 eV and d2 = 0.14 eV.


Applying Poisson’s equation as for inorganic semiconductors, using a depletionwidth of 15 nm and the reported relative dielectric constant of CuPc and F16CuPc(about 5) [30], the charge carrier density in CuPc/F16CuPc organic heterojunctioncan be estimated according to the following equation:

L21 ¼ee0kT2q2N1

ð1:50Þ

Fig. 1.25 Ideal interfacialelectronic structure ofCuPc/F16CuPc heterojunction

Fig. 1.26 Real schematicenergy level diagram ofCuPc/F16CuPc organicheterojunction. Reprintedfrom [24]


where q is the elementary charge, e0 is the permittivity in a vacuum, e is the relativedielectric constant of organic semiconductor, k is the Boltzman constant, and T isthe temperature. Here, L1 is concluded to be 3 nm according to the accumulationlayer thickness of 15 nm because over 90% of the space charges in the accumu-lation region reside within 3L1 of the interface. Thus, the charge density N1 iscalculated as 1.15 � 1017 /cm3, which is over five orders of magnitude larger thanthat in pure organic films. The excessive carrier density is attributed to the extrinsiccarrier transfer between organic materials. For the case of CuPc/F16CuPc hetero-junction, when they come into contact, CuPc serves as a donor while F16CuPcbecomes an acceptor. The electrons would move from CuPc to F16CuPc across theinterface and holes would be created in CuPc due to the higher Fermi level of CuPcthan that of F16CuPc. Consequently, the accumulation of holes and electrons at theinterface leads to substantial band bending in both the CuPc and F16CuPc layers, asdescribed in Fig. 1.26. The excessive carrier density also results in a high con-ductivity of this organic heterojunction, which is the important property of accu-mulation organic heterojunctions.

Figure 1.27 shows the real energy band diagram for the interface ofCoPc/F16CuPc accumulated-type organic heterojunction [31]. The estimatedinterface dipole value of 0.16 eV is obtained by using photoemission technology.The formation of this accumulation organic heterojunction can well explain theenhanced conductivity of heterojunction. Clearly, when CoPc and F16CuPc come incontact, since the surface potential of CoPc (*5.16 eV) is higher than that ofF16CuPc (*5.73 eV), therefore CoPc serves as donor while F16CuPc becomes anacceptor. As a result, CoPc layer near the interface becomes hole rich and F16CuPclayer becomes electron rich, i.e., there is an accumulation of charge carrier at theinterface. Thus, the total charge density in the accumulation region has been esti-mated from the equation N = 2eVD/eW

2, whereW is the width of accumulation regionand VD is the built-in potential [32]. By taking W = 20 nm, e * 2.43 � 10−11 F/m,VD = 0.568 eV. Then, the value of N was estimated to be *2.1 � 1017 /cm3. Thisindicates that all the charge carriers injected due to formation of interface are freeowing to the high crystalline nature of heterojunction films. From the applicationpoint of view, this study provides a method where injection efficiency can beimproved by making highly conductive heterojunction at the electrode. In addition,heterojunction films can withstand very high current density of *103 A/cm2

without any noticeable degradation, which is encouraging for their applications.ZnPc/C60 organic heterojunction can also form accumulation junction well [33].

Figure 1.28 gives the energy level diagram of ZnPc/C60 organic heterojunctiondetermined by UPS. It can be seen that ZnPc/C60 organic heterojunction shows asmall interface dipole (eD), i.e., 0.06 eV, which is estimated by subtracting the bandbanding of both layers from the total the secondary electron (SE) cutoff shift. Bandbending is obtained from the HOMO level shift of each layer, and the values aredetermined to be 0.30 and 0.43 eV at the ZnPc and C60 layers, respectively. Thisband bending corresponds to the formation of a depletion layer in ZnPc and anaccumulation layer in C60 due to charge transfer.


Pentacene/C60 organic heterojunction is also a kind of typical accumulationjunction. Figure 1.29 shows the complete energy level diagram of pentacene/C60

organic heterojunction [34]. The interface dipole at the pentacene/C60 interface isobtained to be 0.11 eV. The LUMO offset is 0.89 eV and the HOMO offset is 1.29eV. It is clearly seen that the HOMO and LUMO offsets at the pentacene/C60

interface are smaller for the C60 on the pentacene layer compared to that for thepentacene on the C60 layer, indicating different amounts of charge redistribution.The dense and uniform nature of the pentacene/C60 interface is a result of C60 beingdeposited on pentacene rather than the opposite case, which enhanced chargeredistribution on the interface and induced the better energy level alignment. Itexplains the smaller HOMO and LUMO offsets, and the larger interface dipole inpentacene/C60 organic heterojunction.

However, it is found that the surface property of the used substrate has certaineffect on the interfacial electronic structure. The interfacial electronic structureshown in Fig. 1.29 is obtained by deposited pentacene/C60 on Au. Aspentacene/C60 heterojunction is deposited on MoO3 surface, the energy level dia-gram is showed in Fig. 1.30 [35]. As seen, the band bending shows slightly

Fig. 1.27 Schematic energyband diagram ofCoPc/F16CuPc organicheterojunction. Reprintedfrom [31]

Fig. 1.28 Energy banddiagram of ZnPc/C60 organicheterojunction. Reprintedfrom [33]


difference. In this case, the HOMO onset of C60 is measured to be 1.65 eV belowthe EF at the beginning of the C60 layer growth, while to be 1.85 eV after enoughcoverage of C60. On the other hand, the HOMO onset of the underneath pentacenefilm is 0.37 eV below the EF with no deposition of C60, while it is measured to be0.18 eV with enough overlayer of C60. Thus, a net interface dipole of 0.12 eV isobtained at the pentacene/C60 interface. It is experimentally demonstrated that the

Fig. 1.29 Energy leveldiagram of pentacene/C60

organic heterojunction on Au.Reprinted from [34]

Fig. 1.30 Energy leveldiagram of pentacene/C60

organic heterojunction onMoO3/ITO. Reprinted from[35]


small energy offset at the pentacene/C60 heterojunction makes it easy to transferelectrons from pentacene to C60 even under a small applied bias, facilitating theoccurrence of charge generation. The band bending observed in both pentacene toC60 is beneficial to exciton dissociation and charge transport in opposite direction.

1,4,5,8,9,11-Hexaazatriphenylenehexacarbonitrile (HAT-CN) is a N-type or-ganic material well to construct accumulation junction due to its low LUMO level.Figure 1.31 depicts the energy level diagram of a-NPD on HAT-CN/indium tinoxide (ITO) by using the results from UPS and XPS [36]. The HOMO and vacuumlevels are extracted from the HOMO onset and SECO cutoff in UPS results,respectively. The LUMO position of a-NPD is given by using the charge transportgap of 4.0 eV. As shown in Fig. 1.31, the a-NPD HOMO pinning phenomenon,which is controlled by gap states, occurs only near the interface (<1.6 nm, in regioni) with the HOMO onset position located at 0.3 eV below the Fermi level. Thesurface dipole induced by the spontaneous electron transfer is also shown in region iwith its contribution to the energy level shift (D) of 0.4 eV. The constant IP ofa-NPD films over the thickness of 1.6 nm (in region ii) with exhibiting HOMOband bending behavior is observed, indicating weaker charge transfer from the gapstates in upper a-NPD film to the underlying HAT-CN film to gradually achievethermodynamic equilibrium. Upon the bias operation in tandem OLEDs, theelectrons from HOMO region of a-NPD could be initially easily injected toLUMO-derived unoccupied gap states of HAT-CN due to a energy offset of 0.4 eVat the HAT-CN/a-NPD interface (region i) and leave holes in the upperHOMO-tailing region of a-NPD, which then transport through band bending regionof the a-NPD film. Obviously, these generated electrons and holes are consequentlytransported into the adjacent emission units to insure that the tandem devices workeffectively when using it as charge generation layer.

Fig. 1.31 Energy leveldiagram of a-NPD/HAT-CNorganic heterojunction onITO. Reprinted from [36]


Depletion junction based on organic semiconductors is well realized inm-MTDATA/BCP system. Electrons in BCP (N-type) transfer to m-MTDATA(P-type), leading to depletion of mobile majority carriers near the junction.Figure 1.32 gives the energy level diagram of the m-MTDATA/BCP interfaceshowing the charge transfer under thermal equilibrium [25]. It can be observed thatthe HOMO peaks of both BCP and m-MTDATA are gradually shifted away fromthe Ef by 0.23 eV and 0.14 eV, respectively. The direction of energy level bendingof the presented organic heterojunction suggests that the BCP (N-type material) onthe right have given electron to the m-MTDATA (P-type material) on the left. Thisimplies that both materials are losing their majority carriers at the contact. Suchmobile carrier depletion at the P/N junction is common in inorganic materials.

1.5 Brief Description of Organic HeterojunctionApplication in Organic Light-Emitting Diodes

OLEDs are attracting much attention for lighting and display [37]. An OLED is aLED in which the emissive layer is a film of organic semiconductor that emits lightin response to an electric current. It has a simple multilayer structure where anorganic semiconductor thin film is sandwiched with an anode and a cathode. Whenthe voltage is applied between the electrodes, the organic film gives luminescence.The typical device structure is illustrated schematically in Fig. 1.33. It can be seenthat a good OLED generally consists of at least five different layers with differentfunctionalities: two charge transport layers (hole transport layer (HTL) and electron

Fig. 1.32 Energy leveldiagram of m-MTDATA/BCPorganic heterojunction underthermal equilibrium.Reprinted from [25]


transport layer (ETL)), which transport the charges toward the emission layer(EML), and two charge blocking layers (hole blocking layer (HBL) and electronblocking layer (EBL)) that confine the charges in the EML.

As depicted in Fig. 1.34, the light generation in OLEDs includes four basicprocesses: (1) the injection of charge carriers (holes and electrons) from therespective electrodes into the organic layers, (2) the transport of charge carriers inthe organic layers, (3) the recombination of charge carriers on emission moleculeswith the creation of an excited state (excitons) and possible diffusion of theseexcitons, and (4) the decay of excitons accompanied by the emission of photons.Therefore, the emissive efficiency inside the OLEDs is fully decided by these fourprocesses.

Obviously, in order to obtain high efficiency OLEDs, the electrons and holesshould be effectively injected into the emissive layers. The low energy barriers atthe electrode/organic film interfaces are desirable for efficient charge injection andare generally a prerequisite for high device performance. This indicates that the lowand high work function metals, therefore, have to be employed in the cathode andanode to facilitate the injection of electrons and holes, respectively. However, thisresults in drawbacks, including the diffusion of metal ions, such as indium, from acommon ITO anode into the emissive layers of OLEDs and the accumulation ofspace charges at the interface due to the injected barriers between electrodes andorganic semiconductors, which leads to the degradation of device performance overtime [38]. Moreover, the low work function metals are very sensitive to moistureand oxygen in the air, which often form detrimental quenching sites near theinterface between the emissive layer and the cathode [39]. By using air- andchemistry-stable high work function metals, such as Au, Ag, and Cu, as the

Fig. 1.33 Typical devicestructure of a multi-layerOLED

Fig. 1.34 Scheme of the fourbasic processes involved inthe light emission of anOLED


cathode, the degradation effect caused by moisture and oxygen in the air can beavoided. Unfortunately, it has been proven that the electrons are very difficult toeffectively inject due to their higher injection barrier. Even when an interfacial layer[40] or N-doped organic layer is introduced [41], the low work function metals alsohave to be used as the cathode contact to guarantee the effective injection ofelectrons. Likewise, the anode must also be a high work function metal to realizethe effective hole injection, even when inserting an interfacial layer or P-dopedorganic layer [42]. This means that the device performance is strongly dependent onthe work function of metal electrodes in conventional OLEDs, which is difficult toresolve. More importantly, the instability caused by defects and high space electricfield due to charge accumulation at the interfaces between electrodes and organics isdetrimental to the efficiency and lifetime of OLEDs [43]. This problem is generallyalso very difficult to control and resolve in the design of conventional OLEDs dueto the limitations of the working principle.

Furthermore, it is also necessary for us how to realize highly efficient excitonutilization in order to obtain high efficiency OLEDs. Besides the use of high effi-ciency phosphorescence and thermally activated delayed fluorescence materials,which can realize the emission of 100% excitons, including singlet and tripletexcitons, the design of device structure to effectively confine excitons in emissivezone is also very important. Among, tandem structure is an efficient method toincrease excitons, thus doubling the current efficiency and quantum efficiency of thefabricated OLEDs [44]. A tandem OLED is a device that vertically stacks a numberof emissive single units via charge generation layers (CGLs) and can convert oneinjected electron into multiple photons, thus achieving more brightness and currentefficiency with lower current density. The CGL, obviously, plays a critical role inthe realization of high-performance tandem OLEDs since it functions as both aninternal anode and cathode to generate intrinsic charge carriers and to facilitateopposite electron and hole injection into the adjacent sub-OLEDs. Many studieshave been conducted to improve the performance of CGLs, most of which havebeen focused on the development of the doped-N/doped-P organic junction anddoped-N organic/transition metal oxide junction for both optimum optical charac-teristics and electrical properties because they allow for the realization of veryefficient tandem devices [45]. However, problems remain. First, the sophisticatedand high-cost doping (Li or Cs dopant) process is always required to increase theconductivity of the N-type layer in a CGL. In this case, direct contact between thislayer and the emissive layer should be avoided to prevent exciton quenching by thehighly active alkaline-metal dopants. Furthermore, it is well known that powerefficiency (PE) is one key to the commercial realization of a lighting source.However, since the driven voltage consumed by conventional tandem OLEDsscales in a linear fashion with the number of electroluminescent (EL) units, theresulting power consumption to obtain the same luminescence would be the samefor both the single-unit and tandem OLEDs, which means that the power efficiencycannot be greatly increased for such tandem devices. Though several factors,including a reduced quenching effect from electrodes, an improved chargerecombination balance, and a reduced charge-exciton quenching effect, may

1.5 Brief Description of Organic Heterojunction Application … 33

contribute to the increase in PE [46], tandem OLEDs still cannot give a satisfactoryperformance in PE enhancement because they require a high driving voltage. Inreality, a large PE improvement in tandem devices can be achieved only if the CGLhas excellent charge generation, transport, and injection/extraction capabilities,which allow for a negligible voltage drop across the CGL.

Recently, Ma et al. found [47, 48] that intrinsic organic semiconductor hetero-junctions, a bilayer structure composed of a P-type organic semiconductor and anN-type organic semiconductor, can be used as the CGLs to significantly enhance thepower efficiency of the fabricated tandem OLEDs. It is clearly shown that the noveldesign concept of organic semiconductor heterojunction-based CGLs is superior tothat of conventional CGLs. The utilization of organic semiconductor heterojunctionCGLs in tandem OLEDs enhances not only doubly the luminance and currentefficiency but also significantly the power efficiency. The effect of organic semi-conductor heterojunctions as CGL on improving the power efficiency was firstlyverified in red, green, and blue tandem OLEDs based on C60/pentacene (H2Pc,ZnPc, and CuPc) organic semiconductor heterojunctions as CGLs. The experi-mental results clearly showed that the power efficiency of all the devices wasenhanced with respect to the single-unit device, indicating the universal method ofthe design concept of organic semiconductor heterojunctions as CGLs forimproving efficiency, especially the power efficiency, of tandem OLEDs. Themechanism investigation demonstrated that organic semiconductor heterojunctionsallow for an interfacial electron redistribution to supply high-density free charges,thus efficiently decreasing the voltage drop across it, and also significantlyenhancing the exciton recombination. It can be seen that the role of organicheterojunctions as CGL is not only related to the mobility of used organic semi-conductors but also strongly dependent on energy level position between P-typeand N-type organic semiconductors [49]. It is clearly shown by experimental andtheoretical investigation that the charge generation in organic semiconductorheterojunctions is a tunneling process, which can be well explained by proposedtunneling model [50].

As we can see, CGLs in tandem OLEDs can effectively generate charges andrealize the injection of charges into respective EL units under external electricfields. Like metals, CGLs play the important role of electrodes, although they arefloated within the devices. Theoretically, CGLs can serve as electrodes to realize theinjection of both electrons and holes, but they are completely different from themetal electrodes in conventional OLEDs. In the case of CGLs, the injected chargesoriginate from the generated charges in CGLs, and the injection is directly from theCGLs into the EL units. This predicts that CGLs should be able to be used as chargeinjectors to realize charge injection instead of metal electrodes, whereas theadvantage is that the charge injection is far from the instable metal/organic inter-faces. This will greatly improve the stability of the fabricated OLEDs. Recently, theelectrode-independent charge injection by using organic semiconductor hetero-junctions as injectors in OLEDs was indeed realized, and a comparable EL per-formance with that of conventional OLEDs was obtained. More importantly, thedevice stability is greatly improved [51]. The detailed studies presented on the


transport properties of electrons and holes in single-carrier devices based on organicsemiconductor heterojunctions by current density–voltage (J–V) measurements atvarious temperatures [52] have demonstrated that Fowler–Nordheim (F–N) tunnelmodel can be used to well demonstrate the J–V properties, indicating that the chargeinjection based on organic semiconductor heterojunction is a tunnel process. Theimpedance measurements found that the lower unoccupied molecular orbit energylevel and higher electron mobility properties are very necessary for the usedelectron-transport materials to reduce the space charge region width, thus realizinghighly efficient electron injection.

Clearly, organic semiconductor heterojunctions play an important role in con-trolling the operation and performance of multilayered organic optoelectronicsdevices, such as OLEDs, organic photovoltaic devices (OPVs), and organicfield-effect transistor (OTFTs). And the organic semiconductor heterojunctions asCGLs and charge injectors in OLEDs enhance greatly the performance of thefabricated OLEDs, especially the improvement in power efficiency and stability.Moreover, the realization of organic semiconductor heterojunctions as chargeinjectors in OLEDs also breaks through the working principle of traditional OLEDs.It has been experimentally demonstrated that electronic structures and frontierorbital energy offsets at the junction interface essentially determine the function-alities of the organic semiconductor heterojunctions, such as chargeinjection/transport properties, and charge separation processes. Insight into energylevel alignments of organic semiconductor heterojunctions is of fundamentalimportance for continuous development of these devices.

References

1. C. Kittel, Introduction to Solid State Physics, 7th edn. (Wiley, 2005)2. D.H. Yan, H.B. Wang, B.X. Du, Introduction to Organic Semiconductor Heterojunctions

(Wiley Science Press, 2008)3. A.I. Gubanov, Zh. Tekh. Fiz. 21, 304 (1951)4. A.I. Gubanov, Zh. Eksper. Teor. Fiz. 21, 79 (1951)5. H. Kroemer, Proc. IRE 45, 1535 (1957)6. R.L. Anderson, IBM J. Res. Dev. 4, 283 (1960)7. R.L. Anderson, Solid Stae Electron. 5, 341 (1962)8. B.L. Sharma, R.K. Purohit, Semiconductor Heterojunction (Pergamon Press, 1974)9. W. Shockley, The theory of pn junctions in semiconductors and pn junction transistors. Bell

Syst. Tech. J. 28, 435 (1949)10. S.S. Perlman, D.L. Feucht, Solid State Electron. 7, 911 (1964)11. R.H. Rediker, S. Stopek, J.H.R. Ward, Solid State Electron. 7, 621 (1964)12. P.C. Newman, Electron. Lett. 1, 265 (1965)13. U. Dolega, Z. Naturf. 18a, 653 (1963)14. A.R. Riben, D.L. Feucht, Solid State Electron. 9, 1055 (1966)15. A.R. Riben, D.L. Feucht, Int. J. Electron. 20, 583 (1966)16. R.L. Anderson, Solid State Electron. 5, 341 (1962)17. I.G. Hill, D. Milliron, J. Schwartz, A. Kahn, Appl. Surf. Sci. 166, 354 (2000)18. S.R. Forrest, Chem. Rev. 97, 1793 (1997)19. H. Va´zquez, W. Gao, F. Flores, A. Kahn, Phys. Rev. B 71, 041306(R) (2005)

1.5 Brief Description of Organic Heterojunction Application … 35

20. Y.L. Gao, H.J. Ding, H.B. Wang, D.H. Yan, Appl. Phys. Lett. 91, 142112 (2007)21. J.X. Tang, C.S. Lee, S.T. Lee, J. Appl. Phys. 101, 064504 (2007))22. J.X. Tang, K.M. Lau, C.S. Lee, S.T. Lee, Appl. Phys. Lett. 88, 232103 (2006)23. L. Yan, N.J. Watkins, S. Zorba, Y. Gao, C.W. Tang, Appl. Phys. Lett. 81, 2752 (2002)24. K.M. Lau, J.X. Tang, H.Y. Sun, C.S. Lee, S.T. Lee, Appl. Phys. Lett. 88, 173513 (2006)25. T.W. Ng, M.F. Lo, S.T. Lee, C.S. Lee, Appl. Phys. Lett. 100, 113301 (2012)26. S. Hiller, D. Schlettwein, N.R. Armstrongc, D. Wöhrle, J. Mater. Chem. 8, 945 (1998)27. H.B. Wang, X.J. Wang, H.C. Huang, D.H. Yan, Appl. Phys. Lett. 93, 103307 (2008)28. H.B. Wang, Z.T. Liu, T.W. Ng, M.F. Lo, C.S. Lee, D.H. Yan, S.T. Lee, Appl. Phys. Lett. 96,

173303 (2010)29. H.B. Wang, X.J. Wang, D.H. Yan, Appl. Phys. Lett. 93, 113303 (2008)30. M. Gorgoi, D.R.T. Zahn, Org. Electron. 6, 168 (2005)31. A.K. Debnath, A. Kumar, S. Samanta, R. Prasad, A. Singh, A.K. Chauhan, P. Veerender, S.

Singh, S. Basu, D.K. Aswal, S.K. Gupta, Appl. Phys. Lett. 100, 142104 (2012)32. B.G. Streetman, S.K. Banerjee, Soild State Electronic Devices (PHI Learning Pvt. Ltd., New

Delhi, 2010)33. S.H. Park, J.G. Jeong, H.-J. Kim, S.-H. Park, M.-H. Cho, S.W. Cho, Y. Yi, M.Y. Heo, H.

Sohn, Appl. Phys. Lett. 96, 013302 (2010)34. S.J. Kang, Y. Yi, C.Y. Kim, K. Cho, J.H. Seo, M. Noh, K. Jeong, K.H. Yoo, C.N. Whang,

Appl. Phys. Lett. 87, 233502 (2005)35. X.L. Liu, C.G. Wang, C.C. Wang, I. Irfan, Y.L. Gao, Org. Electron. 17, 325 (2015)36. J.P. Yang, F. Bussolotti, Y.Q. Li, X.H. Zeng, S. Kera, J.X. Tang, N. Ueno, Org. Electron. 24,

120 (2015)37. Q. Wang, D.G. Ma, Chem. Soc. Rev. 39, 2387 (2010)38. W. Brutting, C. Adachi, Physics of Organic Semiconductors, 2nd edn. (Wiley-VCH Verlag

GmbH & Co. KGaA, Weinheim, 2012)39. I.D. Parker, Y. Cao, C.Y. Yang, J. Appl. Phys. 85, 2441 (1999)40. H. Ma, H.L. Yip, F. Huang, A.K.F. Jen, Adv. Funct. Mater. 20, 1371 (2010)41. K. Walzer, B. Maennig, M. Pfeiffer, K. Leo, Chem. Rev. 107, 1233 (2007)42. X.F. Qiao, J.S. Chen, X.L. Li, D.G. Ma, J. Appl. Phys. 107, 104505 (2010)43. F. So, D. Kondakov, Adv. Mater. 22, 3762 (2010)44. H.D. Sun, Y.H. Chen, J.S. Chen, D.G. Ma, IEEE. J. Sel. Top. Quantum Electron. 22, 7800110

(2016)45. M.K. Fung, Y.Q. Li, L.S. Liao, Adv. Mater. 28, 10381 (2016)46. L.S. Liao, K.P. Klubek, Appl. Phys. Lett. 92, 223311 (2008)47. Y.H. Chen, D.G. Ma, J. Mater. Chem. 22, 18718 (2012)48. Y.H. Chen, Q. Wang, J.S. Chen, D.G. Ma, D.H. Yan, L.X. Wang, Org. Electron. 13, 1121

(2012)49. Y.H. Chen, H.K. Tian, Y.H. Geng, J.S. Chen, D.G. Ma, D.H. Yan, L.X. Wang, J. Mater.

Chem. 21, 15332 (2012)50. Q.X. Guo, H.D. Sun, J.X. Wang, D.Z. Yang, J.S. Chen, D.G. Ma, J. Mater. Chem. C 4, 376

(2016)51. Y.H. Chen, D.G. Ma, H.D. Sun, J.S. Chen, Q.X. Guo, Q. Wang, Y.B. Zhao, Light Sci Appl.

5, e16042 (2016)52. Q.X. Guo, H.D. Sun, D.Z. Yang, X.F. Qiao, J.S. Chen, T. Ahamad, S.M. Alshehri, D.G. Ma,

Adv. Mater. Interfaces 3, 1600081 (2016)


Chapter 2Electrical Properties of OrganicSemiconductor Heterojunctions

2.1 Current–Voltage Characteristics

The current–voltage (I-V) characteristics can be able to explain the electrical con-duction mechanism of heterojunctions, also be generally used to determine thebuilt-in junction potential and energy discontinuities in band edges at the interfaceof heterojunctions. The various heterojunction models for the I-V characteristics ofheterojunctions based on inorganic semiconductors have been developed [1]. TheI-V characteristics of organic semiconductor heterojunctions are often similar tothose of inorganic heterojunctions. As a consequence, the theoretical models basedon inorganic heterojunctions have been extended to organic heterojunctions [2].However, the models derived for inorganic heterojunctions are based on energyband structure where thermal and optical excitation results in delocalized freecharge carriers. In contrast, organic semiconductors are generally characterized byhopping transport and tightly bound, localized exciton states that require significantenergy to dissociate into free charge carriers. Therefore, these descriptions do notrigorously apply to the physics of organic heterojunctions.

Forrest et al. derived the ideal diode equation specifically for the case of organicheterojunctions [3]. Explicitly treating polaron pair generation, recombination, anddissociation at the heterojunctions, they developed a current density–voltage(J-V) characteristic similar in form to the Shockley but differing in several keyaspects. In this polaron pair (PP) model, it is assumed that the recombination andgeneration (if excitons are considered) govern the currents in the system. As shownin Fig. 2.1a, the average polaron pair separation, a0, thus defines the “volume” ofthe heterojunction region, the current outside of this region is unipolar with purehole and electron currents flowing in the P and N bulk, respectively. Figure 2.1bshows the processes that occur within the organic heterojunction volume. Therecombination of polaron pairs is described via


37

Fig. 2.1 a Energy level diagram showing the anode and cathode work functions, WFa and WFc,and their associated injection barriers /a and /c, respectively. The interfacial gap, DEHL, is theenergy difference between the highest occupied molecular orbital energy of the donor and thelowest unoccupied molecular orbital energy of the acceptor. Current is unipolar in the donor (Jp)and acceptor (Jn) layers and is determined from generation/recombination in the heterojunctionregion, roughly defined by the spatial extent, a0, of the polaron pair distribution at the interface.b Processes occurring within the heterojunction region. Excitons diffuse, with current density, JX,to the heterojunction and undergo charge transfer to form polaron pairs. These may recombine, atrate kPPr, or dissociate with rate, kPPd, as determined by the Onsager–Braun model (C. L. Braun,J. Chem. Phys., 80, 4157 (1984)). The current density, J, contributes to the interfacial free electron(nI) and hole (pI) densities, which bimolecularly recombine to form polaron pairs at rate krec.Reprinted from [3]

JXa0

� kPPr n� neq� �� kPPdnþ krecnIpI ¼ 0 ð2:1Þ

and for free carriers:

kPPdn� krecnIpI þ JXqa0

¼ 0 ð2:2Þ

where steady-state conditions are assumed. Here, n is the PP density, JX is theexciton current density diffusing to the interface, J is the charge current densityflowing through the device, q is the electron charge, and nI and pI are the interfacialfree electron and hole densities, respectively.

38 2 Electrical Properties of Organic Semiconductor Heterojunctions

Solving Eq. (2.1) for the PP density and substituting the result into Eq. (2.2)give

J ¼ qa0kreckPPr

kPPd þ kPPr

� �nIpI � kPPd

kPPd;eqnI;eqpI;eq

� �� qJX

kPPdkPPd þ kPPr

� �ð2:3Þ

where neq = krecnI,eqpI,eq/kPPd,eq is used from Eq. (2.2). The subscript eq indicatesthe thermal equilibrium value in the absence of bias or illumination. Similar to theShockley equation, quasi-equilibrium is assumed. Hence, the carrier densities at theinterface (nI, pI) and contacts (nC, pC) are related via

nI ¼ nC expdAq Va � Vbið Þ

kbT

� �ð2:4aÞ

and

pI ¼ pC expdDq Va � Vbið Þ

kbT

� �ð2:4bÞ

where dD + dA = 1 are the fractions of the potential dropped across the donor(D) and acceptor (A) layers, respectively. Here, Va is the applied bias, kb isBoltzmann’s constant, and T is the temperature. These relations are strictly validonly when J = 0, but are a good approximation at low current when J is muchsmaller than either of its drift or diffusion components.

Use of Eq. (2.4a, b) in Eq. (2.3) yields

J ¼ qa0krecnCpC 1� gPPdð Þ exp � qVbi

kbT

� �

� expqVa

kbT

� �� kPPdkPPd;eq

� � qgPPdJX

ð2:5Þ

where ηPPd = kPPd/(kPPd + kPPr) is the PP dissociation probability. Assumingdetailed balance of the charge density adjacent to an injecting contact, then

nC ¼ f Fc; Tð ÞNLUMO exp �/c=kbTð Þ ð2:6Þ

where NLUMO is the density of states (DOS) at the acceptor LUMO, and Fc is theelectric field at the cathode contact. The analogous relation involving the injectionbarrier, /a, (see Fig. 2.1a) exists for holes at the anode with NHOMO as the DOS atthe donor HOMO. The term f (Fc, T) is dominated by Schottky barrier lowering,which can be neglected since it is near unity except for the case of high field and/orlow temperature here. Thus, the ideal diode equation for an organic heterojunctionin the absence of traps can be obtained

2.1 Current-Voltage Characteristics 39

J ¼ qa0krecnCpCNHOMONLUMO 1� gPPdð Þ exp �EHL

kbT

� �

� expqVa

kbT


� � qgPPdJX

¼ Js0 expqVa

kbT


� � qgPPdJX

ð2:7Þ

where DEHL = /a + /c + qVbi from Fig. 2.1a.Under forward bias, kPPd is similar to or less than kPPd,eq, and the current density

increases exponentially with an ideality factor n = 1. In this case, Eq. (2.7) reducesto the familiar

J ¼ Js0 expqVa

kbT

� �� 1

� � qgPPdJX ð2:8Þ

which is frequently used to model organic heterojunction solar cells. As expected,the photocurrent is directly proportional to the PP dissociation efficiency, whichdiminishes with increasing forward bias.

It is well-known that the significant molecular disorder of organic semicon-ductors leads to a broad density of states with an approximately Gaussian energeticdistribution at the HOMO and LUMO levels [4]. Evidence also suggests thatground-state interactions between the donor and acceptor further broaden thisdistribution near the heterojunction interface [5]. The low-energy tail of the DOScan be modeled as an exponential distribution of traps [6]. This changes the kineticsof recombination at the interface and introduces an ideality factor n > 1 in the diodeequation.

Assuming an exponential trap distribution with characteristic trap temperature,Tt,A, in the acceptor, the trapped (nt) and free (n) electron densities are relatedapproximately via

nt � HA expEFn � ELUMO

kbTt;A

� �� HA

nNLUMO

� �1=lA

ð2:9Þ

where HA is the density of trap states at the acceptor LUMO, EFn is its electronquasi-Fermi energy, ELUMO is the LUMO energy of the acceptor, and lA = Tt,A/T. Asimilar relationship holds for the trapped hole density, pt, in the donor. Assumingthat the trapped carrier density significantly exceeds the free carrier density,Eq. (2.3) becomes

J ¼ qa0kPPr

kPPd þ kPPr

� �krec;n nIpI � kPPd

kPPd;eqnI;eqpI;eq

� ��

þ krec;p pInIt � kPPdkPPd;eq

pI;eqpIt;eq

� �� qJX

kPPdkPPd þ kPPr

� � ð2:10Þ


where recombination now occurs primarily at trap states, and krec,n and krec,p are therate constants describing recombination at the heterojunction between a free elec-tron in the acceptor (nI) with a trapped hole in the donor (pIt), and vice versa.

Using Eqs. (2.4a, b), (2.9), and (2.10) gives

J ¼ qa0 1� gPPdð Þ krec;nNLUMO

HD exp �aD=kbTð Þ

� exp qVa=nDkbTð Þ � kPPdkPPd;eq

� �þ krec;nNLUMOHA

� exp �aD=kbTð Þ exp qVa=nAkbTð Þ � kPPdkPPd;eq

� �� qgPPdJX

ð2:11aÞ

where

aD ¼ DEHL

nDþ lD � 1

lDdD/c � dA/að Þ ð2:11bÞ

and

aA ¼ DEHL

nAþ lA � 1

lAdA/a � dD/cð Þ ð2:11cÞ

The ideality factors nA and nD are given by

nA ¼ lAdD lA � 1ð Þþ 1

ð2:12aÞ

and

nD ¼ lDdA lD � 1ð Þþ 1

ð2:12bÞ

More compactly, Eq. (2.11a–c) becomes the ideal diode equation in the presenceof traps

J ¼ JsD exp qVa=nDkbTð Þ � kPPdkPPd;eq

� �

þ JsA exp qVa=nAkbTð Þ � kPPdkPPd;eq

� �� qgPPdJX

ð2:13aÞ

which simplifies to

J ¼ JsD exp qVa=nDkbTð Þ � 1½ � þ JsA exp qVa=nAkbTð Þ � 1½ � � qgPPdJX ð2:13bÞ


when kPPd �kPPd,eq under forward bias (cf. Eq. (2.8)). Here JsD and JsA are the darksaturation currents given by the prefactors in Eq. (2.11a–c).

The ideal diode equation (2.13a, b) is modified to include the effect of seriesresistance, and Rs can be written as

J ¼ JsD exp q Va � JRsð Þ=nDkbTð Þ � 1½ �þ JsA exp q Va � JRsð Þ=nAkbTð Þ � 1½ � � qgPPdJX

ð2:14Þ

Figure 2.2a shows a set of dark J-V characteristics calculated using Eq. (2.11a–c)over the temperature range from 120 to 296 K. The used parameters are listed inTable 2.1 for organic semiconductors. The series resistance of Rs = 1 X cm2 (inwhich case, Va is replaced by Va − JRs) and assumed that most of the potential isdropped across the donor layer, resulting in different ideality factors according toEq. (2.12a–c). Figure 2.2b gives the ideality factors and their associated dark sat-uration currents. It can be seen that both ideality factors increase with temperaturedecrease, and the quantity n ln(Js), where the argument in the logarithm is implicitlynormalized to 1 A/cm2, is nonlinear when plotted versus 1/kbT. This contrasts withprevious analyses based on the Shockley equation [2], which predict a linear rela-tionship for this quantity with a slope equal to −DEHL/2.

Figure 2.3a, b shows the forward bias J-V characteristics of both ITO/CuPc/C60/BPhen/Al and ITO/SubPc/C60/BPhen/Al devices taken over a range of temperaturesfrom 114 to 296 K. The fitting curves by Eq. (2.14) are also given. It can be seen

Fig. 2.2 a Calculated dark J-V characteristics over the temperature range from 120 to 300 K in20 K increments using Eq. (2.11a–c). The slope in reverse bias is due to the field-dependentdissociation of thermally generated polaron pairs. In forward bias, three regions are apparent. AtVa < 0.3 V, trap-limited recombination involving free acceptor electrons and trapped donor holesdominates, and the slope is proportional to the donor ideality factor, nD. At higher bias, the inverseprocess dominates (i.e., free donor holes recombine with trapped acceptor electrons) and the slopeis proportional to the acceptor ideality factor, nA. Series resistance (Rs) limits the current atVa > 0.8 V. b Diode parameters n and n ln(Js) corresponding to the dark currents in (a). Bothideality factors increase with decreasing temperature, though nA does so only slightly. Reprintedfrom [3]


that Eq. (2.14) fits the experimental data over the entire temperature range well,demonstrating the validity of Eq. (2.14) to accurately describe the J-V character-istics of organic heterojunction-based devices.

As we know, a notable feature of inorganic heterojunctions is the rectifyingcharacteristics, i.e., the current is very small at reverse voltage, until reversebreakdown. This is the basic property of a depletion-type heterojunction. As shownin Figs. 2.2a and 2.3, organic heterojunctions with depletion space charge regionalso show the same rectifying characteristics and can be explained well by themodified Shockley Eq. (2.14). However, the heterojunctions formed by two organicsemiconductors become accumulation-type junction, and the J-V characteristicscannot be modeled well by thermionic emission processes. For the case of accu-mulation junctions, the electrons accumulate at the N-type semiconductor, and theholes accumulate at the P-type semiconductor; therefore, a large number of free

Table 2.1 Model parameter values for calculation here. Reprinted from [3]

Parameter Value

Donor thickness = acceptor thickness 40 nm

ΔEHL 1.2 eV

vbi 0.5 V

Tt,a = Tt,D 1000 K

HA= HD 1018/cm3

NHOMO = NLUMO 1021/cm3

dA 0.1

a0 1.5 nm

kppr 1 /ls

krec,n = krec,p = ql/e e/e0 = 3, l = 10−3 cm2/V s

Rs 1 X cm2

Fig. 2.3 Dark current density versus forward voltage for a CuPc/C60 and b SubPc/C60 devicesrecorded for T = 296, 275, 247, 218, 193, 171, 155, 145, 128, and 114 K. Bold red lines indicatefits to Eq. (2.14) in the text. Thin black lines connect the data points. Reprinted from [3]


charge carriers exist in the space charge region [7], and besides the large currentproperty at forward bias, it is experimentally shown [8] that the reverse bias alsoproduces large current. Figure 2.4 shows the J-V characteristics of ITO/MoO3

(3 nm)/20 wt% MoO3:TAPC (50 nm)/m-MTDATA (15 nm)/HAT-CN (15 nm)/3wt% Cs2CO3:BPhen (50 nm)/Cs2CO3 (1 nm)/Al heterojunction device 1 andITO/MoO3 (3 nm)/20 wt% MoO3:TAPC (50 nm)/m-MTDATA (15 nm)/3 wt%Cs2CO3:BPhen (50 nm)/Cs2CO3 (1 nm)/Al non-heterojunction device 2. It can beseen that m-MTDATA/HAT-CN heterojunction device 1 produces a highly sym-metric current property at forward and reverse bias, and the current is much largerthan that of non-heterojunction device 2. As shown, the J-V characteristic ofnon-heterojunction device 2 can be well described by Shockley equation, where theslope of the linear region corresponds to the ideality factor n = 2, typical Shockleydiode property. However, heterojunction device 1 does show a clear linear straightregion in J-V characteristic, indicating that the electrical property cannot beexplained only by using thermionic emission. To elucidate this, the reverse bias J-Vcharacteristics of heterojunction device 1 under different temperatures were tested.As shown in Fig. 2.5, a relation of ln(J) � T is obtained well, a typical tunnelingprocess [9]. This also indicates that the charge generation ofm-MTDATA/HAT-CN heterojunction is a tunneling process.

The tunneling equation for heterojunctions under reverse bias was proposed byRiben et al. as follows [10]

Jr ¼ G0Va exp �U Vd þVað Þ�1=2h i

ð2:15Þ

where Jr is the reverse current density, G0 is a constant determined by the nature ofthe material, U is a linear variable of temperature, Vd is the built-in potential, and Va

is the applied reverse voltage (written as a positive value). Initially, the Zenertunneling equation well described the electron tunneling from the valence band ofthe P-type semiconductor to the conduct band of the N-type semiconductor instaggered gap-type heterojunctions. Figure 2.6 gives the double logarithm J-V plotof m-MTDATA/HAT-CN heterojunction device 1 under various temperatures. Thecurves are well fitted by Eq. (2.15) at high voltage region, further demonstrating

Fig. 2.4 J-V characteristicsof heterojunction device 1 andnon-heterojunction device 2.The fitting curves by tunnel(red solid line) and thermionicemission (black solid line)models are given. Reprintedfrom [8]


that the current process of m-MTDATA/HAT-CN organic heterojunction is Zenertunneling.

It can be seen that the current at high voltage region agrees with the tunnelingmodel, but deviates at low bias region. The deviation has been attributed thecontribution of thermionic emission, and the fact that the Fermi–Dirac distributionis not a step function at temperatures higher than absolute zero [11, 12]. Therefore,a modified Fowler–Nordheim (F-N) tunneling model is described by [12]

ln I=F2� � ¼ �P1

Fþ ln

P2

F

� �� ln sin

P3

F

� �� ð2:16Þ

where P1, P2, and P3 are constant parameters in a constant temperature measure-ment. If the thickness cannot be determined, then the Eq. (2.16) can be written as

ln I=V2� � ¼ �~P1

Vþ ln

~P2

V

� �� ln sin

~P3

V

� �� ð2:17Þ

Equation (2.16) has well suited for fitting the current characteristics ofpentacene/C70organic heterojunction-based devices at different temperatures [13].

Fig. 2.5 Temperature–current density dependence ofheterojunction device 1 underfixed voltages of 0.5 and 1 V.Reprinted from [8]

Fig. 2.6 J-V characteristicsof m-MTDATA/HAT-CNheterojunction device 1 atdifferent temperatures. Thefitting curves by Eq. (2.15)are given by solid lines. Thereverse voltages are absolutevalues. Reprinted from [8]


As shown in Fig. 2.7, the modified F-N tunneling model of Eq. (2.16) fits quitewell with the experimental data at all temperatures, indicating that the chargegeneration processes in pentacene/C70 organic heterojunction are also tunneling.

2.2 Capacitance–Voltage Characteristics

The capacitance–voltage (C-V) characteristics are also an important means tocharacterize the electrical properties of a heterojunction. The measurement of thejunction capacitance (C = dQ/dV) as a function of reverse bias is often used as apowerful experimental method for the analysis of the space charge region potentialand the charge distribution in a heterojunction.

When two semiconductors with opposite conductivity contact, there occurscharge transfer between two semiconductors until the Fermi levels are equalized.This causes the formation of a space charge layer on both sides of the interface. Ifno considering interface states, the expression of junction capacitance per unit areaof an abrupt anisotype heterojunction can be written as [14]

C ¼ qenepNnPp

2 enNn þ epPp� �

" #1=2

VD � Vð Þ�1=2 ð2:18Þ

where Nn and Np are the donor and acceptor concentrations in, and en and ep are thedielectric constants of, N- and P-type semiconductors, respectively, VD is thebuilt-in junction potential, V is the applied voltage and q is the electronic charges. Itis clear that a plot of C−2 against applied reverse voltage V is linear and itsextrapolated intercept on the voltage axis gives the built-in junction potential VD.

If considering interface states and electric dipole, then the junction capacitanceper unit area of an abrupt anisotype heterojunction can be expressed as [15]

Fig. 2.7 Tunneling currentcharacteristics of ITO/MoO3

(3 nm)/TAPC: MoO3 (30%,50 nm)/pentacene (30 nm)/C70 (30 nm)/Li2CO3 (1 nm)/Bphen: Li2CO3 (3%, 45 nm)/Li2CO3 (1 nm)/Al at differenttemperatures. The black solidlines are the fitting curves byEq. 2.16. Reprinted from [13]


C ¼ B1 1þ f ðVÞ qQIs

dVþ q/m

dV

� �VD � V � /m � B2Q

2Is

� ��1=2 ð2:19aÞ

where

f ðVÞ ¼ 2B2QIs þ 2 B2enNn=epPp� �1=2

VD � V � /m � B2Q2Is

� ��1=2 ð2:19bÞ

B1 ¼ qenepNnPp


" #1=2

ð2:19cÞ

B2 ¼ 2q enNn þ epPp� �� 1 ð2:19dÞ

and QIS and /m are the net charge on the interface states and the electric dipole,respectively.

If the net charge on the interface states is independent on the applied voltage andthe electric dipole is absent, then the expression (2.19) reduces to

C ¼ qenepNnPp


" #1=2

VD � V � B2Q2Is

� ��1=2 ð2:20Þ

It can be seen that this expression also has a linear relation of C-2 versusV similar to Eq. (2.18), but the extrapolated intercept of this plot on the voltage axisgives rise to an apparent built-in junction voltage (VD-B2Q

2IS) instead of VD.

Ma et al. studied the capacitance characteristics of pentacene/C70organicheterojunction [13]. The current–capacitance–voltage characteristics ofpentacene/C70 organic heterojunction-based device are shown in Fig. 2.8. It can beseen that in the forward direction, the capacitance begins to drop drastically at about0.5 V because of the significant injection and recombination of carriers, which justcorresponds to the exponential increase of the current. Similarly, the breakdown ofcurrent and the severe drop of capacitance both happen at the reverse voltage of

Fig. 2.8 Current–capacitance–voltagecharacteristics of ITO/MoO3

(3 nm)/TAPC: MoO3 (30%,50 nm)/pentacene (30 nm)/C70 (30 nm)/Li2CO3 (1 nm)/Bphen: Li2CO3 (3%, 45 nm)/Li2CO3 (1 nm)/Al device at287 K. Reprinted from [13]

2.2 Capacitance-Voltage Characteristics 47

about 1.5 V, which should be caused by the large amount of generated charges inthe pentacene/C70 heterojunction. It should be noticed that the large reverse currentshould mainly come from the charge generation instead of the injection from theelectrode under reverse voltage, because the doped hole and electron transport layercould effectively block the injection of electrons and holes, respectively. Figure 2.9gives 1/C2 against V relation of pentacene/C70organic heterojunction-based device.A linear relation of the inverse capacitance square versus the reverse voltage isclearly seen, which corresponds well with the Mott–Schottky relation [16]. Similarto Eq. (2.18), Mott–Schottky relation is generally written as

1C2 ¼

2 Vb � Vð ÞeNAere0A2 ð2:21Þ

where Vb is the built-in voltage, N is the density of free charge carriers, e0 is thepermittivity of free space, er is the relative dielectric constant, and A is the activearea. Therefore, the density of free charge carriers N in space charge region could becalculated from the slope, in the range of 1019/cm3, which is quite high for organicheterojunctions [17]. The high free carrier density guarantees the large tunnelingcurrent density and thus the possibility of efficient charge generation inpentacene/C70organic heterojunction.

It has been proven experimentally that organic bulk heterojunctions possess theproperty of high conductivity [18], not only as a highly efficient charge generationlayer (CGL) to fabricate high-efficiency tandem OLEDs, but also as excellenthole-transporting layer to realize the large injection of holes in OLEDs [19]. It canbe seen that the high conductivity of organic bulk heterojunctions may be wellstudied by C-V measurement. Figure 2.10 shows 1/C2 against V relation ofHAT-CN/HAT-CN:TAPC/TAPC organic heterojunction-based device. Accordingto the classical Mott–Schottky theory, the relation of C versus V forHAT-CN/HAT-CN:TAPC/TAPC structure device can be written as

Fig. 2.9 1/C2 againstV relation of ITO/MoO3

(3 nm)/TAPC: MoO3 (30%,50 nm)/pentacene (30 nm)/C70 (30 nm)/Li2CO3 (1 nm)/Bphen: Li2CO3 (3%, 45 nm)/Li2CO3 (1 nm)/Al device at1 kHz. Reprinted from [13]


1C2 ¼

2 Vb � Vð ÞeNAere0A2 þ d2i

ere0Að Þ2 ð2:22Þ

where e is the charge of one electron, di is the thickness of intrinsic layer, NA is thedensity of ionized free charges. Thus, NA is estimated to be in the range of 1019/cm3

by Eq. (2.22), which is several orders higher than a typical intrinsic organicsemiconductor (usually less than 1015/cm 3) [20]. This high free charge density doesconfirm that HAT-CN/HAT-CN:TAPC/TAPC organic heterojunction efficientlygenerates the more free charges, thus making it highly conductive and preferable forhole transporting.

The C-V measurement can also be used to well determine the electronic struc-tures and energy level alignment at the interface of organic heterojunctions by theextracted data of space charge width, built-in potentials, and vacuum-level shiftinformation [21]. Kim et al. studied the interface energy level alignment of a dopedorganic heterojunction using this method [22]. Figure 2.11 gives the C-V charac-teristic of ITO/15 mol% Cs2CO3-doped BPhen (20 nm)/HAT-CN (20 nm)/Aldevice. The large capacitance implies that the depletion width in the junction is very

Fig. 2.10 Mott–Schottkyplot of ITO/HAT-CN(10 nm)/HAT-CN:TAPC (30wt%) (100 nm)/TAPC(10 nm)/Al device at 5 kHz.Reprinted from [19]

Fig. 2.11 C-V characteristicof ITO/Cs2CO3-dopedBPhen/HAT-CN/Al device(measured at 1 kHz, and anapplied AC bias voltage of10 mV). Reprinted from [22]

2.2 Capacitance-Voltage Characteristics 49

narrow. The device may be simply assumed as a planar capacitor. Therefore, thedepletion width (W) is written as

W ¼ ere0AC0

ð2:23Þ

where er is the relative dielectric constant of the Cs2CO3-doped BPhen layer (3.5 fororganic), er is the permittivity of free space, C0 is the capacitance at 0 V, and A isthe area of the device. Thus, the depletion width W of 6.6 nm in Cs2CO3-dopedBPhen layer is obtained by correlating the capacitance at 0 V.

The built-in potential in Cs2CO3-doped BPhen side at 0 V is estimated using theMott–Schottky equation

w2 ¼qn22er2e0

W22 ð2:24Þ

where w2 is the built-in potential, q is the electric charge, n2 is the carrier density.Here n2 is 6.36 � 1018/cm3 [23]. Then, the built-in potential w2 is calculated to be0.71 eV. The built-in potential in HAT-CN side at 0 V can be given by [24]

w1\2ktq

er2n2er1n1

w2

� �1=2ð2:25Þ

where k is the Boltzmann constant and T is the temperature. Here, the free carrierdensity n1 of HAT-CN is 6.39 � 1019/cm3, then the upper limit of the built-inpotential w1 on HAT-CN is then estimated from Eq.2.25 as 0.06 V. Finally, thevacuum-level shift (D) at the HAT-CN/Cs2CO3-doped BPhen junction can then beestimated from the LUMO levels of HAT-CN and Cs2CO3-doped BPhen usingEq. (2.26)

D ¼ LUMOHAT�CN � LUMOCs2CO3:BPhen �Xi¼1;2

qwi ð2:26Þ

By calculation, D is 2.73 eV. Figure 2.12 gives the energy level alignment atthermal equilibrium based on the information for HAT-CN/Cs2CO3-doped BPheninterface.

Fig. 2.12 Energy leveldiagram of theHAT-CN/Cs2CO3-dopedBPhen interface determinedby the C-V method. Reprintedfrom [22]


2.3 Charge Transport Properties

The transport of charges in semiconductors is the fundamental process of anyoptoelectronic devices. The correct description and understanding on the transportare the basis for the optimization of optoelectronic devices. In inorganic semi-conductors, strong covalent bonds hold atoms together with well ordered config-urations. The energy band in this case extends continuously in the bulk, andtherefore, the delocalized charges can freely move along the band with a relativehigh mobility, as depicted in Fig. 2.13a. However, for most organic semiconduc-tors, the weak intermolecular forces between molecules are predominant. In thiscase, discrete energy band structure is dominant in the bulk. The freely propagationwave of charges as usually seen in inorganic semiconductor no longer exists. Inorganics, the molecular orientation and energetic profile are intrinsically disordered.As a result, the charge transport in organic semiconductors becomes a hoppingprocess that involves thermionic emission and tunneling of carriers between

Fig. 2.13 a Band and b hopping transport mechanism in different type of semiconductors

2.3 Charge Transport Properties 51

localized sites, as shown in Fig. 2.13b. This is an activated process; the mobilityincreases with increasing temperature and is a field dependent. It can be readilypredicted that the carrier mobility in most organic semiconductors is much smallerthan that in their inorganic counterparts.

There are several proposed models, mainly including the empirical Poole–Frenkel formula [25], the Gaussian disorder model (GDM) [26], and correlateddisorder model (CDM) [27], to describe the temperature- and field-dependentcharge transport in disordered organic semiconductors.

The empirical Poole–Frenkel is simply written as

l F; Tð Þ ¼ l0 exp �E0 � bffiffiffiffiF

p

kTeff

� �ð2:27Þ

with b ¼ffiffiffiffiffiffie3pee0

qand T�1

eff ¼ T�1 � T�10

where F is the electric field, E0 is the thermal activation energy at F = 0, k is theBoltzmann constant, T0 is the temperature at which Arrhenius plots of l at variousF intersect, and l0 is the mobility at T0. b is called the Poole–Frenkel coefficient,and it is in some cases reported to be temperature-dependent [28].

A reasonable assumption for the disorder based on the central limit theorem is aGaussian distribution. This has led to the development of GDM. The generalbehavior of the mobility as a function of both temperature and electric field in thepresence of diagonal and off-diagonal disorder is given as

l F; Tð Þ ¼ l0 exp � 2r3kT

� �2" #

� exp C r02 � R2� � ffiffiffiffi

Fp� �

if R� 1:5exp C r02 � 2:25ð Þ ffiffiffiffi

Fp� �

if R� 1:5

(ð2:28Þ

where C is a numerical constant and r′ = r/kT. r is the standard deviation of thedispersion in energy levels. R is the standard deviation associated with the posi-tional disorder. At high fields, l saturates, because the gain in electrostatic energycompensates the disorder. Using the GDM, it was shown that the experimentallyobserved Poole–Frenkel field dependence is a sign for charge carrier hopping in adisordered system. Still, at low fields ��105 V/m the predicted GDM produces adifferent field dependence than the experimentally observed Poole–Frenkelbehavior.

CDM introduced site correlations, that is, neighboring sites influence eachothers’ energy rather than being independent of each other, which reproduces thePoole–Frenkel-type field dependence at low electric fields and matches the GDM athigh fields. CDM has the following formation [29]

l F; Tð Þ ¼ l0 exp � 3r0d5

� �2

þC0 r03=2d � C� � ffiffiffiffiffiffiffiffi

eaFrd

r" #ð2:29Þ


where rd′ = rd/kT, rd = 2.35ep/ea2. a is the minimal charge–dipole separation. p isthe independently and randomly oriented dipole of moment. The model describes anon-dispersive mobility in correlated (e.g., dipolar) media, where C0 = 0.78, andC = 2. The parameter l0 may have additional temperature dependence due to otherless correlated sources of energy disorder or polaron effects.

The result of hopping transport caused by disorder generally leads the chargetransport in organic semiconductors to be space charge limited. The space chargelimited current (SCLC) obeys Child’s law [30]

J ¼ 98ee0l F; Tð ÞV

2

d3ð2:30Þ

where e is the dielectric constant, e0 is the vacuum permittivity, V the bias voltage,and d the thickness of organic layer.

Figure 2.14 shows the J-V characteristics of ITO/HAT-CN(10 nm)/HAT-CN:TAPC (30 wt%) (X nm)/TAPC(Y nm)/Al (device A: X = 50 nm, Y = 10 nm;device B: X = 100 nm, Y = 10 nm, and device C: X = 150 nm, Y = 10 nm) [19]. Itcan be seen that there are two regions in this log(J)-log(V) plot: ohmic conductanceregion with slope of 1 at low voltage and SCLC region with slope of about 2 at highvoltage. This indicates that the charge transport in HAT-CN/HAT-CN:TAPC/TAPC organic heterojunction is bulk-limited. Importantly, it is found thatincreasing the thickness of HAT-CN:TAPC does not reduce the current in devices,indicating that HAT-CN:TAPC bulk heterojunction is highly conductive. This alsofurther demonstrates that HAT-CN/HAT-CN:TAPC/TAPC organic heterojunctionpossesses excellent charge transport properties.

Figure 2.15 shows the J-V characteristics of Device 1: ITO/MoO3 (0.75 nm)/NPB (20 nm)/HAT-CN (60 nm)/NPB (20 nm)/MoO3 (5 nm)/Al and Device 2:ITO/MoO3 (0.75 nm)/NPB (100 nm)/MoO3 (5 nm)/Al [31]. They showed excel-lent SCLC property at high voltage region. However, it is clear that by insertion ofthe HAT-CN layer between NPB layers, the hole carrier mobility was remarkablyincreased for more than one order of magnitude, which has been attributed to the

Fig. 2.14 J-V characteristicsof ITO/HAT-CN/HAT-CN:TAPC/TAPC/Al devices withdifferent thicknesses ofHAT-CN:TAPC layer.Reprinted from [19]

2.3 Charge Transport Properties 53

carrier recombination at NPB/HAT-CN/NPB interfaces through coulombicinteraction.

2.4 Charge Generation Properties

In inorganic semiconductor heterojunctions, the space charge region at theheterojunction interface is formed by charge diffusion due to the high mobility andfree charge carrier density of inorganic semiconductors. In this case, the formedspace charge region is generally depleted, where the holes are presented on the sideof N-type semiconductor and the electrons are on the side of P-type semiconductor,and the charges are immobile. Therefore, the space charge region is highly resistant.However, the formation of the space charge region is a charge transfer process inorganic semiconductor heterojunctions due to the low mobility and free chargecarrier density in organic semiconductors. In addition to the formation ofdepletion-type heterojunctions like inorganic semiconductors, an accumulation-typeheterojunction can also be formed [7]. When the Fermi level of P-type organicsemiconductors is higher than that of N-type organic semiconductors, the electronswill be transferred from P-type semiconductors to N-type semiconductors, thus theside of N-type semiconductors accumulates electrons and the side of P-typesemiconductors accumulates holes. The accumulation-type organic heterojunctionsform a highly conductive space charge region due to the existence of a largenumber of free charge carriers. As shown in Fig. 2.16, the charge transfer pins theFermi level at the heterojunction interface, which is dominated by the gap statestailed from the HOMO onset of P-type organic semiconductors, resulting in that thecharges fill the band gap, therefore the conductivity is greatly enhanced. Theexperimental results have shown that a large transfer of charges between twoorganic semiconductors is a promising way to obtain high mobility and highlyconductive organic films, which are interesting not only for practical applicationsbut also for understanding the operation mechanisms in organic optoelectronicdevices [32–35].

Fig. 2.15 J-V characteristicsof ITO/MoO3/NPB/HAT-CN/NPB/MoO3/Al Device 1 and ITO/MoO3/NPB/MoO3/Al Device 2.Reprinted from [31]


The important applications of organic semiconductor heterojunctions are chargeinjectors and CGLs in OLEDs [36, 37]. It is clearly demonstrated that the chargegeneration behavior is independent of the work function of the contact electrodesbut was strongly dependent on the energy level alignment between the HOMO ofP-type organic semiconductors and the LUMO of N-type organic semiconductors.Therefore, by controlling the energy level, we can realize the accumulation-typeorganic

heterojunction with high conductance. It can be seen that using theaccumulation-type organic heterojunction as charge injectors can realizehigh-efficiency OLEDs independent on the work function of used metal electrodes,as CGLs can realize high power efficiency tandem OLEDs. The achievement of thegood performance in OLEDs has been attributed to the larger free charge generationand the high conductivity. Adachi et al. have verified this point [38]. Figure 2.17 isthe energy level diagram of F16CuPc/m-MTDATA organic heterojunction-baseddevices that is used to study. Obviously, in forward bias, both electrons and holescannot be injected into the organic layers because the work function of Al is notaligned to the energy levels of both the m-MTDATA LUMO and the F16CuPcHOMO. However, as shown in Fig. 2.18, a very large current is observed, clearlyindicating the large charge generation from the interface between m-MTDATA andF16CuPc heterojunction. It can be seen that the device current is greatly reduced asusing a-NPD instead of m-MTDATA. This behavior has been ascribed to thedifference in the HOMO levels of a-NPD (5.6 eV) and m-MTDATA (5.0 eV). Thisleads to the number of transferred charges from a-NPD to F16CuPc negligible. This

Fig. 2.16 Schematic diagram of charge transfer states in HAT-CN/a-NPD organic heterojunctionbefore contact and after contact. The density of states are assumed to be Gaussian

2.4 Charge Generation Properties 55

also implies that the charge separation at the F16CuPc/a-NPD interface is mainlyinduced by the external electric field. With respect to the energy level alignment,the work function of the contacts is unimportant, even when very low work functionmetals such as calcium (2.9 eV) are used as the anode, the current is almost thesame as that for the aluminum anode.

Figure 2.19 shows the I-V characteristics of F16CuPc (20 nm)/m-MTDATA(20 nm) heterojunction and m-MTDATA (30 nm) single-layer devices. Clearly, thesingle-layer m-MTDATA film shows a current as low as 2 � 10−11 A even at 50 V,and its conductivity is calculated to be of the order of 10−10 S/cm, which is a typicalvalue for organic semiconductors that have few charge carriers. On the other hand,the F16CuPc/m-MTDATA heterojunction film shows a current five orders ofmagnitude higher. As shown in the inset of Fig. 2.19, an ohmic behavior isobserved, implying that the carrier density at the F16CuPc/m-MTDATA hetero-junction interface is very high. By calculation, the carrier density at theF16CuPc/m-MTDATA heterojunction interface is as high as about 1 � 1018/cm3,

Fig. 2.17 Energy leveldiagram ofF16CuPc/m-MTDATAorganic heterojunction-baseddevices. Reprinted from [38]

Fig. 2.18 J-V characteristics of devices composed of Al/m-MTDATA/F16CuPc/Al,Al/m-MTDATA/F16CuPc/Ca/Al and Al/a-NPD/F16CuPc/Al. Reprinted from [38]


whereas that for a single layer of F16CuPc is about 1 � 1012/cm3, and the con-ductivity also reaches 0.001 S/cm.

References

1. B.L. Sharma, R.K. Purohit, Semiconductor Heterojunction (Pergamon Press, 1974)2. B.P. Rand, D.P. Burk, S.R. Forrest, Phys. Rev. B 75, 115327 (2007)3. N.C. Giebink, G.P. Wiederrecht, M.R. Wasielewski, S.R. Forrest, Phys. Rev. B 82, 155305

(2010)4. H. Bässler, Phys. Status Solidi B 175, 15 (1993)5. P. Sreearunothai, A.C. Morteani, I. Avilov, J. Cornil, D. Beljonne, R.H. Friend, R.T. Phillips,

C. Silva, L.M. Herz, Phys. Rev. Lett. 96, 117403 (2006)6. K.C. Kao, W. Hwang, Electrical Transport in Solids (Pergamon Press, Oxford, 1981)7. D.H. Yan, H.B. Wang, B.X. Du, Introduction to Organic semiconductor Heterojunctions

(Wiley, 2010)8. H.D. Sun, Q.X. Guo, D.Z. Yang, Y.H. Chen, J.S. Chen, D.G. Ma, ACS Photon. 2, 271–279

(2015)9. A.R. Riben, D.L. Feucht, Solid State Electron. 9, 1055 (1966)

10. A.R. Riben, D.L. Feucht, Int. J. Electron. 20, 583 (1966)11. A.J. Heeger, I.D. Parker, Y. Yang, Synth. Met. 67, 23 (1994)12. M. Koehler, I.A. Huümmelgen, Appl. Phys. Lett. 70, 3254 (1997)13. Q.X. Guo, H.D. Sun, J.X. Wang, D.Z. Yang, J.S. Chen, D.G. Ma, J. Mater. Chem. C 4, 376

(2016)14. R.L. Anderson, IBM J. Res. Dev. 4, 238 (1960)15. J.P. Donnelly, A.G. Milnes, I.E.E.E. Trans, Electron. Dev. 14, 63 (1967)16. B. van Zeghbroeck, Principles of Semiconductor Devices (University of Colorado, 2007)17. H. Kleemann, B. Luüssem, K. Leo, J. Appl. Phys. 111, 123722 (2012)18. L. Burtone, J. Fischer, K. Leo, M. Riede, Phys. Rev. B 87, 045432 (2013)19. H.D. Sun, Y.H. Chen, L.P. Zhu, Q.X. Guo, D.Z. Yang, J.S. Chen, D.G. Ma, Adv. Electron.

Mater. 1, 1500176 (2015)20. K. Walzer, B. Maennig, M. Pfeiffer, K. Leo, Chem. Rev. 107, 1233 (2007)21. S. Lee, J.-H. Lee, J.-H. Lee, J.-J. Kim, Adv. Funct. Mater. 22, 855 (2012)

Fig. 2.19 I-V characteristicsof m-MTDATA single layerand F16CuPc/m-MTDATAheterojunction on asemilogarithmic scale. Theinset shows the I-Vcharacteristics on a linearscale for the same data.Reprinted from [38]

2.4 Charge Generation Properties 57

22. S. Lee, J.-H. Lee, K.H. Kim, S.-J. Yoo, T.G. Kim, J.W. Kim, J.-J. Kim, Org. Electron. 13,2346 (2012)

23. J.-H. Lee, D.-S. Leem, H.-J. Kim, J.-J. Kim, Appl. Phys. Lett. 294, 123306 (2009)24. R.L. Anderson, Solid-State Electron. 5, 341 (1962)25. J. Frenkel, Phys. Rev. 54, 647 (1938)26. H. Bässler, Phys. Stat. Solidi B: Bas. Res. 175, 15–56 (1993)27. Y. Gartstein, E. Conwell, Chem. Phys. Lett. 245, 351 (1995)28. W. Bruütting, S. Berleb, A.G. Mückl, Org. Electron. 2, 1 (2001)29. S.V. Novikov, D.H. Dunlap, V.M. Kenkre, P.E. Parris, A.V. Vannikov, Phys. Rev. Lett. 81,

4472 (1998)30. M.A. Lampert, P. Mark, Current Injection in Solids (Academic Press, New York, 1970)31. W.S. Jeon, J.S. Park, L. Li, D.C. Lim, Y.H. Son, M.C. Suh, J.H. Kwon, Org. Electron. 13,

939 (2012)32. G. Yu, J. Gao, J.C. Hummelen, F. Wudl, A.J. Heeger, Science 270, 1789 (1995)33. J. Wang, H.B. Wang, X.J. Yan, H.C. Huang, D.H. Yan, Appl. Phys. Lett. 87, 093507 (2005)34. H. Alves, A.S. Molinari, H. Xie, A.F. Morpurgo, Nature Mater. 7, 574 (2008)35. Y.H. Chen, Q. Wang, J.S. Chen, D.G. Ma, D.H. Yan, L.X. Wang, Org. Electron. 13, 1121

(2012)36. Y.H. Chen, D.G. Ma, H.D. Sun, J.S. Chen, Q.X. Guo, Q. Wang, Y.B. Zhao, Light Sci. Appl.

5, e16042 (2016)37. Y.H. Chen, D.G. Ma, J. Mater. Chem. 22, 18718 (2012)38. T. Sakanoue, T. Irie, C. Adachi, J. Appl. Phys. 105, 114502 (2009)


Chapter 3Organic Semiconductor Heterojunctionsas Charge Injector in OrganicLight-Emitting Diodes

3.1 Basic Condition as Charge Injector

It is well-known that organic light-emitting diodes (OLEDs) are driven by injectedcharges from an anode and a cathode. Therefore, the proper energy level matchingbetween electrodes and organic charge-transport layers is necessary to obtain highlyefficient charge injection at a low voltage to obtain high-efficiency OLEDs [1].However, in conventional OLEDs, the charge carriers are directly injected into theorganic transport layers from electrodes and the injection barriers between theorganic transport layers and electrodes are unavoidable due to the mismatchbetween the work function of metal electrodes and the energy level ofcharge-transport layers, which greatly affects the performance of fabricated OLEDs[2]. Even though introducing interface layers and doped transport layers to furtherreduce the injection barriers, the low work function metals as cathode and the highwork function metals as anode have to be used, greatly limiting the selection ofelectrode metals [3]. This also means that the device performance is stronglydependent on the work function of metal electrodes in conventional OLEDs. Moreseriously, the instability caused by defects and high space electric field due tocharge accumulation at the interface between electrodes and organics are detri-mental to the efficiency and lifetime of OLEDs [2], which is very difficult to controlin the design of conventional OLEDs due to the limitations of the workingprinciple.

Many experiments have shown that organic semiconductor P/N heterojunctionsshow somewhat different characteristics from crystalline inorganic P/N junctions.One of the most important characteristics of the inorganic P/N junctions is recti-fication, which is actually a core technology in the semiconductor industry.However, organic semiconductor P/N heterojunctions show poor rectifying char-acteristics and have a high current under reverse bias. Under the reverse biascondition, the charges are generated at the junction interface and are transported viatunneling through the narrow depletion layer for several nanometers [4, 5]. This


59

interesting behavior of the organic semiconductor P/N heterojunctions has led themto be frequently utilized as a charge generation layer (CGL) and a recombinationlayer in tandem OLEDs and OPVs, respectively [6, 7].

As we see, organic semiconductor heterojunctions as CGLs in tandem OLEDscan effectively generate charges and realize the injection of charges into respectiveEL units under external electric fields. Similar to the metal electrodes, CGLs playthe important role of electrodes, although they are floated within the devices. Thisindicates theoretically that organic semiconductor heterojunctions should serve asthe electrodes to realize the injection of both electrons and holes but they arecompletely different from the metal electrodes in conventional OLEDs. For the caseof organic semiconductor heterojunctions, the injected charges are originated fromthe generated charges in heterojunction interface, and the injection is directly fromthe organic semiconductor heterojunctions into the EL units. In order to satisfy theinjection requirement, organic semiconductor heterojunctions as charge injectorsstill need to meet several requirements. First, the organic semiconductor hetero-junctions should be highly conductive in vertical direction. The use of organicsemiconductor heterojunctions as charge injectors is inevitable to increase thethickness of the whole device, resulting in higher driving voltage. The high con-ductivity can effectively avoid the voltage drop. Second, the effective chargegeneration in organic semiconductor heterojunction interface must be guaranteed.The organic semiconductor heterojunctions serve as electrode instead of metal, andthe charge carriers are all from the organic semiconductor heterojunctions for use indevice operation. Therefore, the charge generation must be close or even higherthan the injected charges from external electrodes. Third, a facile charge injection isalso needed since the generated charges should be fastly and effectively injectedinto adjacent emissive layers.

Lee et al. [8] used an organic P/N junction consisted of a ReO3-doped copperphthalocyanine (CuPc)/Rb2CO3-doped 4,7-diphenyl-1,10-phenanthroline (BPhen)as electron injector to successfully fabricate high-efficiency inverted OLEDs withITO cathode. The forward bias in the OLEDs corresponds to the reverse bias in theP/N junction. The voltage loss for generating electrons and holes was diminished,and the P/N junction showed very efficient electron and hole generation under areverse bias, reaching 100 mA/cm2 at 0.3 V. This resulted in the symmetric J-Vcharacteristics under the forward and reverse bias region, indicating that thisjunction is indeed highly conductive. Interestingly, the inverted OLEDs withorganic P/N junctions were found to have the same J-V-L characteristics inde-pendently of the work function of the bottom cathodes, in contrast to the otherdevices. This indicates that the organic P/N junction realized highly efficientelectron injection even though using high work function metal ITO as the cathode.

The high electric conductivity was also observed in the multi-alternatingm-MTDATA/F16CuPc junctions [9] and NPB/HAT-CN [10]. Them-MTDATA/F16CuPc film showed a conductivity up to S = 4 � 102 S/cm, and thehole mobility of NPB/HAT-CN film reached 5.3 � 10−1 cm2/V s. Clearly, theincrease in the number of alternating units (n) results, respectively, in an increase incapacitance and a decrease in resistance of the stack. This result suggests the

60 3 Organic Semiconductor Heterojunctions as Charge Injector …

increasing charges accumulated at the heterojunctions leading to reduction inoverall device resistance. It can be seen that when this multi-alternating unit appliednear the ITO anode of OLEDs, the device efficiency is greatly enhanced and thestability is also significantly improved.

Recently, Ma et al. simultaneously applied the pentacene/C60 (or C70) as holeinjector and electron injector to fabricate high-efficiency OLEDs [11, 12]. It can beseen that not only the efficiency, but also the stability is greatly improved. Mostsignificantly, the impressive performance can be achieved despite using an air- andchemistry-stable high work function metal, such as Au, Ag, or Cu, as the electriccontact, which has been suggested to be very difficult with conventional OLEDs.Because the novel charge injection architecture created here is based on a funda-mental physical understanding of semiconductor heterojunction theory, organicsemiconductor heterojunctions as charge injectors should be generally applicable toa wide range of phosphorescent and fluorescent devices and different coloreddevices, including white devices. It is believed that this finding offers anunprecedented versatility and a solid theoretical basis in the design of organicsemiconductor heterojunctions, thus greatly facilitating the further improvement inOLED performance for practical applications, which will ideally inspire furtherwork.

3.2 As Hole Injector for High-Efficiency OrganicLight-Emitting Diodes

Low energy barriers at electrode/organic film interfaces are desired for efficientcharge injection and are generally a prerequisite to high performance of OLEDs. Inthat regard, the interlayer of molybdenum trioxide (MoO3) has recently generatedconsiderable interest for hole injection enhancement and efficiency improvement inOLEDs [13, 14]. Except for direct ohmic contact formation between ITO and MoO3,a mechanism that holes extraction at MoO3/organic hole-transporting layer interfacedue to charge transfer rather than injection from anode has been demonstrated [15]. Itcan be seen that organic semiconductor heterojunctions possess a similar workingprocess and it should be predicted that using proper organic semiconductorheterojunction as injector can realize efficient hole injection. Ma et al. [11] devel-oped C60/pentacene organic heterojunction and used it as the hole injector to fab-ricate high-efficiency OLEDs. It has been clearly demonstrated that the charges canbe effectively generated due to the charge transfer from pentacene to C60 and thegenerated holes and electrons are then extracted and injected into the respective ELunits upon an external bias, finally leading to the light emission. Figure 3.1 showsthe structures of the conventional OLEDs as a reference (Fig. 3.1a) and the C60/pentacene organic heterojunction injector-based OLEDs (Fig. 3.1b).

Completely different from the reference OLEDs with metals as the chargeinjectors, a C60/pentacene organic heterojunction is located on the side of ITO in

3.1 Basic Condition as Charge Injector 61

the studied devices, where the ITO only plays a role of electric contact. In the fabri-cated OLEDs, a common metal organic phosphor of bis(2-phenylpyridine)iridiumacetylacetonate (Ir(ppy)2(acac)) is doped into a host of 4,4′,4″-tri(N-carbazolyl)triphenylamine (TCTA) as the emissive layer, which is sandwiched between thehole/exciton-blocking layer of 2,2′,2″-(1,3,5-benzenetriyl) tris-(1- phenyl-1H-benzimidazole) (TPBi) and the electron/exciton-blocking layer of TCTA. AP-type doped layer of TCTA:MoO3 and an N-type doped layer of TPBi:Li2CO3

are employed as the hole-injection/transporting layer and electron-injection/transporting layer, respectively.

To clarify that the injected holes are originated from the generated charges in theC60/pentacene heterojunction rather than those injected from the external ITOelectrode in the C60/pentacene heterojunction-based OLEDs, the hole-only deviceswith different hole injection structures, as shown in Fig. 3.2a, are fabricated.Figure 3.2b shows the current density–voltage (J-V) characteristics of these devices.It is clearly seen that ITO/C60 (20 nm)/TCTA:MoO3(40 nm)/TCTA (100 nm)/Al(device H-1) hardly shows any current flow despite the high bias voltage of 20 V.The extremely low currents in this device H-1 should be attributed to the largeinjection and transport barrier between ITO (*4.7 eV) and C60 (HOMO * 6.2 eV)[16] for holes and between Al (*4.3 eV) and hole-transporting organic TCTA(lowest unoccupied molecular orbital (LUMO) * 2.7 eV) [17] for electrons. Thisdemonstrated that the hole injection from ITO/C60 and the electron injection fromAl/pentacene are impossible. However, when inserting a pentacene layer in deviceH-1, the hole-only device of ITO/C60 (20 nm)/pentacene (10 nm)/TCTA:MoO3

(40 nm)/TCTA (100 nm)/Al) (device H-2) shows very large current. Since noexternal holes can be injected into the device from ITO electrode, as demonstrated indevice H-1, it is exclusively proven that the large current is obviously due to thegenerated charges in the C60/pentacene organic heterojunction under the externalelectric field induction. This strongly indicates that the C60/pentacene organicheterojunction is indeed an extremely effective charge injector. To further evaluate

Fig. 3.1 Schematic diagram of the fabricated OLEDs: a conventional OLED with a structure ofITO/TCTA:MoO3(70 nm)/TCTA(10 nm)/TCTA:Ir(ppy)2(acac)(20 nm)/TPBi(10 nm)/TPBi:Li2CO3(40 nm)/Al(120 nm) and b C60/pentacene organic heterojunction-based OLED with astructure of ITO/C60(20 nm)/pentacene(10 nm)/TCTA:MoO3(70 nm)/TCTA(10 nm)/TCTA: Ir(ppy)2(acac)(20 nm)/TPBi(10 nm)/TPBi:Li2CO3(40 nm)/Al(120 nm). Reprinted from [11]


the large current injection characteristics of the C60/pentacene organic heterojunc-tion, the hole-only device of ITO/TCTA:MoO3 (40 nm)/TCTA (130 nm)/Al (deviceH-3) was fabricated for a comparison. For the purpose of confirming the sameelectric field intensity, the total thickness of device H-3 was designed to be the sameas that of device H-2. As shown in Fig. 3.2b, the hole injection from ITO anode indevice H-3 is very effective due to the introduction of a P-type doped layer TCTA:MoO3, which are widely used in conventional OLEDs to enhance hole injection[18]. However, in comparison, the current in device H-3 is still less than that indevice H-2, further demonstrating the validity of C60/pentacene organic hetero-junction as charge injector.

Figure 3.3 displays the current density–luminance–voltage (J-L-V) and effi-ciency–luminance (E-L) characteristics of the fabricated green OLED with C60/pentacene organic heterojunction as hole injector and that of the conventionalOLED. The C60/pentacene-based device shows an approximate same turn-onvoltage as that conventional device, but works at low current density and higherluminance although it has large device thickness, indicating that the introduction ofC60/pentacene organic heterojunction is highly conductive and effectively generatescharges. As shown in Fig. 3.3b, c, the C60/pentacene-based OLED emits highefficiencies with respect to the conventional OLED. The maximum current effi-ciency and power efficiency arrive at 75.4 cd/A and 76.4 lm/W, respectively, andyet remain at 75.3 cd/A and 72.1 lm/W at 1000 cd/m2 luminance, which are higherthan those in the conventional OLED, indicating the highly efficient charge injec-tion property of the C60/pentacene organic heterojunction as charge injector,superior to the case of charge injection directly from electrodes.

Fig. 3.2 Device structures and J-V characteristics of three different hole-only devices: a devicestructures, b J-V characteristics. Reprinted from [11]

3.2 As Hole Injector for High-Efficiency Organic Light-Emitting Diodes 63

When changing C60 layer thickness in the C60/pentacene organic heterojunction,as shown in Fig. 3.4, the EL performance of the resulting OLEDs in drivingvoltage, luminance, and efficiency exhibits the dependence on C60 layer thicknessand the 20 nm C60 layer thickness leads to the best device efficiency. This shouldfurther demonstrate that the injected holes are originated from the charge carriersgenerated at the C60/pentacene organic heterojunction rather than those injectedfrom ITO anode.

Besides the C60/pentacene organic heterojunction that is well used as the holeinjector, actually many other heterojunction systems such as C60/CuPc, C60/ZnPc,C60/H2Pc [19], C70/pentacene [20], HAT-CN/TAPC, HAT-CN/m-MTDATA [21],and bulk heterojunctions, like C60:ZnPc [22], and HAT-CN/HAT-CN:TAPC/TAPC[23], can be also used as the hole injector to realize effective hole injection inOLEDs because they are excellent CGLs to effectively generate charges, whichhave got a good verification in tandem OLEDs by using them as CGLs.

As we know, one important condition of using an organic semiconductorheterojunction as charge injector is its high conductivity. Kwon et al. found [10]that multi-alternating NPB/HAC-CN organic heterojunction structure can provide

Fig. 3.3 EL performances of conventional OLED and OLED based on C60/pentacene organicheterojunction as hole injector: a J-V-L characteristics, b current efficiency as a function of currentdensity characteristics, and c power efficiency as a function of current density characteristics.Reprinted from [11]


high current conduction with high mobility and the current conduction wasimproved by increasing the P/N junctions made of intrinsic P-type hole transportlayer and N-type electron transport layer. Figure 3.5 shows the J-V characteristicswhich are originated from the hole current flows of three devices. It is clearly shownthat at a given constant voltage of 1.0 V, the current densities are 5.03 mA/cm2,12.92 mA/cm2, and 41.76 mA/cm2 for Devices A, B, and C, respectively. The holecurrent conduction in Devices B and C is dramatically improved by 257 and 830%compared with Device A at 1 V.

The high hole current conduction in multi-junction devices indicates that themulti-junction films possess higher hole mobility, which can be well determined byspace charge limited current (SCLC) model [24]. Figure 3.6a shows the SCLCbehaviors of hole- and electron-only devices D, E, F, G, and H. It is shown thatDevices G and H gave 5.3 ± (0.3) � 10−1 cm2/Vs and 9.9(±0.4) � 10−2 cm2/Vs ofhole mobilities at the 0.3 MV/cm electric field, respectively, which are higher than2.2 ± (0.2) � 10−4 cm2/Vs (hole mobility) for NPB (Device D), 6.4 ± (0.3) �10−5 cm2/Vs (hole mobility) for 2-TNATA (Device E), and 3.3 ± (0.1) � 10−2

cm2/Vs (electron mobility) for HAT-CN (Device F) at 0.3 MV/cm (see Fig. 3.6b).

Fig. 3.4 EL performance of OLEDs based on C60/pentacene organic heterojunction as holeinjector at different C60 layer thicknesses: a J-V-L characteristics, b current efficiency as a functionof current density characteristics, and c power efficiency as a function of current densitycharacteristics. The device structure: ITO/C60(X nm)/pentacene(10 nm)/TCTA:MoO3(70 nm)/TCTA(10 nm)/TCTA: Ir(ppy)2(acac)(20 nm)/TPBi(10 nm)/TPBi:Li2CO3(40 nm)/C60(20 nm)/pentacene(10 nm)/Al(120 nm), X = 10, 20, 30, 40. Reprinted from [11]


Fig. 3.5 J-V characteristics of hole-only devices A, B, C (Device A: ITO/HAT-CN (40 nm)/NPB(60 nm)/MoO3 (5 nm)/Al; Device B: ITO/MoO3 (0.75 nm)/NPB (30 nm)/HAT-CN (40 nm)/NPB(30 nm)/MoO3 (5 nm)/Al; Device C: ITO/MoO3 (0.75 nm)/NPB (20 nm)/HAT-CN (20 nm)/NPB(20 nm)/HAT-CN (20 nm)/NPB (20 nm)/MoO3 (5 nm)/Al). Reprinted from [10]

Fig. 3.6 a J-V characteristics of devices D, E, F, G, H. (Dotted lines mean ideal SCLC curves),b mobility characteristics of P-only, N-only, and P/N/P devices (Device D: ITO/MoO3 (0.75 nm)/a-NPB (100 nm)/MoO3 (5 nm)/Al; Device E: ITO/MoO3 (0.75 nm)/2-TNATA (100 nm)/MoO3

(5 nm)/Al; Device F: ITO/HATCN/Al; Device G: ITO/MoO3 (0.75 nm)/a-NPB (20 nm)/HAT-CN(60 nm)/a-NPB (20 nm)/MoO3 (5 nm)/Al; Device H: ITO/MoO3 (0.75 nm)/2-TNATA (20 nm)/HATCN (60 nm)/2-TNATA (20 nm)/MoO3 (5 nm)/Al). Reprinted from [10]


Such high current conduction with high mobility has been attributed to the carrierrecombination at P/N/P interfaces through coulombic interaction.

Because of the high conduction and hole mobility properties, using themulti-alternating NPB/HAC-CN organic heterojunction as hole injector will reducethe operational voltage and improve the efficiency of the fabricated OLEDs.Figure 3.7a, b shows the J-V-L and efficiency characteristics of the fluorescent blueDevices I, J, and K, respectively. It is shown that the operational voltage at1000 cd/m2 luminance is decreased to 4.3 V in Device K with increasing P/N/P junctions from 5.2 V in Device I, whereas the current density is increased from6.46 to 16.07 mA/cm2 at 4.0 V. As shown in Fig. 3.7b, the power efficiency at1000 cd/m2 was 2.98 lm/W in Device K (four P/N junctions), which is *21.6%

Fig. 3.7 a J-V characteristics of the fabricated blue fluorescent devices I, J, and K. The inset givesluminance versus voltage characteristics of the blue fluorescent devices I, J, and K. b Powerefficiency as a function of luminance of the devices I, J, and K. The inset shows EL spectra ofdevices I, J, and K at 1000 cd/m2 luminance. The device structures are Device I: ITO/HATCN(40 nm)/NPB (40 nm)/MADN: BD-1 (5%, 40 nm)/TmPyPb (10 nm)/Liq (1 nm)/Al (100 nm);Device J: ITO/DNTPD (20 nm)/HATCN (40 nm)/NPB (20 nm)/MADN: BD-1 (5%, 40 nm)/TmPyPb (10 nm)/Liq (1 nm)/Al (100 nm); Device K: ITO/DNTPD (10 nm)/HATCN (20 nm)/NPB (20 nm)/HATCN (20 nm)/NPB (10 nm)/MADN: BD-1 (5%, 40 nm)/TmPyPb (10 nm)/Liq(1 nm)/Al (100 nm). Reprinted from [10]

improved result compared to that of Device I (2.45 lm/W). This result indicates thatthe charge recombination interfaces in the multiple P/N junctions are very effectivein reducing driving voltage and improving efficiency.


The highly electric conductivity behavior was also observed in the multilayerstack of m-MTDATA/F16CuPc films [25]. As shown in Fig. 3.8, the current densityin the 30-unit device is 267 mA/cm2, which is about three orders of magnitude higherthan that in the 1-unit device. This current increase has been attributed to thereduction of film resistance. The impedance data show that the circuit resistance issignificantly dropped by three orders of magnitude from *0.2 MX to *0.4 kXwhen the number of alternating units is increased from one to six, also accompaniedby the increase of capacitance, keeping a total thickness of 300 nm. This resultsuggests the increasing charges accumulated at the heterojunctions, leading tothe reduction in overall film resistance. It can be seen that the application of the highconductive units in OLEDs results in the stability enhancement, as shown in Fig. 3.9.

Figure 3.10 gives the generation and recombination processes of charges in P/Nand P/N/P organic heterojunctions. It is shown that the charge generation behaviorcan be generated because it separates holes and electrons at the junction as a P/N junction interface with these P and N materials is made. When P/N/P multipleinterfaces are stacked, the hole current can flow without any serious barrier as

Fig. 3.8 J-V characteristic of 30-unit device (solid-line), single unit device (dashedline) and30-units with SubPc interlayer device (dotted-line). The device configuration: ITO/[m-MTDATA/F16CuPc] � 30 (300 nm)/BPhen (8 nm)/Al, ITO/m-MTDATA (150 nm)/F16CuPc(150 nm)/BPhen (8 nm)/Al, and ITO/[m-MTDATA (4 nm)/SubPc (1 nm)/F16CuPc(4 nm)/SubPc (1 nm)] � 29/m-MTDATA (4 nm)/SubPc (1 nm)/F16CuPc (4 nm)/BPhen (8 nm)/Al. Reprinted from [25]

Fig. 3.9 Luminance of theOLEDs with and without themulti-alternating unit as afunction of time. Deviceconfiguration: ITO/[m-MTDATA (5 nm)/F16CuPc (5 nm)] � 0 or7/NPB (70 nm)/Alq3(60 nm)/LiF (1 nm)/Al.Reprinted from [25]


shown in Fig. 3.10 (right) due to enough small barrier. In this configuration, oneinterface (right side of HAT-CN layer) can work as a charge generation part and theother interface (left side of HATC-N layer) can work as a charge recombinationpart. Obviously, the multiple intrinsic P/N/P junctions can provide non-radiativecoulombic interaction and thus carrier transportation is improved significantly.

According to Langevin recombination theory, the current density (J) attracted bycoulombic interaction between holes and electrons within Langevin radius could beexpressed by the following equation [26]:

J ¼ nheltF ð3:1Þ

where e is the unit charge, nh is the hole densities, F is the electric field, andltðlt ¼ le þ lhÞ is the total mobility. le and lh are the electron and hole mobility,respectively. In general, hole or electron conduction in P- or N-type organicsemiconductor materials uses only hole mobility or electron mobility. However, inthis equation we expect that both mobilities of holes and electrons can contribute tomake the current conduction in organic semiconductor devices. If recombinationinterface is generated by P/N junctions in organic semiconductor devices, we maygenerate enhanced current conduction from both contributions of hole and electronmobilities within Langevin radius.

The coulombic capture radius can be expressed by Langevin:

rc ¼ e2=ð4pere0kBTÞ ð3:2Þ

where e0 is the vacuum permeability, er is the relative dielectric constant of thesemiconductor, T is the temperature, and kB is the Boltzmann’s constant, respec-tively. This relation is valid if we suppose a much smaller mean-free path compared

Fig. 3.10 Generation and recombination processes in P/N and P/N/P organic heterojunctions.Reprinted from [10]


to the thermal capture radius because charge transport normally takes place byhopping between molecules and the mean-free path is on the order of the intersitedistance (a) of *1–2 nm. At room temperature and with a relative dielectricconstant (er) of *3, typical for organic semiconductors, the thermal capture radius(rc) is *18.5 nm. The fact that the increased mobility values of Devices G and Hgiven in Fig. 3.6b is much higher than the intrinsic electron mobility of HATCN,indicating that the current conduction by both contributions of hole and electronmobility is existed within Langevin coulombic capture radius in NPB/HAT-CNorganic heterojunction system.

3.3 As Electron Injector for High-Efficiency OrganicLight-Emitting Diodes

Similarly, as an electron injector, organic semiconductor heterojunctions shouldgenerate charge carriers with minimal extra voltage at the junction and the gener-ated electrons should be effectively injected into the emissive unit under reversebias. As shown above, C60/pentacene organic heterojunction can be used as a holeinjector. Actually, it is a charge generation for both electrons and holes. Therefore,the C60/pentacene organic heterojunction is also used for electron injector.Figure 3.11 is the structure of the fabricated green OLEDs based on the C60/pentacene organic heterojunction as electron injector. Completely different from theOLEDs with metals as electron injector, a C60/pentacene organic heterojunction islocated on the side of Al cathode, whereas the Al cathode only plays a role ofelectric contact.

To clarify that the injected electrons originate from the generated charges in theC60/pentacene organic heterojunction rather than from those injected from theexternal electrode (Al) in C60/pentacene organic heterojunction-based OLEDs,the electron-only devices (see Fig. 3.12a) are fabricated. Figure 3.12b shows the

Fig. 3.11 Schematic diagram of the fabricated green OLEDs based on the C60/pentacene organicheterojunction as electron injector. This device structure is ITO/TCTA:MoO3 (70 nm)/TCTA(10 nm)/TCTA: Ir(ppy)2(acac) (20 nm)/TPBi (10 nm)/TPBi:Li2CO3 (40 nm)/C60 (20 nm)/pen-tacene (10 nm)/Al (120 nm)


J-V characteristics of these devices. It is clearly observed that the electron-onlydevice of ITO/TPBi (100 nm)/TPBi:Li2CO3 (40 nm)/pentacene (10 nm)/Al (deviceE-1) shows hardly any current flow despite the high bias voltage of 20 V betweenITO positive bias and Al negative bias. The extremely low current in this deviceshould be attributed to the large electron injection barriers between Al (*4.3 eV)and hole-transporting organic TCTA layer (lowest unoccupied molecular orbital(LUMO) *2.7 eV) and pentacene (LUMO * 0 eV). This demonstrates that theelectron injection from Al/pentacene interface is impossible. However, wheninserting a C60 layer in device E-1, the electron-only device of ITO/TPBi (100 nm)/TPBi:Li2CO3 (40 nm)/C60 (20 nm)/pentacene (10 nm)/Al (device E-2) shows verylarge electron current, even higher current than that of conventional electron-onlydevice of ITO/TPBi (130 nm)/TPBi:Li2CO3 (40 nm)/Al (device E-3). Because noexternal charge carriers can be injected into the device from Al electrode in deviceE-2, it is definitively proven that the large current is obviously due to the generatedcharges in the C60/pentacene organic heterojunction under the external electric fieldinduction. This strongly indicates that the C60/pentacene organic heterojunction isan extremely effective electron injector.

The performances of the resulting OLEDs with and without C60/pentacene or-ganic heterojunction as the electron injector are shown in Fig. 3.13. The maximumcurrent efficiency and power efficiency arrive at 72.7 cd/A and 72.6 lm/W,respectively, and yet remain at 72.5 cd/A and 67.7 lm/W at 1000 cd/m2 luminancein heterojunction-based OLEDs, which are higher than those in the conventionalOLEDs, indicating the highly efficient electron injection property of the C60/pen-tacene organic heterojunction as electron injector, superior to the case of electroninjection directly from electrode.

Fig. 3.12 Device structures (a) and J-V characteristics (b) of three electron-only device E-1, E-2,and E-3. Reprinted from [11]

3.3 As Electron Injector for High-Efficiency Organic … 71

Figure 3.14 depicts the EL performances of the OLEDs based on C60/pentaceneorganic heterojunction as electron injector at different pentacene thicknesses. It isshown that the current, luminance and efficiency show the dependence on pen-tacene layer thickness, indicating that the injected electrons are indeed from theC60/pentacene organic heterojunction as electron injector and greatly efficient.

It can be seen that the electron injection via organic heterojunction is completelydifferent from the direct injection by metal electrode. In a conventional device, asshown in Fig. 3.15a, the electrons are usually injected over the injection barrier bythermionic emission or a tunneling mechanism from the electrode to the organiclayer. To reduce the injection barrier, a thin interfacial layer and an electrical dopinglayer are widely used [13, 14]. However, the injection current still depends on thework function of the used electrodes. Evidently, this problem can be resolved by theP/N organic heterojunction structure supplying holes to the electrode and electronsto the emission layer simultaneously from the interface of the junction, as shown inFig. 3.15b. Organic P/N junctions are known to generate electrons and holes underreverse bias by tunneling of electrons from the HOMO level of the P-layer to theLUMO level of the N-layer through a narrow depletion layer at the junctions [4]. Animportant result of the structure using P/N organic heterojunctions is that the

Fig. 3.13 EL performances of the resulting OLEDs with and without C60/pentacene organicheterojunction as electron injector. a J-V-L characteristics, b current efficiency as a function ofluminance characteristics, and c power efficiency as a function of luminance characteristics.Reprinted from [11]


electrons can be injected and supplied to an emission unit independently of the workfunction of the bottom cathode. This has been experimentally confirmed by usinginverted OLEDs with three different cathodes: poly(3,4-ethylenedioxythiophene)–polystyrenesulfonic acid, UV-O3-treated ITO, and non-treated ITO electrodes withdifferent work functions, respectively [8]. The results are shown in Fig. 3.16. It isinterestingly seen that the inverted OLEDs with organic P/N junctions were found tohave the same J-V-L characteristics independently of the work function of the bottomcathodes, in contrast to the other devices.

Fig. 3.14 EL performances of the OLEDs based on C60/pentacene organic heterojunction aselectron injector at different pentacene thicknesses. a J-V-L, b current efficiency as a function ofcurrent density, and c power efficiency as a function of current density characteristics. Devicestructure is ITO/TCTA:MoO3(70 nm)/TCTA(10 nm)/TCTA: Ir(ppy)2(acac)(20 nm)/TPBi(10 nm)/TPBi:Li2CO3(40 nm)/C60(20 nm)/pentacene(X nm)/Al(120 nm), X = 5, 10, 15, 20.Reprinted from [11]

Fig. 3.15 Schematicdiagrams of the electroninjection mechanism in a anormal electrode/organicjunction and b anelectrode/P/N organicjunction


Obviously, the remarkable property of organic heterojunction independentelectron injection on the work function of the bottom electrode is very useful tofabricate high-efficiency inverted OLEDs. It is well-known that inverted OLEDswith a bottom cathode have attracted increasing attention for display applicationsbecause of their easy integration with the N-type transistors based on low-cost andhighly uniform amorphous silicon (a-Si), and transparent amorphous oxide semi-conductors (TAOSs) [27]. Up to date, indium tin oxide (ITO) has been widely usedas the transparent electrode, but the dogged issue of using ITO as the cathode ininverted OLEDs is the large energy barrier for electron injection due to its highwork function, resulting in the poor performance of the fabricated inverted OLEDs[28]. As shown above, using organic semiconductor heterojunctions as electroninjector can resolve this issue due to their highly efficient electron injection prop-erty. Ma et al. fabricated high-efficiency red, green, and blue inverted OLEDs basedon an intrinsic P-type organic/bulk heterojunction/intrinsic N-type organic com-posite junction structure as electron injector on ITO cathode [29]. It can be seen thatthe state-of-art red, green, and blue OLEDs achieved the maximum efficiencies of14.1% (25.0 cd/A, 27.2 lm/W), 22.4% (86.8 cd/A, 97.0 lm/W), and 14.3%(37.5 cd/A, 34.8 lm/W), respectively. Figure 3.17 shows the J-V characteristics offour electron-only devices with ITO/m-MTDATA (10 nm)/m-MTDATA :HAT-CN (2 : 1) (15 nm)/HAT-CN (10 nm)/Be(pp)2:Li2CO3 (3%, 30 nm)/Be(pp)2(100 nm)/Al (device E1), ITO/m-MTDATA (10 nm)/Be(pp)2:Li2CO3 (3%,30 nm)/Be(pp)2 (100 nm)/Al (device E2), ITO/HAT-CN (10 nm)/Be(pp)2:Li2CO3

(3%, 30 nm)/Be(pp)2 (100 nm)/Al (device E3), and ITO/m-MTDATA (10 nm)/m-MTDATA : HAT-CN (2 : 1) (30 nm)/HAT-CN (10 nm)/Be(pp)2:Li2CO3 (3%,15 nm)/Be(pp)2 (100 nm)/Al (device E4), here ITO acted as the cathode and Al asthe anode. The structures simulate the real sequence in the inverted OLEDs. It isclearly shown that the current density of devices E1 and E4 is much higher than

Fig. 3.16 J-V and L-V characteristics of the fabricated inverted OLEDs with five different cases ofITO electrode treatment. The value of the work function of the bottom ITO cathode therein isshown in right side. Reprinted from [10]


those of the devices E2 and E3, indicating that m-MTDATA/m-MTDATA:HAT-CN/HAT-CN heterojunction possesses stronger electron injection ability.Furthermore, it can be seen from the comparison of current characteristics ofdevices E1 and E4 that the conductivity of MTDATA:HAT-CN is higher than Be(pp)2:Li2CO3. This is also the reason why the m-MTDATA/m-MTDATA:HAT-CN/HAT-CN organic heterojunction supplies more electron current inheterojunction-based devices than that in conventional devices.

Indeed, the m-MTDATA/m-MTDATA:HAT-CN/HAT-CN organic heterojunc-tion as electron injector shows the independent electron injection capability on thework function of the used cathode. Figure 3.18 shows the EL performance of thefabricated blue inverted OLEDs with the m-MTDATA/m-MTDATA:HAT-CN/HAT-CN heterojunction electron injector on ITO cathode treated bydifferent treating conditions. Here, Devices F1, F2, and F3 correspond to the casesof UV-O3-treated ITO, pristine ITO, and UV-O3-treated ITO covered by 2 nm Al,respectively. Clearly, these devices show the same turn-on voltage as low as 2.6 V,and the current density, luminance, and device efficiency only changed slightly withthe different treating conditions, corresponding to the different work functions ofITO cathode. This obviously demonstrates that the used organic heterojunction aselectron injector on ITO cathode plays a very important role in reducing the deviceoperational voltage, thus improving device performance due to its highly efficientcharge generation and injection effects.

As we know, the amount of generated charges is determined by the properties ofthe used P-type and N-type organic materials in the organic heterojunctions, notonly their mobilities, but also their relative energy level positions [30]. Figure 3.19shows the EL performances of two green-inverted OLEDs based onm-MTDATA/m-MTDATA:HAT-CN/HAT-CN (green device 1) and TAPC/TAPC:HAT-CN/HAT-CN (green device 2) heterojunctions as the electron injector. It isshown that the m-MTDATA/m-MTDATA:HAT-CN/HAT-CN heterojunction leadsto higher device efficiency than the TAPC/TAPC:HAT-CN/HAT-CN, indicating

Fig. 3.17 J-V characteristics of the electron only devices with m-MTDATA/m-MTDATA:HAT-CN/HAT-CN, m-MTDATA and HAT-CN as the injectors. The inset gives the schematicillustration of the electron-only devices. Reprinted from [29]


Fig. 3.18 EL characteristics of the fabricated blue inverted OLEDs with them-MTDATA/m-MTDATA:HAT-CN/HAT-CN heterojunction electron injector on ITO cathodetreated by different treating conditions. a J-L-V, b CE-L, c PE-L, and d EQE-L. Here, Devices F1,F2, and F3 correspond to the cases of UV-O3 treated ITO, pristine ITO and UV-O3 treated ITOcovered by 2 nm Al, respectively. Reprinted from [29]

Fig. 3.19 a J-L-V, b CE-L, c PE-L and d EQE-L characteristics of the resulting green invertedOLEDs based on m-MTDATA/m-MTDATA:HAT-CN/HAT-CN (Green device 1) andTAPC/TAPC:HAT-CN/HAT-CN (Green Device 2) heterojunctions. Reprinted from [29]


that m-MTDATA should be more suitable than TAPC for the charge generation andinjection in HAT-CN-based heterojunction. To demonstrate the impact, twoheterojunction-type devices with ITO/m-MTDATA (10 nm)/m-MTDATA :HAT-CN (2 : 1, 50 nm)/HAT-CN (10 nm)/Al and ITO/TAPC (10 nm)/TAPC :HAT-CN (2 : 1, 50 nm)/HAT-CN (10 nm)/Al are fabricated. Here, the forward biasrefers to ITO as cathode and Al as anode, and the reverse bias refers to ITO asanode and Al as cathode. As shown in Fig. 3.20, the m-MTDATA-based hetero-junction generates larger current than the TAPC-based heterojunction. Clearly,although TAPC has a higher hole mobility of *10−2 cm2/Vs than m-MTDATA(*10−5 cm2/Vs), which is directly related to the charge generation in the formedheterojunction, the larger energy level difference between the LUMO of HAT-CN(*4.8 eV) and the HOMO of TAPC (*5.4 eV) than that of m-MTDATA(*5.1 eV) greatly reduces the charge generation. Obviously, the construction oforganic heterojunctions as charge injectors must consider the match of the energylevel positions between P- and N-type organic semiconductors.

3.4 As Hole and Electron Injectors for High-EfficiencyOrganic Light-Emitting Diodes

As shown above, organic heterojunctions as charge injectors possess excellentinjection capability of holes as well as electrons. This also indicates thathigh-efficiency OLEDs can be achieved by simultaneously using organic hetero-junctions, respectively, as hole injector and electron injector in the fabricatedOLEDs. Figure 3.21 shows the schematic diagrams of the resulting green phos-phorescent OLEDs without (a) and with (b) C60/pentacene organic heterojunctionas the charge injectors against both side of ITO anode and Al cathode [11]. Unlikeconventional OLEDs with metals as charge injectors, a C60/pentacene organic

Fig. 3.20 J-V characteristics of ITO/m-MTDATA/m-MTDATA:HAT-CN/HAT-CN/Al (blacksquares) and ITO/TAPC/TAPC:HAT-CN/HAT-CN/Al (blue triangles) devices. The inset showsthe schematic illustration of two devices. Reprinted from [29]


heterojunction is located on each side of ITO anode and Al cathode in the studieddevices (Fig. 3.21b), where the ITO and Al only play the role of electric contact.The holes and electrons are generated by charge transfer from pentacene to C60.And the generated holes and electrons are then extracted and injected intorespective EL units upon an external bias and finally lead to the light emission.

Figure 3.22 displays the J-L-V (left) and efficiency–current density (right)characteristics of the OLEDs with and without C60/pentacene organic heterojunc-tion as charge injectors. It can be observed that the OLEDs with heterojunctioncharge injectors show higher EL efficiency. The maximum current efficiency andpower efficiency arrive at 75.9 cd/A and 76.0 lm/W, respectively, and remain at75.6 cd/A and 72.1 lm/W at 1000 cd/m2 luminance, which are higher than those inconventional OLEDs. This is attributed to the highly efficient charge injectionproperty of the C60/pentacene organic heterojunction as charge injectors, which issuperior to the direct charge injection from electrodes. Moreover, the C60/pentaceneorganic heterojunction-based device also works at low current density, indicatingmore balanced charge transport and recombination in the device.

To demonstrate the stable metals as electrodes in organic heterojunctioninjectors-based OLEDs, the green phosphorescent OLEDs based on C60/pentaceneorganic heterojunction as charge injectors with different high work function metalsof Au (*5.1 eV), Ag (*4.4 eV), and Cu (*4.7 eV) to replace the low workfunction metal Al as the cathode contact. As shown in Fig. 3.23, although thecurrent density and luminance at the same voltage show certain variations, it isimpressive that they show almost same efficiency and the maximum current effi-ciency and power efficiency can reach 73.2 cd/A and 72.9 lm/W for Ag electrodedevice, 74 cd/A and 68 lm/W for Cu electrode device, and 72.9 cd/A and69.1 lm/W for Au electrode device. This also further proves the electrode work

Fig. 3.21 Schematic diagram of the fabricated OLEDs without (a) and with (b) C60/pentaceneorganic heterojunction as the charge injectors against both side of ITO anode and Al cathode.a ITO/TCTA:MoO3(70 nm)/TCTA(10 nm)/TCTA: Ir(ppy)2(acac)(20 nm)/TPBi(10 nm)/TPBi:Li2CO3(40 nm)/Al(120 nm). b ITO/C60(20 nm)/pentacene(10 nm)/TCTA:MoO3(70 nm)/TCTA(10 nm)/TCTA: Ir(ppy)2(acac)(20 nm)/TPBi(10 nm)/TPBi:Li2CO3(40 nm)/C60(20 nm)/pentacene(10 nm)/Al(120 nm). Reprinted from [11]


function-independent injection property of organic heterojunctions as chargeinjectors. Obviously, the introduction of organic heterojunctions as a new way ofcharge injection would greatly enlarge the choice of device electrode withoutconsidering the extreme match between the work function of electrode and theenergy level of charge transport layer in OLEDs.

The same high-efficiency green phosphorescent OLEDs based on C70/pentaceneorganic heterojunction as hole and electron injectors have also been realized [12].The maximum power efficiency, current efficiency, and external quantum efficiencyreach 80.2 lm/W, 72.8 cd/A, and 19.2% and yet keep 76.1 lm/W, 72.6 cd/A, and19.0% at 1000 cd/m2 luminance, respectively, indicating that C70/pentacene or-ganic heterojunction is also excellent charge injector to realize highly efficient holeand electron injection.

To demonstrate the hole and electron injection mechanism of organic hetero-junctions, the hole-only device D1: ITO/C70(5 nm)/pentacene(10 nm)/MoO3(3 nm)/TAPC:MoO3(10%, 50 nm)/TAPC (20 nm)/Al and the electron-onlydevice D2: ITO/BPhen(20 nm)/BPhen: Li2CO3(3%, 50 nm)/Li2CO3(1 nm)/C70(5 nm)/pentacene(10 nm)/Al are fabricated, and the ordinary hole-only deviceD3: ITO/MoO3(3 nm)/TAPC:MoO3(10%, 50 nm)/TAPC (20 nm)/Al and the

Fig. 3.22 J-L-V (left) andefficiency–current density(right) characteristics of theOLEDs with and without C60/pentacene organicheterojunction as chargeinjectors. Reprinted from [11]

3.4 As Hole and Electron Injectors for High-Efficiency … 79

electron-only device D4: ITO/BPhen(20 nm)/BPhen: Li2CO3(3%, 50 nm)/Li2CO3(1 nm)/Al are also fabricated as comparison [12]. It is noticed that TAPCand Bphen could effectively block the electron and hole injection, respectively.Therefore, the injected holes and electrons could only be through C70/pentacene andthen reach the adjacent transport layer in D1 and D2. The J-V characteristics ofdevices D1, D2, D3, and D4 at room temperature are shown in Fig. 3.24. As wesee, the current densities of D1 and D2 are comparable to those of D3 and D4,which demonstrates the efficient charge injection from the C70/pentacene organicheterojunction. What impressed us most is that the balance of hole current andelectron current could be further improved when the C70/pentacene organicheterojunction is employed as the charge injectors, which should contribute to theremarkable performance of fabricated OLEDs.

To further elucidate the injection mechanism, the J-V characteristics of devicesD1 and D2 under different temperatures are measured, as shown in Fig. 3.25. Thecurrent density–temperature characteristics of devices D1 and D2 at fixed voltagesfrom 137 to 287 K are also summarized in Fig. 3.25. Usually, the Shockleyequation is used to describe the J-V characteristics of organic heterojunctions,which predicts a linear relation between log(J) and 1/T [31]:

Fig. 3.23 a J-V-L characteristics, b current efficiency as a function of current densitycharacteristics, and c power efficiency as a function of current density characteristics of thegreen phosphorescent OLEDs based on C60/pentacene organic heterojunction as both hole andelectron injectors with Au, Ag, Cu, and Al metal contact electrodes. The device structure isITO/C60(20 nm)/pentacene(10 nm)/TCTA:MoO3(70 nm)/TCTA(10 nm)/TCTA: Ir(ppy)2(acac)(20 nm)/TPBi(10 nm)/TPBi:Li2CO3(40 nm)/C60(20 nm)/pentacene(10 nm)/Au, Ag, Cu or Al(120 nm). Reprinted from [11]


Fig. 3.24 J-V characteristics of single-carrier devices D1, D2, D3, and D4. Reprinted from [12]

Fig. 3.25 Current density–temperature characteristics of devices D1 (a) and D2 (b) at fixedvoltages. J-V characteristics of device D1 (c) and D2 (d) at different temperatures, where the bluesolid lines are the fitting results by Eq. 3.4. Reprinted from [12]


J ¼ J0 expqVnkT

� �� 1

� �ð3:3Þ

where J0 is the saturation current density, q is the charge quantity of an electron, k isthe Boltzmann constant, T is the temperature, and n is the ideal factor. However, theresults, as shown in the inset in Fig. 3.25a, b, deviate seriously from the Shockleyrelation. Instead, a clear linear relation between log(J) and T could be observed,which is a strong indication of a tunnel process [21]. Furthermore, as shown inFig. 3.25c, d, the J-V characteristics of devices D1 and D2 could be perfectly fittedby the modified Fowler–Nordheim (F–N) Tunnel equation [32] as follows:

lnJV2

� �¼ �P1

Vþ ln

P2

V

� �� ln sin

P3

V

� �� ð3:4Þ

where J is the current density, V is the applied voltage, and P1, P2, and P3 areconstants that are to be determined. This indicates that the electron and holeinjection of C70/pentacene organic heterojunction is a tunnel process.

In the charge injection processes of organic heterojunctions, we should say thatthe electron injection from N-type layer to its adjacent electron-transporting layerhas to consider because there generally exists a larger energy level differencebetween them; e.g., the LUMO energy level of BPhen (–3.0 eV) is much higherthan C70 (–4.0 eV). This means that the electrons on C70 have to overcome anenergy barrier of about 1.0 eV before reaching the emission unit. Therefore,Li2CO3-doped BPhen is used as the electron-transporting layer to reduce theelectron injection barrier, thus enhancing the electron injection. As shown, neitherthe decrease of power efficiency nor the increase of turn-on voltage caused by thisinterface energy barrier is observed by comparing all devices fabricated above. Thismeans that the electron injection from C70 into BPhen:Li2CO3 is very efficient. Inorder to further demonstrate if the electron-transporting layer has an effect on theelectron injection, two green phosphorescent OLED devices A3 and A5, respec-tively, with BPhen:Li2CO3 and Bepp2:Li2CO3 electron-transporting layers arefabricated. Figure 3.26 gives their EL performance characteristics. It can be seenthat device A5 shows much higher voltage and lower luminance than device A3 at afixed current density. As a result, the device A5 emits much lower power efficiencythan device A3, indicating the used electron-transporting layer materials indeedhave significant effect on device performance.

To elucidate the difference, two electron-only devices E1: ITO/BPhen: Li2CO3

(3%, 50 nm)/Li2CO3 (1 nm)/C70 (5 nm)/Al and E2: ITO/Bepp2: Li2CO3 (3%,50 nm)/Li2CO3 (1 nm)/C70 (5 nm)/Al are designed, where the electron injectionprocess from C70 to its adjacent electron-transport layer is simulated. Figure 3.27displays the J-V characteristics of devices E1 and E2. Under forward voltage, onlythe electrons could be injected from C70, whereas the holes are effectively blockedby the doped BPhen or Bepp2 layer. It can be seen that the current density is greatlylowered when BPhen is replaced by Bepp2, which is a strong evidence that a large


energy barrier is introduced. The result is further confirmed by capacitance–voltage(C–V) measurements, as shown in Fig. 3.27. The drastic decrease of the capacitancein device E1 happens at the voltage of about 0.2 V, which is caused by the effectiveinjection of electrons from C70 to doped BPhen layer [33]. This corresponds well tothe sharp increase of the current density. Therefore, the electron injection barrier indevice E1 is almost negligible. Contrarily, the capacitance in device E2 increasesslowly before the peak value of 1.4 V, which should be caused by the electronaccumulation at the interface between C70 and doped Bepp2, because the electronshave to overcome a larger barrier as injected. After that, the capacitance begins todecrease with voltage increase, accompanied by a noticeable decrease of currentdensity.

Fig. 3.26 Comparison of EL performance of devices A3 and A5. Device A3 is ITO)/MoO3

(3 nm)/TAPC:MoO3 (50 nm, 10%)/TAPC(10 nm)/TCTA:Ir(ppy)2 acac (4 nm, 8%)/BPhen:Ir(ppy)2acac (8 nm, 8%)/BPhen (10 nm)/BPhen:Li2CO3 (40 nm, 3%)/Li2CO3 (1 nm)/C70 (5 nm)/pentacene (10 nm)/Al, and device A5 is ITO)/MoO3 (3 nm)/TAPC:MoO3 (50 nm, 10%)/TAPC(10 nm)/TCTA:Ir(ppy)2 acac (4 nm, 8%)/BPhen:Ir(ppy)2acac (8 nm, 8%)/BPhen (10 nm)/Bepp2:Li2CO3 (40 nm, 3%)/Li2CO3 (1 nm)/C70 (5 nm)/pentacene (10 nm)/Al. Reprinted from [12]

Fig. 3.27 J-V and C-V characteristics of twoelectron-only devices E1:ITO/BPhen: Li2CO3 (3%,50 nm)/Li2CO3 (1 nm)/C70

(5 nm)/Al and E2:ITO/Bepp2: Li2CO3 (3%,50 nm)/Li2CO3 (1 nm)/C70

(5 nm)/Al. Reprinted from[12]


Generally, as the doped electron-transporting layer and the C70 layer contact, adepletion layer should be formed by charge transfer. Therefore, the thickness of thedepletion layer would greatly influence the electron injection efficiency. In order toconfirm the difference of the depletion layer in the case of BPhen: Li2CO3 andBepp2: Li2CO3, the capacitance–frequency (C–F) characteristics of devices E1 andE2 are measured, as shown in Fig. 3.28, which could be simply equaled to a parallelplate capacitor with the equation of

C0 ¼ ere0Ax

ð3:5Þ

where C0 is the capacitance at 0 V, er is the relative dielectric constant, e0 is thepermittivity of free space, A is the area of device, and w is the width of the depletionregion. A relative dielectric constant of 3.5 was used for the organic materials [34].Figure 3.28 shows the capacitance and the phase angle / as the function of fre-quency from 20 Hz to 2 MHz at 0 V in devices E1 and E2 with the phase angle

/ ¼ arctanXR

ð3:6Þ

where X and R are the imaginary part and the real part of the device impedance,respectively. Phase angle close to −90° corresponds to the capacitance character ofthe devices, which is governed by the charge depletion region. The increase of / atlow frequency is caused by the leakage current, whereas the increase at high fre-quency should be attributed to the device’s series resistor effect caused by highfrequency [35]. Therefore, the capacitances of 113 and 37 nF at 5 kHz and 500 Hzfor devices E1 and E2 are chosen to calculate the width of the depletion region,where the phase angles are mostly close to −90°. Thus, the widths of the depletionregion are estimated to be 4.4 and 13.4 nm by Eq. 3.5 for devices E1 and E2,respectively. It can be seen that Bepp2 causes wider depletion width than BPhen,which should be related to the device efficiency. It is well-known that Gaussian or

Fig. 3.28 C–F and phase angle-frequency characteristics of two electron-only devices E1:ITO/BPhen: Li2CO3 (3%, 50 nm)/Li2CO3 (1 nm)/C70 (5 nm)/Al and E2: ITO/Bepp2: Li2CO3 (3%,50 nm)/Li2CO3 (1 nm)/C70 (5 nm)/Al. Reprinted from [12]


exponential density of tail states (DOTS) dominates the interface energy levelbending at organic semiconductor interfaces [36]. Therefore, it is concluded that thedifference in the depletion region thickness caused by BPhen and Bepp2 should beoriginated from the distinct distribution of DOTS and the molecular order in thedoped Bphen and Bepp2.

According to the determined depletion widths, Fig. 3.29 depicts the difference ofelectron injection processes from C70 to the doped Bphen layer and the dopedBepp2 layer. It can be seen that the electrons on C70 have to overcome a muchthicker depletion region in device E2 than in device E1. On the other hand, theLUMO energy level of Bepp2 (−2.6 eV) is higher than that of BPhen (−3.0 eV),thus the energy barrier h2 that the electrons have to overcome is higher than h1.Therefore, the electrons on C70 are also more difficult to get across h2 in the case ofBepp2 than h1 in the case of BPhen. The comprehensive results are that the electroninjection is much more difficult from C70 into Bepp2: Li2CO3 than into BPhen:Li2CO3, thus leading to the low injection current density in device E2 and the lowpower efficiency in device A5. For the case of BPhen:Li2CO3 electron transportlayer, it is very clear that the thin depletion width and low injection barrier are veryfavorable in the electron tunnel, which is so efficient that the voltage drop on theinterface is negligible and thus guarantees the sufficient electron injection and thehigh power efficiency in device A3. Obviously, in order to realize further effectiveelectron injection, the electron-transporting layer adjacent to the organic hetero-junction charge injector has to carefully design and choice. Similarly, it is foundthat the thin MoO3 layer adjacent to pentacene also plays an important role inimproving the hole injection efficiency, which is very necessary to introduce it inthe fabrication of high performance OLEDs based on organic heterojunctions ascharge injectors.

Figure 3.30 gives the simple working processes of OLEDs based onpentacene/C70 organic heterojunction as both hole and electron injectors. As shown,electrons and holes accumulate on N-type C70 and P-type pentacene, respectively,

Fig. 3.29 Electron injection processes from C70 to its adjacent electron-transporting layer ofa BPhen: Li2CO3 and b Bepp2: Li2CO3. Reprinted from [12]


in the vicinity of the C70/pentacene interface. When an external electric field isapplied to the ITO and Al electrodes, the accumulated holes at the interface of theC70/pentacene are injected into the emissive layer across the hole-transportinglayers, and the accumulated electrons move toward the ITO. Simultaneously, theaccumulated electrons at the interface of C70/pentacene on the Al electrode side areinjected into the emissive layer across the electron-transporting layers, and theaccumulated holes move toward the Al. Then, the injected holes and electrons intothe emissive layer form excitons that subsequently emit light upon recombination.Clearly, the injected holes into the emissive layer result from the electron extractionfrom the pentacene HOMO through the C70 LUMO and then into the ITO, insteadof the hole transit from the ITO through the C70 HOMO, while the injected elec-trons into the emissive layer result from the hole extraction from the pentaceneHOMO through the C70 LUMO and then into the Al, instead of an electron transitfrom the Al through the pentacene LUMO. It can be seen that the injection mannerof electrons and holes in the devices based on organic heterojunction as the chargeinjectors is obviously different from that in conventional devices with metalinjectors, placing the charge injection far from the problematic electrode interfaces.More importantly, as proven above, the generated charges in organic heterojunc-tions are determined by the electric field on the heterojunction but are not related tothe work function of the used electrode metals. This means that the generated holeson the side of the anode should be approximately equal to the generated electronson the side of the cathode. As a result, a more balanced hole–electron recombi-nation is realized. The improved balance in the generated charge carriers alsoprevents the excess charges from accumulating, whereas the redistribution of theelectric field on organic heterojunctions also reduces the electric field intensity inthe emissive region in devices, greatly suppressing the quenching effect of localfield on the emissive excitons. All of these superior properties shown in OLEDswith organic heterojunctions as charge injectors guarantee a high efficiency and along lifetime of the fabricated OLEDs.

Fig. 3.30 Working mechanism of OLEDs based on pentacene/C70 organic heterojunction as bothhole and electron injectors. Reprinted from [12]


References

1. S. Braun, W.R. Salaneck, M. Fahlman, Adv. Mater. 21, 1450 (2009)2. F. So, D. Kondakov, Adv. Mater. 22, 3762 (2010)3. C. Carter, M. Brumbach, C. Donley, R.D. Hreha, S.R. Marder, B. Domercq, S. Yoo, B.

Kippelen, N.R. Armstrong, J. Phys. Chem. B 110, 25191 (2006)4. M. Kröer, S. Hamwi, J. Meyer, T. Dobbertin, T. Riedl, W. Kowalsky, H.-H. Johannes, Phys.

Rev. B 75, 235321 (2007)5. S. Hamwi, J. Meyer, M. Kröer, T. Winkler, M. Witte, T. Riedl, A. Kahn, W. Kowalsky, Adv.

Funct. Mater. 20, 1762 (2010)6. Y.H. Chen, D.G. Ma, J. Mater. Chem. 22, 18718 (2012)7. T. Ameri, N. Lia, C.J. Brabec, Energy Environ. Sci. 6, 2390 (2013)8. J.-H. Lee, J.W. Kim, S.-Y. Kim, S.-J. Yoo, J.-H. Lee, J.-J. Kim, Org. Electon. 13, 545 (2012)9. H.W. Mo, M.-F. Lo, Q.D. Yang, T.-W. Ng, C.S. Lee, Adv. Funct. Mater. 24, 5375 (2014)

10. W.S. Jeon, J.S. Park, L. Li, D.C. Lim, Y.H. Son, M.C. Suh, J.H. Kwon, Org. Electron. 13,939 (2012)

11. Y.H. Chen, D.G. Ma, H.D. Sun, J.S. Chen, Q.X. Guo, Q. Wang, Y.B. Zhao, Light Sci. Appl.5, e16042 (2016)

12. Q.X. Guo, H.D. Sun, D.Z. Yang, X.F. Qiao, J.S. Chen, T. Ahamad, S.M. Alshehri, D.G. Ma,Adv. Mater. Interfaces 3, 1600081 (2016)

13. H. You, Y.F. Dai, Z.Q. Zhang, D.G. Ma, J. Appl. Phys. 101, 026105 (2007)14. F.X. Wang, X.F. Qiao, T. Xiong, D.G. Ma, Org. Electron. 9, 985 (2008)15. J. Meyer, S. Hamwi, M. Kröger, W. Kowalsky, T. Riedl, A. Kahn, Adv. Mater. 24, 5408–

5427 (2012)16. S. Yoo, B. Domercq, B. Kippelen, Appl. Phys. Lett. 85, 5427 (2004)17. S. Reineke, F. Lindner, G. Schwartz, N. Seidler, K. Walzer, B. Lussem, K. Leo, Nature 459,

234 (2009)18. K. Walzer, B. Maennig, M. Pfeiffer, K. Leo, Chem. Rev. 107, 1233 (2009)19. Y.H. Chen, Q. Wang, J.S. Chen, D.G. Ma, D.H. Yan, L.X. Wang, Org. Electron. 13, 1121

(2012)20. Q.X. Guo, H.D. Sun, J.X. Wang, D.Z. Yang, J.S. Chen, D.G. Ma, J. Mater. Chem. C 4, 376

(2016)21. H.D. Sun, Q.X. Guo, D.Z. Yang, Y.H. Chen, J.S. Chen, D.G. Ma, ACS Photon. 2, 271 (2015)22. Y.H. Chen, J.S. Chen, D.G. Ma, D.H. Yan, L.X. Wang, F.R. Zhu, Appl. Phys. Lett. 98,

243309 (2011)23. H.D. Sun, Y.H. Chen, L.P. Zhu, Q.X. Guo, D.Z. Yang, J.S. Chen, D.G. Ma, Adv. Electron.

Mater. 1, 1500176 (2015)24. K.C. Kao, W. Hwang, Electrical Transport in Solids (Pergamon Press, Oxford, 1981)25. H.W. Mo, M.F. Lo, Q.D. Yang, T.W. Ng, C.S. Lee, Adv. Funct. Mater. 34, 5375 (2014)26. M. Pope, C.E. Swenberg, Electronic Processes in Organic Crystals (Oxford University Press,

New York, USA, 1982)27. K. Nomura, H. Ohta, A. Takagi, T. Kamiya, M. Hirano, H. Hosono, Nature 432, 488 (2004)28. J.S. Chen, C.S. Shi, Q. Fu, F.C. Zhao, Y. Hu, Y. Feng, D.G. Ma, J. Mater. Chem. 22, 5164

(2012)29. X.L. Wang, C.S. Shi, Q.X. Guo, Z.B. Wu, D.Z. Yang, X.F. Qiao, T.A. Ahamad, Saad M.

Alshehri, J.S. Chen, D.G. Ma, J. Mater. Chem. C 4, 8731 (2016)30. Y.H. Chen, H.K. Tian, Y.H. Geng, J.S. Chen, D.G. Ma, D.H. Yan, L.X. Wang, J. Mater.

Chem. 21, 15332 (2011)31. N.C. Giebink, G.P. Wiederrecht, M.R. Wasielewski, S.R. Forrest, Phys. Rev. B 82, 155305

(2010)

References 87

32. M. Koehler, I.A. Hümmelgen, Appl. Phys. Lett. 20, 3254 (1997)33. L. Zhang, H. Nakanotani, C. Adachi, Appl. Phys. Lett. 103, 093301 (2013)34. S. Lee, J.-H. Lee, J.-H. Lee, J.-J. Kim, Adv. Funct. Mater. 22, 855 (2012)35. P. Pahner, H. Kleemann, L. Burtone, M.L. Tietze, J. Fischer, K. Leo, B. Lüssem, Phys. Rev.

B 88, 195205 (2013)36. I. Lange, J.C. Blakesley, J. Frisch, A. Vollmer, N. Koch, D. Neher, Phys. Rev. Lett. 106,

216402 (2011)


Chapter 4Organic Semiconductor Heterojunctionsas Charge Generation Layer in TandemOrganic Light-Emitting Diodes

4.1 Basic Condition as Charge Generation Layer

Since tandem OLEDs can obtain the same luminance under several folds of lowercurrent density (depend on the number of EL units), its lifetime can be significantlylengthened. Furthermore, the tandem structure provides the feasibility that EL unitsof different colors can be vertically stacked together for color tuning and whiteemission [1]. Due to the unique advantages over traditional single-unit OLEDs,tandem OLEDs have been attracting extensive research interests ever since theirfirst appearance in 2003 [2].

A typical tandem OLED is fabricated by vertically connecting several individualelectroluminescent (EL) units together in series via interconnectors called chargegeneration layers (CGLs), with the entire device driven by a single power source, asshown in Fig. 4.1. When certain voltage is applied on the electrodes, each ELunit lights up individually under the same current that flows through the wholedevice. From the viewpoint of the simplest terms, the current efficiency and theexternal quantum efficiency of the tandem devices would be the sum of each ELunit, while the power efficiency might be inferior to the single EL unit ones,considering the extra voltage drop across the interconnector and the interfaces itbrings in. Obviously, the used CGLs play an important role, and the rationalinterconnector selection and design are crucial to the performance of tandemOLEDs. A good CGL must meet several requirements: efficient hole-electrongeneration, minimal energy barrier for charge injection, high conductivity, trans-parent in visible spectral range, high operational stability, and easy for deposition.In fact, the researchers are working hard toward this goal, and most publishedpapers on tandem OLEDs are focusing on interconnector design and its influenceon device performance. Various types of interconnectors have been put forward, ofwhich the most widely used structures include metal oxide/hole transport layer,doped-N/doped-P heterojunctions, and N/P organic heterojunctions [3].


89

Figure 4.2 shows an illustrated energy diagram of a tandem OLED with twoemissive units under a forward bias. In this operational mode, holes are injectedfrom the anode, and electrons are injected from the cathode. Meanwhile, electronsand holes are generated at the interface of two layers in CGL. As a result, the

Fig. 4.1 Structure of tandem OLEDs with three emissive units

Fig. 4.2 Schematic energy band diagram of a tandem OLED with two EL units under forwardbias voltage. The energy diagram is based on the assumption that all of the vacuum levels arealigned (Schottky–Mott model)

90 4 Organic Semiconductor Heterojunctions as Charge Generation …

generated electrons in CGL are injected and transferred to the lowest unoccupiedmolecular orbit (LUMO) of the emissive layer (EML) in EL unit 1 and recombinewith the externally injected holes at the highest occupied molecular orbit (HOMO)of the EML to emit light. At the same time, the generated holes at CGL are injectedand transferred to the HOMO of the EML in EL unit 2 and recombine with theexternally injected electrons at the LUMO of the EML to emit light. It can be seenthat the charge generation and charge injection are two important processes thathave to consider as designing an efficient CGL. Generally, the CGLs include twolayers consisted of a strong electron acceptor and an electron donor, and the strongelectron acceptor shows a lower LUMO level, thus an effective charge transfer fromdonor to acceptor occurs, and finally generating electrons and holes. This meansthat the energy level difference between the LUMO of acceptor and the HOMO ofdonor should be as small as possible to guarantee the efficient charge transfer,which is also a key for the design of efficient CGLs. Additionally, the energydifference between the LUMO level of the ETL in the first EL unit and the HOMOlevel of the HTL in the second EL unit should also as small as possible, thusgenerated electrons and holes can effectively inject into respective EL unit withoutany resistance.

Since the used electron acceptors in CGLs have rather deep LUMOs, which isnear, even lower than the HOMOs of most common hole transporting materials(HTMs), the charge generation and the hole injection are relatively easy. However,the electron transporting materials (ETMs) in OLEDs usually have a LUMO levelaround 3.0 eV; therefore, the electron injection from the LUMO of electronacceptor layer into the adjacent ETL is hard unless the vacuum-level bending isinvolved. So the design of proper electron acceptor layer/ETL interface is critical tothe performance of tandem OLEDs. At present, N-doped ETL is a commonly usedsolution to increase the band bending at the interface, such as doping Li, Cs, andtheir compounds of Liq, Li2CO3, and Cs2CO3 in ETL [3]. Another popular methodto increase the band bending at the interface is adding a thin layer of metal (about1 nm) Al or Ag in between [4]. It can be clearly seen that the addition of a thin layermetal can improve the device performance significantly, especially in power effi-ciency, because the voltage drop at the electron injection interface is reduced. Themechanism of enhancing the charge generation by inserted metal is thought to bethat the metal cluster can induce more gap states to assist charge tunneling [5].

The first CGL that was used to fabricate tandem OLEDs is ITO and V2O5 asinterconnector between two EL units [2]. Ever since, various interconnectorsemploying transition metal oxides (TMOs) have been reported, of which MoO3 andWO3 are frequently used [6, 7]. The typical TMO-type interconnector structure is toinsert a thin layer of TMO between the HTL of one sub-EL unit and the ETL ofanother sub-EL unit. Now it is generally considered that the charges are generatedat the interface of TMO and adjacent HTM [8]. For a typical TMO/HTM system,the TMO serves as a strong electron acceptor, and when the electrons are drawnfrom the HOMO of HTM to the LUMO of TMO, the holes would be left at theHOMO side of HTM, thus electrons and holes are generated in pairs. Afterward, thegenerated electrons and holes would separate and transport in opposite directions

4.1 Basic Condition as Charge Generation Layer 91

under certain bias, and finally are injected to adjacent EL units, respectively. Forthis case of TMOs as CGL, there yet exists degradation at TMO/ETL interface,which has attributed to the electric-field-induced migration of metal ions towardETL/TMO interface when thermally evaporating TMOs on ETL due to hightemperature.

Subsequently, different CGLs with further efficient charge generation propertyare successfully developed, such as doped-N/doped-P heterojunctions, N/P bilayerheterojunctions, N:P bulk heterojunctions, and N/N:P/P composited heterojunctions[9–12]. The following are introduced separately.

4.2 Doped-N/Doped-P Heterojunction as ChargeGeneration Layer for High-Efficiency TandemOrganic Light-Emitting Diodes

The first CGL employing a doped-N/doped-P organic heterojunction to fabricatetandem OLEDs was reported in 2004 by Liao et al. [9]. They used Li:Alq/FeCl3:NPB as CGL to achieve a tandem OLED with current efficiency of 81.7 cd/A fortwo-unit device and 136.3 cd/A for three-unit device using Ir(ppy)3 as emitter,which are about two and three times higher than single-unit one, proving thevalidity of the doped-N/doped-P organic heterojunction as CGL. This type of CGLis considered to offer several advantages over the TMOs, including excellent opticaland electrical properties, as well as the ease of fabrication by thermal evaporation.In the CGL structures, the used P- and N-type dopants play an important role incharge generation. Among, some electron-withdrawing materials with low LUMOlevel such as 2,3,5,6-tetrafluoro-7,7,8,8-tetracyanoquinodimethane (F4-TCNQ),1,4,5,8,9,11-hexaazatriphenylenehexacarbonitrile (HAT-CN), MoO3, and ReO3 arealways used as the P-type dopants, and some alkali metals and their compounds areused as the N-type dopants [3]. Similar to the TMO CGLs, the performance ofdoped-N/doped-P organic heterojunction CGLs is closely related to the energy levelalignment at the junction interface. Due to the preserve of a large space chargedensity in the doped systems, a significant energy shift at the interface can arise,which is essential to the charge generation in doped-N/doped-P organic hetero-junction CGLs.

A model for the charge generation in doped-N/doped-P organic heterojunctionCGLs was presented as direct tunneling of charges from the HOMO states of thedoped-P layer to the LUMO states of the doped-N layer [13], which has been provenby conducting current-voltage (I-V) test and Kelvin probe (KP) measurements in aseries of devices with different interface structures (Fig. 4.3b). The forward profile ofthe I-V plot is considered to be a result of recombination by the injected charges,while the reverse profile can be understood as the result of charge generation insidethe devices. The charge injection of the devices is very efficient due to the Ohmiccontact, so the symmetry of the I-V characteristics of CGL structures can be taken as


an indicator for the efficiency of charge generation at the doped-N/doped-P interface.It is found that both N-doping and P-doping are important for the generationof charges, and only the interface with N-doping/P-doping structure shows thebest symmetry, thus efficient charge generation, as shown in Fig. 4.3a. Thelow-temperature I-V characteristics of the devices with doped-N/doped-P interface,both forward and reverse bias I-V curves, are found to be independent of temperature.Since the forward charge injection of such Ohmic contact device is usually inter-preted as Fowler–Nordheim tunneling, it is natural to assume that the charge gen-eration under reverse bias should also be tunneling. The assumption is supported bythe Kelvin probe measurements which show a rather thin charge depletion layer atthe doped-N/doped-P interface, as well as very close to the P-doped material HOMOand N-doped material LUMO level alignment due to the vacuum-level bending,suggesting that the electron tunneling would be very easy in this case, As shown inFig. 4.4b. In contrast, the undoped interface exhibits rather thick charge depletionlayer (shown in Fig. 4.4a), which will make the charge generation via tunnelingdifficult, therefore the necessity of doping is well explained. A schematic diagram

Fig. 4.3 a I-V characteristics of devices with different interface structures. b Schematic diagramsof devices with different interface structures. Reprinted from [13]

4.2 Doped-N/Doped-P Heterojunction as Charge Generation Layer … 93

given in Fig. 4.5 depicts the energy level alignment and compares the involvedphysical processes for forward and reverse bias.

From the above, it can been seen that the charge generation efficiency of thedoped-N/doped-P organic heterojunction CGLs mainly depends on the energy levelalignment at the junction interface; therefore, sometimes the interface modificationssuch as a thin metal layer [14] or a HAT-CN layer [15] are employed to increase thevacuum-level shifting, so as to improve the device performance.

Figure 4.6a shows the current density–luminance–voltage (J-L-V) curves of thefabricated green tandem OLEDs with a thin metal interlayer in CGL [14], where theCGL structures are PEGDE:Al (1:2, 4.5 nm)/none, Ag (1 nm), or Au (1 nm)/NPB:F4-TCNQ (10:1, 15 nm) named as CGL, Ag in CGL, and Au in CGL in figure,respectively. It is clearly seen that the tandem OLEDs with an additional Ag or Auinterlayer in the CGL present the relatively higher current density and brightness atthe low bias. The turn-on voltage largely reduces from 9.1 V for the tandem deviceapplying CGL without the metal interlayer to 6.6 V for tandem device with Ag orAu interlayer in CGL, suggesting that the metal interlayer effectively improves theelectrical properties at the junction interface in CGL as well as the device perfor-mance. The tandem OLEDs with Ag and Au interlayer in CGL, respectively, givethe current efficiency of 47.1 cd/A and 51.4 cd/A, and the power efficiency of 8.9lm/W and 10.5 lm/W at 10 mA/cm2 current density, which are much higher than37.2 cd/A and 6.08 lm/W of the tandem device without the metal interlayer inCGL, as shown in Fig. 4.6b. Figure 4.7 depicts the mechanism for the transport orgeneration of charge carriers in CGLs with and without metal interlayer under theelectrical bias. The transport of charge carriers has to go through a large energybarrier in the junction interface at PEGDE:Al (N-CGL)/NPB:F4-TCNQ (P-CGL)layer, causing the obvious current delay and high turn-on voltages in tandemOLEDs (Fig. 4.7a). However, while a thin Ag or Au interlayer (1 nm) is introducedin CGL (in Fig. 4.7b), the transport of charge carriers could go through the surfacestates in Ag or Au interlayer. The energy levels of those surface states are located at

Fig. 4.4 Work function obtained from KP measurements as a function of the film thickness,relative to the ITO substrate’s work function for a an undoped-P/N heterojunction (inset:progression of the work function for a thicker TPBi film) and b a doped-P/N heterojunction.Reprinted from [13]


the middle of the junction barrier, functionalizing well as the mediums to theinjection or transport of electrons and holes to CGL layer. Additionally, thecluster-like morphology of the 1-nm metal interlayer exhibits the large surface area,which provides a high density of generation sites for charge carriers as induced bythe applied electrical bias. Accordingly, the accumulation of charge carriers in CGLis markedly suppressed by Ag or Au interlayer in CGL, thus reducing the biasvoltage and enhancing the efficiency of the resulting tandem OLEDs.

The utilization of HAT-CN with low LUMO level in CGLs can further reduceoperation voltage, enhance power efficiency, and improve stability of the resultingtandem OLEDs [16]. The deep mechanism investigation demonstrates [15] that thecharge generation is yet a tunneling process at the interface between HAT-CN anddoped-P layer, which is quite similar to that in the doped-N/doped-P heterojunction.However, there is a subtle difference between them: In the case of the doped-N/doped-P heterojunction CGLs, the electron tunneling occurs from the highestoccupied molecular orbital (HOMO) of doped-P layer to the LUMO of doped-N layer, which is the rate-limiting step of the charge-generation process, but in the

Fig. 4.5 Energy level alignment for a doped-N/doped-P organic heterojunction CGL-baseddevice: a no external bias, b under forward bias (recombination), and c under reverse bias (chargegeneration). Reprinted from [13]

4.2 Doped-N/Doped-P Heterojunction as Charge Generation Layer … 95

CGLs consisting of doped-N/HAT-CN/doped-P, the charge carriers are easilygenerated at the junction of the doped-P layer and HAT-CN, then the electrontunneling occurs from the LUMO of HAT-CN to the LUMO of doped-N layer (N/Njunction), which is the rate-limiting step of the charge-generation process. Thedifference in the charge generation in the two cases is schematically illustrated inFig. 4.8a, b. The tunneling in the N/N junction can be very easy due to the lowbarrier height compared to the tunneling in the N/P junction. Because the electroninjection at the HAT-CN/doped-N junction limits the charge generation in theCGLs rather than the charge generation itself at the doped-P/HAT-CN junction, thecharge carrier generation in the CGLs depends on the vacuum-level shift at theHAT-CN/doped-N and the free carrier density of the doped-N layer rather than theLUMO energy level of the doped-N layer. This means that the choice ofelectron-transport materials in the doped-N layer is very important to form out-standing charge generation efficiency in CGLs.

Fig. 4.6 a J-L-V and b LE-J curves of tandem OLEDs based on CGLs without and with a thin Agor Au interlayer (mark as Ag in CGL and Au in CGL). The efficiency property of a single-unitOLED is also given in (b). Reprinted from [14]

Fig. 4.7 Mechanisms for the transport or generation of charge carriers in CGLs under theelectrical bias. a CGL without the metal interlayer. There is a large energy barrier in PEGDE:Al/NPB:F4-TCNQ junction interface. b CGL with Ag or Au interlayer. The transport or generationof charge carriers could go through the surface states in Ag or Au interlayer. Reprinted from [14]


4.3 N/P Bilayer Heterojunction as Charge GenerationLayer for High-Efficiency Tandem OrganicLight-Emitting Diodes

Due to some drawbacks of TMO and doped-N/doped-P heterojunction CGLsmentioned above, and the power efficiencies of the resulting tandem OLEDs basedon them are far from satisfactory, therefore, the design of other CGLs is verynecessary. So far, organic heterojunctions as CGLs seem to be the best choice forthe architecting of high power efficiency tandem OLEDs because of the advantagesof organic heterojunctions as CGLs including non-doping, high optical transmis-sion, and efficient charge generation, thus resulting in easy for deposition, moregood stability, and good electrical properties [17].

The first report about organic heterojunctions as CGLs was given by Lai et al. in2007, and they used copper hexadecafluorophthalocyanine (F16CuPc)/copperphthalocyanine (CuPc) organic heterojunction as CGL to successfully achieve atandem OLED with double current efficiency of the single-unit device [18].However, the operating voltage of the device increased significantly (more thantriple times) due to the large injection barrier at F16CuPc and ETL interface,resulting in the power efficiency unsatisfied. Similar to other CGLs, the organicheterojunction CGLs also need effective vacuum-level bending at the organicheterojunction CGL/ETL interface to insure good electron injection. In 2008, Liaoet al. reported high efficiency tandem OLEDs based on HAT-CN/NPB hetero-junction as CGL [19]. They employed an N-doped Li:BPhen as ETL which canresult in a large vacuum-level bending at the ETL/HAT-CN interface to solve theelectron injection problem. The tandem devices not only showed a double currentefficiency of the referencing single-unit device, but also exhibited higher powerefficiency, indicating the great potential of organic heterojunctions as CGLs.

Fig. 4.8 Energy level alignment diagrams of a doped-TAPC/HAT-CN/doped BPhen CGL andb doped-TAPC/doped- BPhen CGL. Reprinted from [15]

4.3 N/P Bilayer Heterojunction as Charge Generation Layer for High Efficiency … 97

Organic heterojunctions as CGLs usually contain two organic layers deposited insequence to form donor–acceptor system, and the charges are generated at theinterface under bias. The acceptor is usually a N-type organic material with bigelectron affinity, thus low LUMO. The choice of such organic semiconductorssuitable for deposition is restricted. The commonly reported ones include F4TCNQ,C60, C70, and HAT-CN. The donor is usually a P-type material with good holetransport property and proper HOMO level. Because the dielectric constant of theorganic semiconductor is usually low and the non-covalent electronic interactionsbetween organic semiconductors are weak compared to inorganic semiconductors,two types of anisotype organic heterojunctions may be formed: accumulation anddepletion organic heterojunctions. Considering the energy levels of acceptor anddonor used in organic heterojunctions, such as C60/pentacene and HAT-CN/m-MTDATA, the Fermi level of the intrinsic P-type materials is higher than that of theN-type those, so when they contact, the electrons would flow from P-side to N-sideto form an accumulation-type organic heterojunction [20]. It has been experimen-tally demonstrated that the formation of accumulation-type organic heterojunctionsis important for the performance improvement of resulting tandem OLEDs.Different from the depletion junction in typical doped-N/doped-P organic hetero-junction CGLs, where the depletion of free carriers will cause lower mobility, theaccumulation-type organic heterojunctions will accumulate free holes at the P-sideand electrons at the N-side, thus exhibit rather high mobility (several magnitudeshigher than intrinsic layers) at both sides. One significant advantage of such organicheterojunctions is that the accumulation of free carriers ensures the efficient chargegeneration, meanwhile, the bulk voltage drop across the organic heterojunctionscould be reduced to a minimum because of the high conductivity. Due to theseadvantages, it is reasonable to expect the better performances from the tandemOLEDs utilizing organic heterojunctions as CGLs, especially in terms of powerefficiency improvements are more prominent [10].

It is well-known that semiconductors are defined by their unique electric con-ductivity behaviors and can be classified into P-type and N-type semiconductors.A semiconductor heterojunction is the interface that occurs between P-type andN-type semiconductors, and it is always advantageous to engineer the electronicenergy bands in many solid state devices, including semiconductor light-emittingdiodes, lasers, solar cells, and transistors. In fact, the concept of semiconductorheterojunctions had already been proposed and the energy band profiles follow theAnderson model [21]. To date, all inorganic optoelectronic devices are based onthis kind of semiconductor heterojunction. The most familiar heterojunction type ininorganic semiconductors is the depletion mode. In this type of heterojunction, adepletion junction is formed on either side of the heterojunction interface, namelythe positive charges are accumulated on the N-type side and the negative chargesare accumulated on the P-type side in the depletion region. In this case, the internalelectric field is opposite to the external field, and the charges in the depletion regionare immovable. However, the case of a heterojunction consisting of two organicsemiconductors is somewhat different. The organic semiconductor heterojunctions(i.e., N-type and P-type organic semiconductors) is a promising electronic system


for charge recombination in OLEDs, for charge separation in organic photovoltaiccells (OPVs), and for charge accumulation in organic field-effect transistors(OFETs) owing to the energy mismatch between the frontier orbital of the twoorganic semiconductors. Because the dielectric constant of the organic semicon-ductors is usually low and the non-covalent electronic interaction between organicsemiconductors is weak compared to inorganic semiconductors, two kinds ofaccumulation and depletion organic heterojunctions are formed. The depletionheterojunction in organic semiconductors is similar to that in inorganic semicon-ductors. However, the accumulation of heterojunction is completely different. Inthis case, the positive and negative charges are accumulated on the P-type andN-type sides, respectively, of the organic semiconductors to form the space chargeregion. We call this phenomenon a heterojunction effect, which has been welldemonstrated in OFETs and OPVs exhibiting highly efficient device performance.The direction of the built-in voltage is from the P-type region to the N-type region.More importantly, the accumulated charges within the space charge region aremovable. The accumulation of high-density free charge carriers results in a highconductivity along the junction direction.

To be able to effectively achieve and manipulate these processes in organicheterojunctions, C60/pentacene organic heterojunction is selected as an example toelucidate these processes. On the basis of thermal emission theory, the electrontransfer from pentacene to C60 can be achieved since pentacene has a higher Fermilevel than C60 in the flat band state (Fig. 4.9 left). Also noted is that this chargetransfer in turn contributes to the interfacial energy level equilibrium. Benefitingfrom the charge redistribution, the electrons and holes can be accumulated onN-type C60 and P-type pentacene, respectively, in the vicinity of the C60/pentaceneinterface (Fig. 4.9 middle). Therefore, high-density free electrons and holes areprovided at the C60/pentacene junction (Fig. 4.9 right), i.e., charge generationoccurs. These generated charge carriers can move away from the interface inopposite direction under an external electric field. This process is beneficial toreducing the voltage drop across the CGL and hence the reduction of the overalldriving voltage during the device operation. Obviously, the relative energy level ofboth semiconductor components is very important for the CGL construction, whichdirectly determines the charge generation.

Fig. 4.9 Proposed working principle of C60/pentacene organic heterojunction CGL. EF: Fermienergy level; LUMO: lowest unoccupied molecular orbital; HOMO: highest occupied molecularorbital. Reprinted from [10]


Fig. 4.10 depicts the schematic diagrams of tandem OLEDs based on N/Pbilayer organic heterojunctions as CGLs. In the organic heterojunction CGLs, C60

is chosen as the N-type layer and either pentacene, ZnPc, CuPc, or H2Pc is used asP-type layer. In order to effectively extract and inject electrons and holes from CGLinto the respective EL unit, thin LiF and MoO3 interfacial layers are introduced onboth sides of CGL. For comparison, the diagram of corresponding single-unitOLEDs is also given in Fig. 4.10 [10].

Figure 4.11 compares the EL characteristics of C60/pentacene bilayer organicheterojunction-based tandem OLEDs and corresponding single-unit OLEDs.Different from the conventional CGLs-based tandem OLEDs, where the operationalvoltage of tandem devices is two times that of single-unit devices, even though theoperational voltage of the resulting tandem OLEDs is higher than that of thesingle-unit OLEDs (Fig. 4.11a), but it is less than two times and graduallydecreases with increasing current densities and luminance. For example, at theluminance of 1000 cd/cm2, the operational voltage is reduced from 10.1 V ofsingle-unit devices to 7.2 V of tandem devices. This demonstrates that C60/pen-tacene organic heterojunction exhibits remarkable charge generation characteristicsas a CGL. As a result, not only is the maximum current efficiency greatly enhancedfrom 15.2 cd/A of single-unit devices to 38 cd/A of tandem devices (Fig. 4.11b),but is the maximum power efficiency also significantly improved 21.9 lm/W oftandem devices from 10.1 lm/W of single-unit devices (Fig. 4.11c). Even at highluminance, the power efficiency enhancement is still significant; for example, under1,000 cd/m2, 10,000 cd/m2, and 38,000 cd/m2, the power efficiency enhancementsare 1.74, 1.81, and 1.9 times that of the optimized single-unit devices, respectively.This undoubtedly gives the best power efficiency improvement reported so far fortandem OLEDs. Actually, when the pentacene is replaced by H2Pc, ZnPc, or CuPc,similar power efficiency improvements can also be achieved besides doubling thecurrent efficiency. Table 4.1 summarizes the detailed EL performances of all tan-dem OLEDs and single-units OLEDs. This demonstrates that organic

Fig. 4.10 Schematic diagrams of single-unit (left) and tandem (right) OLEDs. The molecularstructures of C60, pentacene, ZnPc, CuPc, and H2Pc are shown at the top-left. Reprinted from [10]


heterojunctions as CGLs are a universal concept for the improvement in the powerefficiency of resulting tandem OLEDs.

The improvement of power efficiency in the C60/pentacene heterojunction-basedtandem devices can be evidently provided by investigating the capacitance–voltage(C-V) characteristics of the designed capacitance devices with and without C60/

Fig. 4.11 EL performancesof single-unit OLEDs andtandem OLEDs based on C60/pentacene organicheterojunction as CGL,a current density–luminance–voltage, b currentefficiency-current density, andc power efficiency-luminancecharacteristics. The powerefficiencies of tandem OLEDswith different materialcombinations of CGLs aregiven in inset. Reprintedfrom [10]


pentacene heterojunction. As shown in Fig. 4.12, device 1 without heterojunctionexhibits no change in capacitance with the applied voltages from −20 to 20 V. Thisconstant capacitance indicates that the thick LiF film indeed acts as an effectiveinsulator and completely blocks the charge injection from ITO and Al externalelectrodes in this voltage range. Furthermore, it also indicates that neither the

Table 4.1 Summary of EL performances of single-unit OLEDs and tandem OLEDs with differentorganic heterojunctions as CGLs. Reprinted from [10]

CGL Vt

(V)L (cd/m2) ηCE,max

(cd/A)ηCE(cd/A)

ηPE,max

(lm/W)ηPE(lm/W)

no 2.5 3033 14.9 14.8 10.1 5.1

C60/pentacene 4.9 7322 37.8 36.6 21.9 7.9

C60/ZnPc 4.9 7022 36.1 35.8 21 6.8

C60/CuPc 5.1 6547 33.4 33.4 18.6 6.7

C60/H2Pc 5.3 6527 33.5 33.7 16 5.8

Li:BCP/MoO3 5.5 6131 31.7 30.7 11.6 4.7

Li:BCP/MoO3:NPB

5.5 5998 31 30.4 11.3 4.7

Vt turn-on voltage examined at the luminance of 1 cd/m2. L luminance, ηCE current efficiency, andηPE power efficiency obtained at the current density of 20 mA/cm2. ηCE,max maximum value ofcurrent efficiency. ηPE,max maximum value of power efficiency. No single-unit device

Fig. 4.12 C-V characteristics of different capacitance devices measured at a fixed frequency of1 kHz. The device structures are ITO/LiF(100 nm)/Alq3(30 nm)/NPB(50 nm)/LiF(100 nm)/Al(120 nm) (device 1), ITO/LiF(100 nm)/Alq3(30 nm)/C60(20 nm)/pentacene(15 nm)/NPB(50 nm)/LiF(100 nm)/Al(120 nm) (device 2), ITO/LiF(100 nm)/Alq3(30 nm)/LiF(0.3 nm)/C60(20 nm)/pentacene(15 nm)/MoO3(3 nm)/NPB(50 nm)/LiF(100 nm)/Al(120 nm) (device 3), ITO/LiF(100 nm)/Alq3(30 nm)/Li:BCP (20 nm)/MoO3(3 nm)/NPB(50 nm)/LiF(100 nm)/Al(120 nm) (de-vice 4), and ITO/LiF(100 nm)/Alq3(30 nm)/Li:BCP (20 nm)/MoO3:NPB (40 nm)/NPB(50 nm)/LiF(100 nm)/Al(120 nm) (device 5). Reprinted from [10]


displacement nor the generation of charges within the interface between NPB andAlq3 occurs. Differently, device 2 and device 3 with heterojunction exhibit agradual increase of capacitance with voltage, and the turn-on voltage is reduced indevice 3 due to the introduction of LiF and MoO3 interfacial layers on the bothsides of heterojunction. This indicates that an electric-field-induced charge gener-ation process takes place at the C60/pentacene interface well, and the used LiF andMoO3 interfacial layers greatly extract the charges within heterojunction byreducing injection barriers. Apparently, the power efficiency improvement shouldcome from the efficient and excellent collaboration of charge generation, transportand extraction that occur in the C60/pentacene organic heterojunction. It can be seenthat the buffer-modified C60/pentacene organic heterojunction as CGL is alsosuperior to both the traditional N-doped organic/transition metal oxide and doped-N/doped-P organic junction CGLs. As shown in Fig. 4.12, organic heterojunctionslead to a large capacitance at lower voltage with respect to conventional CGLs,indicating that organic heterojunctions as CGLs generate much more charges at lowelectrical field, thus greatly enhancing the power efficiency of the resulting tandemOLEDs.

The interfacial electronic structures of MoO3 and LiF-modified pentacene/C60

organic heterojunction CGL have been determined well by photoemission spec-troscopy [22]. Figure 4.13 gives the total schematic energy level alignment dia-gram. It is found that a small energy offset at the pentacene/C60 heterojunctionmakes it easy to transfer electrons from pentacene to C60 even under a small appliedbias, facilitating the occurrence of charge generation. The band bending observed inboth pentacene and C60 is beneficial to exciton-dissociation and charge transport inopposite directions. At the MoO3/pentacene interface, the high work function (WF)of MoO3 brings the HOMO onset up to the Fermi level (EF) not only for pentacene

Fig. 4.13 Schematic energy level alignment at the NPB/MoO3/pentacene/C60/LiF/Alq3 interfacesand proposed mechanisms of holes and electrons generation, where the dotted lines denote theenergy levels of HTL and ETL in two adjacent EL units, and the solid dots denote electrons andthe open dots denote holes. The data for VL, LUMO and HOMO denote the values of energy levelwith respect to the EF. Reprinted from [22]


but also for most HTL materials of the adjacent EL unit as this CGL is connectedinto tandem OLEDs. Therefore, holes can be efficiently injected from pentaceneinto this EL unit. Similarly, at the C60/LiF interface, the low WF of the LiF bufferlayer makes the LUMO to pin close to the EF not only for C60 but also for mostETL materials of the other EL unit, which induces the electrons to inject easily fromC60 into that EL unit by tunneling through the thin LiF film. The favorable energylevel alignment can effectively enhance charge generation, transport, and injection.The advantage of MoO3/pentacene/C60/LiF structure is that thus formed CGL cangreatly reduce the voltage drop and thus enhance the power efficiency (PE) of thecorresponding tandem OLEDs.

However, it is found that the power efficiency of the fabricated tandem OLEDscannot be improved when using NPB to replace pentacene in pentacene/C60 organicheterojunction CGL. Figure 4.14 shows the efficiency properties of ndem OLEDsbased on NPB/C60 organic heterojunction as CGL. Clearly, the current efficiency isdoubled, but the power efficiency is reduced with respect to that of single-unitOLEDs (see Fig. 4.11). This means that not all organic heterojunctions can be usedto be as CGLs to fabricate high power efficiency tandem OLEDs. Unlikepentacene/C60 organic heterojunction, as shown in Fig. 4.15, NPB has a lower EF

than that of C60, therefore, as they contact, a depletion junction is formed, where theelectrons are depleted on the side of NPB and the holes are depleted on the side ofC60, forming a reverse built-in field and a different band bending compared topentacene/C60 junction. This will greatly block the effective transport of charges,thus leading to a higher operational voltage drop on this junction. This also furtherdemonstrates that the formation of an accumulation junction in the organicheterojunction-type CGLs is very necessary in order to obtain the improvement inpower efficiency of the resulting tandem OLEDs.

The effects of energy level and mobility of the used organic materials in organicheterojunctions on charge generation and thus device performance have been wellstudied [23]. Here, a homologous series of P-type thiophene organic

Fig. 4.14 Efficiencyproperties of the fabricatedtandem OLEDs based onNPB/C60 organicheterojunction as CGL.Reprinted from [10]


semiconductors NaTn (naphthyl end-capped oligothiophenes, n = 2–6 representsthe number of thiophene units) were used to form organic heterojunctions withN-type organic semiconductor C60 (C60/NaTn). Their molecular structures areshown in the inset of Fig. 4.16 right side. The NaTn organic semiconductorspossess significant differences in HOMO levels (Fig. 4.16 left) and hole mobilities(Fig. 4.16 right) [24], which make them suitable for investigating our hypothesis.Five green tandem OLEDs with NPB/Alq3:C545T/Alq3 active layers based on C60/NaTn organic heterojunction CGLs were fabricated for comparison.

Table 4.2 summarizes the detailed EL performance of all five tandem OLEDsand a single-unit OLED. It can be seen that the current efficiency of all tandemdevices is over 2 times that of the single-unit device at a given current density,demonstrating the C60/NaTn organic heterojunctions functioned as an effectivebipolar CGL. More importantly, the power efficiency of the tandem devices is alsoremarkably enhanced. The maximum power efficiencies from C60/NaT2- to C60/NaT6-based tandem devices are 17.6, 18.2, 21.5, 18.4, and 20.5 lm/W, respectively,which are nearly 1.7–2 times that of the single-unit OLED (10.1 lm/W). Thisconfirms again that the design concept of organic heterojunctions as CGLs is apredominant technique for achieving high power efficiency in tandem OLEDs.

Fig. 4.15 Energy level and space charge region diagrams of NPB/C60 organic heterojunction.Reprinted from [10]

Fig. 4.16 Energy levels relative to C60 (left) and hole mobility of NaTn (right). The molecularstructure of NaTn, n = 2–6 is shown in inset of right. Reprinted from [23]


However, the five tandem devices show a marked difference in luminance andefficiency at the same current density. As shown, NaT4 exhibits the highest effi-ciency, followed by NaT6, NaT5, NaT3, and NaT2.

This means that the energy levels and the mobility of the used P- and N-typeorganic materials in heterojunction CGLs have strong effects on the performance ofthe fabricated tandem OLEDs, which are directly related to the processes of chargegeneration and charge transport. Since the generated charges stem from the electrontransfer from the Fermi level of NaTn to the Fermi level of C60, the position of theHOMOs for NaTn (ranging from 5.43 eV for NaT2 to 5.12 eV for NaT6

(Fig. 4.16)) relative to the LUMO of C60 probably contributes to the differentcapabilities of charge generation in C60/NaTn CGLs. To further elucidate thisproblem, the capacitance devices based on C60/NaTn were fabricated, and theirC-V characteristics were measured, which are shown in Fig. 4.17a. It is clearly seenthat the capacitance is gradually decreased with increasing energy level differencebetween the HOMOs of NaTn and the LUMO of C60 in the saturation regime, asshown in Fig. 4.17b. Because the magnitude of the capacitance can qualitativelygive the charge generation capability, the variation regulation of the capacitancewith the energy levels between the two organic semiconductors used means that theHOMO level of P-type organic molecules close to the LUMO of N-type organicmolecules will generate more charges in the CGLs. Correspondingly, a high ELefficiency will then be obtained. As shown in Fig. 4.17c, the maximum currentefficiency and power efficiency does gradually increase from NaT2 to NaT6 by theenergy level regulation, except for the case of NaT4.

Because the NaT4 possesses the highest hole mobility (0.39 cm2/Vs, Fig. 4.17d)with respect to the others, the achievement of the highest efficiency in tandem OLEDswith NaT4-based CGL indicates that the highmobility of organic semiconductors usedin organic heterojunction CGLs is even more important. The high mobility willrapidly transport the generated charges toward the interfaces adjacent to the EL unitsunder the external electric field, thus facilitating the reduction of operational voltages.As shown in Fig. 4.17d, the current efficiency and power efficiency of the fabricated

Table 4.2 Summary of the EL characteristics of the five TOLEDs based on C60/NaTn organicheterojunction as CGLs and single-unit OLED

CGL Vt

(V)L(cd/m2)

ηcd,Max

(cd/A)ηcd(cd/A)

ηp,Max

(lm/W)ηp(lm/W)

ELpeak

(nm)

C60/NaT2 4.9 6620 33.2 33.1 17.6 6.4 524

C60/NaT3 4.9 7000 35 34.8 18.2 6.7 524

C60/NaT4 4.9 7820 39.1 39.1 21.5 7.8 524

C60/NaT5 4.9 7300 37 36.5 18.4 6.9 524

C60/NaT6 4.9 7440 37.9 37.2 20.5 7.3 524

Single-unit 3.3 3133 14.9 14.8 10.1 5.1 524

Turn-on voltage Vt is tested at 1 cd/m2 (Vt); luminance (L), current efficiency (ηcd), powerefficiency (ηp), and peak of EL spectra (ELpeak) are tested at 20 mA/cm2; maximum currentefficiency (ηcd,Max); maximum power efficiency (ηp,Max)


tandem OLEDs based on NaTn/C60 organic heterojunctions as the CGL indeedincrease with the hole mobility of NaTn from NaT2, NaT3, NaT5, NaT6, to NaT4.Therefore, the organic semiconductors used to construct high-performance organicheterojunction CGLs not only need to have proper energy levels, i.e., closer levelsbetween the HOMO of P-type organic semiconductor and the LUMO of N-typeorganic semiconductor, but also exhibit higher mobility. The results offer adesign/selection rule for organic semiconductors used to construct effective organicheterojunction CGLs, which is useful to fabricate high-performance tandem OLEDs.

To elucidate the mechanism of charge generation in organic heterojunctions,using pentacene/C70 as an example, the following device ITO/MoO3(3 nm)/30 wt%MoO3:TAPC (50 nm)/pentacene (30 nm)/C70 (30 nm)/Li2CO3 (1 nm)/3 wt%Li2CO3:BPhen (45 nm)/Li2CO3 (1 nm)/Al (100 nm) was fabricated, and theI-V characteristics at different temperatures were measured [25]. The P-doping/organic heterojunction/N-doping structure is introduced to simulate the sequencesin a real tandem device (when under reverse bias, the current flows just like thetandem device fabricated above), also to ensure ohmic contact at the electrodes.

Fig. 4.17 a C-V characteristics of capacitance devices based on C60/NaTn organic heterojunctionat a fixed frequency of 1 kHz. The device structures are ITO/LiF(100 nm)/Alq3(30 nm)/LiF(0.3 nm)/C60(20 nm)/NaTn(10 nm)/MoO3(3 nm)/NPB(50 nm)/LiF(100 nm)/Al(120 nm).b Capacitance versus the difference between the HOMOs of NaTn and the LUMO of C60 at 15 Vin the saturation regime. c Maximum power efficiency and current efficiency versus the differencebetween the HOMOs of NaTn and the LUMO of C60. d Characteristics of power efficiency andcurrent efficiency versus the hole mobility of NaTn at the current density of 1 mA/cm2. Reprintedfrom [23]


Considering the high conductivity of MoO3:TAPC and Li2CO3:BPhen dopinglayers, the voltages would mainly drop on the pentacene/C70 heterojunction. Thus,the I-V characteristics will be determined by the electrical properties of thepentacene/C70 organic heterojunction. Figure 4.18 (left) shows the C-V andI-V characteristics of the fabricated device above. In the forward direction, thecapacitance begins to drop drastically at about 0.5 V because of the significantinjection and recombination of carriers, which just corresponds to the exponentialincrease of the current. Similarly, the breakdown of current and the severe drop ofcapacitance both happen at the reverse voltage of about 1.5 V, which should becaused by the large amount of generated charges in the pentacene/C70 hetero-junction. It should be noticed that the large reverse current should mainly comefrom the charge generation instead of the injection from the electrode under reversevoltage, because the doped hole and electron transport layers could effectivelyblock the injection of electrons and holes, respectively. Generally, the C-V curve isreplotted by the Mott–Schottky relation,

1C2 ¼

2 Vb � Vð ÞeNere0A2 ð4:1Þ

where Vb is the built-in voltage, N is the density of free charge carriers, e0 is thepermittivity of free space, er is the relative dielectric constant, and A is the activearea. From Fig. 4.18 (right), a clear linear relation of the inverse capacitance squareversus the reverse voltage is seen, which corresponds well with the Mott–Schottkyrelation. Therefore, N could be calculated from the slope. The value is estimated tobe in the range of 1019 cm−3, which is quite high for organic heterojunctions,guaranteeing the highly efficient charge generation.

Figure 4.19 shows the I-V characteristics of device based on pentacene/C70

organic heterojunction under reverse bias at different temperatures. A linear relationof log (I)-log (V) is well observed, which is a strong hint of tunneling processes[26]. For classical Fowler–Nordheim (F–N) tunneling model proposed by Fowlerand Nordheim, which is described by [27]

Fig. 4.18 I-V and C-V characteristics of device based on pentacene/C70 organic heterojunction at287 K (left) and Mott–Schottky plot of device based on pentacene/C70 organic heterojunction at1 kHz. Reprinted from [25]


I / E2 exp � kE

� �ð4:2Þ

where I is the current, E is the field intensity, and k is a constant, which is related tothe height of the tunneling barrier. According to this equation, a linear relation oflog(I/E2) versus 1/E should be obtained. However, as shown in Fig. 4.20, thecurrent curves are deviated from the linearity at low electrical field, which couldprobably be attributed to the contribution of thermionic emission and deviation ofthe Fermi–Dirac distribution from the step function at temperatures higher thanabsolute zero [28]. Therefore, a modified F–N tunneling model is introduced to fitthe experimental data more precisely [29]

lnIE2

� �¼ �P1

Eþ ln

P2

E

� �� ln sin

P3

E

� �ð4:3Þ

Fig. 4.19 log(I)-log(V)characteristics of device basedon pentacene/C70 organicheterojunction under reversebias at different temperatures.Reprinted from [25]

Fig. 4.20 log(I/E2) versus1/E characteristics of devicebased on pentacene/C70

organic heterojunction underreverse bias at differenttemperatures. Reprintedfrom [25]


where E is electric field. P1, P2, and P3 are constants that have to be determined. Asshown in Fig. 4.19, the modified F–N tunneling model fits quite well with theI–V characteristics at different temperatures. This indicates that the charge gener-ation of pentacene/C70 organic heterojunction is a tunneling process.

It is well-known that HAT-CN is a strong electron accepter with a LUMO as lowas 5.46 eV and a HOMO as high as 9.44 eV [30]. Considering that most commonlyused hole transporting materials’ HOMO are lying between 5–6 eV, accordingly,HAT-CN is suitable to form a good heterojunction CGLs with them. The cases ofNPB/HAT-CN, TAPC/HAT-CN, and TPD/HAT-CN have already been reported byother groups [16, 30, 31]. In order to select an appropriate HTM to match withHAT-CN, a set of devices have been prepared: ITO/MoO3 (3 nm)/20 wt% MoO3:TAPC (50 nm)/HTM (15 nm)/HAT-CN (15 nm)/3 wt% Cs2CO3: BPhen (50 nm)/Cs2CO3 (1 nm)/Al (100 nm), here HTM includes m-MTDTA, TAPC, NPB, andTCTA. Figure 4.21 gives the comparison of J-V properties of these devices [32].Clearly, in all the tested materials, HAT-CN/m-MTDATA heterojunction shows thebest symmetry current property as well as the largest current density at the samebias. This indicates that m-MTDATA/HAT-CN organic heterojunction shouldpossess the best charge generation efficiency than the others. In addition, consid-ering the wide optical band gaps of both materials (all more than 3 eV), the excesslight absorption at visible spectrum range in this organic heterojunction could alsobe avoided when employing it as connector in tandem OLEDs.

Figure 4.22 shows the schematic energy level diagram of HAT-CN andm-MTDATA, and the space charge distribution of HAT-CN/m-MTDATA organicheterojunction. Because HAT-CN possesses higher EF than m-MTDATA, thereforethe holes will accumulate on the side of m-MTDATA and the electron will accu-mulate on the side of HAC-CN as they contact. Obviously, an accumulation-typejunction is well formed. This also further demonstrates that HAT-CN/m-MTDATAorganic heterojunction is indeed a good CGL.

Based on HAT-CN/m-MTDATA organic heterojunction as CGL, high-efficiencyred, green, and blue tandemOLEDswith phosphorescent EL unit using Ir(MDQ)2(acac),

-2 -1 0 1 2

Voltage(V)

1E-5

1E-4

1E-3

0.01

0.1

1

10

100

Cu

rren

t D

ensi

ty(m

A/c

m2 )

Fig. 4.21 Comparison of J-V properties of theHAT-CN-based organicheterojunction devices.Reprinted from [32]


Ir(ppy)2(acac) and FIrpic, respectively, as emitters have been successfully fabri-cated. Table 4.3 summarizes the EL performance of the resulting red, green, andblue tandem OLEDs based on HAT-CN/m-MTDATA organic heterojunction asCGL. Clearly, besides that the current efficiency and external quantum efficiencyare enhanced by over 2 times, the power efficiency of all tandem OLEDs isimproved, further proving the universality of accumulation-type organic hetero-junctions as CGLs to enhance the power efficiency of tandem OLEDs.

Fig. 4.22 Schematic energylevel diagram of HAT-CNand m-MTDATA, and spacecharge distribution ofHAT-CN/m-MTDATAorganic heterojunction.Reprinted from [32]

Table 4.3 Comparison of EL performance of single-unit device and tandem devices based onHAT-CN/m-MTDATA heterojunction as CGL

Device V on(V)

100 cd/m2 1000 cd/m2 5000 cd/m2

PE(lm/W)

CE(cd/A)

EQE(%)

PE(lm/W)

CE(cd/A)

EQE(%)

PE(lm/W)

CE(cd/A)

EQE(%)

Red unit 2.3 47.5 37.0 21.1 43.0 35.8 20.0 32.1 29.6 17.1

Redtandem

4.6 56.3 89.0 53.2 50.0 85.8 52.4 40.2 77.1 45.2

Greenunit

2.4 96.8 80.4 21.9 92.8 81.8 22.2 70.8 72.7 19.5

Greentandem

4.9 119.4 200.8 54.5 110.3 195.2 51.2 99.7 185.7 46.4

Blue unit 2.8 48.9 48.6 21.7 42.7 45.6 20.1 34.2 39.6 17.3

Bluetandem

5.4 52.5 99.0 44.1 47.1 96.2 42.9 39.8 87.7 39.4

The turn-on voltage (Von) and current efficiency (CE), power efficiency (PE), external quantum efficiency(EQE) at 100, 1000, and 5000 cd/m2 luminance, respectively, are given. Reprinted from [32]


The enhancement of current efficiency or external quantum efficiency of tandemOLEDs could be understood from many aspects. First, for a well-designed tandemOLED, the charges are readjusted due to the insertion of CGL, thus sometimes abetter charge balance in tandem devices could be achieved than the referencingsingle-unit devices. In addition, tandem OLEDs are much thicker than thesingle-unit ones, naturally the leakage current would be reduced. Therefore, ahigher current efficiency should be expected. Second, as reported, the fraction ofradiation coupled into surface-plasmon modes would be less when the emitter isfurther away from the cathode, and the maxima forward luminance occurs roughlyin the corresponding antinodes of the metal electrodes due to the interaction ofmicrocavity and surface-plasmon modes [33, 34]. So if the two emitting layers intandem OLEDs can be roughly put at the first and second antinode away from thealuminum cathode, then the brightness will be greatly increased, thus the externalquantum efficiency is enhanced.

However, the improvement of power efficiency in tandem OLEDs has to reducethe operational voltage and increase the brightness. The reduction of operationalvoltage is somehow a relative concept to compare the voltage drop in differentCGLs. In order to evaluate the energy loss in CGLs, we can define the ratio ofenergy conversion rate R as the achieved power efficiency (PE) divided by thetheoretical power efficiency (PT) at a certain EQE for a given spectra,

R ¼ PE

PTð4:4Þ

If 1 W electrical power is completely converted to 1 W luminous power, thenwe can get,

PT ¼ Km �R 780nm380nm L kð ÞVkdkR 780nm380nm L kð Þdk

� EQEn

ð4:5Þ

where Km = 683 lm/W, Vk is the luminous efficiency function which reflects thespectral response of the human eye, L(k) is the EL spectra intensity of OLEDs, andn is the number of EL units in the tandem OLED (for single-unit, n = 1). Theadvantage of using R to evaluate the performance of CGLs is that it can exclude theeffects of EQE related factors, such as charge balance, spectra shifting andout-coupling rate, because R is only related to the percentage of electrical powerconverted to light emission in an OLED. Taken an example [1], two white tandemOLEDs with different CGLs of TPBi:Li2CO3/TCTA:MoO3 and C60/pentacene werefabricated. We can calculate that the former tandem OLED has R = 63%, and thelater tandem OLED has R = 71%, meaning that the C60/pentacene CGL indeed hasless energy loss than the N-doped/P-doped one, considering the energy consumptionof the two tandem OLEDs in other parts is the same.

Theoretically, the R values of the single-unit OLEDs and the tandem OLEDswith two same units as the corresponding single-unit ones should be the


approximately same only when the CGLs can efficiently generate charges. For thetandem OLED based on HAT-CN/m-MTDATA as the CGL, its R = 83.2%, whichis very close to R = 83.5% of the single-unit OLED. This means that HAT-CN/m-MTDATA organic heterojunction is indeed a very optimum CGL.

The charge generation mechanism in HAT-CN/m-MTDATA organic hetero-junction has also been investigated well by the reverse bias J-V characteristics ofITO/MoO3 (3 nm)/20 wt% MoO3: TAPC (50 nm)/m-MTDATA (15 nm)/HAT-CN(15 nm)/3 wt% Cs:BPhen (50 nm)/Cs (1 nm)/Al (100 nm) (device A) at differenttemperatures. As shown in Fig. 4.23, this device gives a linear relation oflnJ * T. This indicates that the charge generation process in HAT-CN/m-MTDATAorganic heterojunction should also be charge tunneling. It is found that the currentproperties can be well explained by the tunneling model proposed by Riben et al. [35]

Jr ¼ G0Va exp �U Vd þVað Þ�1=2h i

ð4:6Þ

Jr is the reverse current density, G0 is a constant determined by the nature ofmaterial, U is the linear variable of temperature, Vd is the built-in potential, Va is theapplied reverse voltage (written as a positive value). The equation is initiallydeveloped by Zener [36] and used to well describe the electron tunneling from thevalence band of a P-type semiconductor to the conduct band of an N-type semi-conductor in staggered gap-type heterojunctions [37, 38]. Figure 4.24 shows the log(J)-log(V) characteristics of device A at different temperatures. It can be clearly seenthat there exist two regions: Under small voltages (*0.1 V), all curves have thesame slope of approximately 1, meaning an ideal ohmic injection, while underhigher voltages, the curves show a power law relation, and it is found that they canbe nicely fitted by Eq. (4.6). It has become obvious that the current property ofm-MTDATA/HAT-CN organic heterojunction is charge tunneling model.Accordingly, a schematic diagram explaining the working principles of HAT-CN/m-MTDATA organic heterojunction is given in Fig. 4.25. The band bending is

Fig. 4.23 Current densitydependence on temperaturesfor device A at a fixed reversebias of 0.5 V. Reprinted from[32]


depicted according to its charge accumulating property at equilibrium. When undercertain bias, the electrons in the HOMO of m-MTDATA would tunnel into theLUMO of HAT-CN, thus generate an electron in the LUMO of HAT-CN and leavea hole in the HOMO of m-MTDATA. Then, the generated holes and electrons gettransported and injected to the adjacent layers under the influence of electric field.Consequently, the efficient charge generation and ohmic injection of HAT-CN/m-MTDATA organic heterojunction are guaranteed by its tunneling nature, furtherresulting in the high energy conversion rate in the tandem OLEDs.

0.01 0.1 1

|Voltage|(V)

0.01

0.1

1

10

100

Cu

rren

t D

ensi

ty(m

A/c

m2 )

137 K 167 K 197 K 227 K 257 K 287 K 297 K equation fit linear fit

slope~1

Fig. 4.24 J-V plot of deviceA and the fitting curves atvarious temperatures. Thereverse voltages are absolutevalues. Reprinted from [32]

-

+

-

Vapp

HOMO

LUMO

LUMO

HOMO

HAT-CN

m-MTDATA

+

Fig. 4.25 A schematic diagram of the charge generation process in m-MTDATA/HAT-CNorganic heterojunction. Reprinted from [32]


4.4 N:P Bulk Heterojunction as Charge Generation Layerfor High-Efficiency Tandem Organic Light-EmittingDiodes

It is found that the charge generation originating from the charge transfer fromP-type organic semiconductor to N-type organic semiconductor is a prerequisite andalso very important to the organic heterojunctions as CGLs. For N/P bilayer organicheterojunction CGL systems, the charge generation occurs at the narrow interfacebetween two organic semiconductors, which may limit the generation of chargesand also cause the accumulation of the generated charges. As known, beside thebilayer heterojunctions, the bulk heterojunctions have widely been used in thefabrication of high-efficiency organic solar cells (OSCs) [39, 40]. The bulk organicheterojunction by blending the donor and acceptor organic materials was firstproposed by Heeger [39] and has been widely used to achieve high efficiency OSCswhere the blend of a P-type organic semiconductor and an N-type organic semi-conductor shows phase states of only several nanometers in microcosmic sizes,leading to an easily effective carrier diffusion. The bipolar charge generationcharacteristics of bulk organic heterojunctions imply that bulk organic hetero-junctions could also be used as excellent CGLs for the fabrication ofhigh-performance tandem OLEDs.

As seen, the bulk organic heterojunctions can indeed be used as CGLs to fabricatehigh-efficiency tandem OLEDs. Figure 4.26 presents the schematic diagram of thefabricated tandem OLED based on ZnPc:C60 bulk organic heterojunction as CGL[41]. One key prerequisite for designing the ZnPc:C60 bulk organic heterojunctionCGL is the effective control of ratios between ZnPc and C60 and the heterojunctionthickness. Considering the film morphology and transport characteristics, the ZnPc:C60 blend is fixed at a weight ratio of 1 : 1, and its thickness is 30 nm by opti-mization. Thus, the structure of the fabricated tandem OLEDs based on ZnPc:C60

bulk organic heterojunction as CGL is ITO/MoO3(6 nm)/NPB(90 nm)/Alq3:C545T

Fig. 4.26 Schematic diagramof the tandem OLED madewith ZnPc:C60 bulkheterojunction as CGL,together with the molecularstructures of ZnPc and C60.The ultra-thin interlayers ofLiF (blue) and MoO3 (red) areintroduced to achieve anefficient electron and holeinjection, respectively.Reprinted from [41]

4.4 N:P Bulk Heterojunction as Charge Generation Layer … 115

(30 nm)/Alq3(30 nm)/LiF(0.3 nm)/ZnPc:C60 (1 : 1, 30 nm)/MoO3(3 nm)/NPB(50 nm)/Alq3:C545T(30 nm)/Alq3(30 nm)/LiF(1 nm)/Al(120 nm).

Figure 4.27 shows the current efficiency-current density (a) and power efficiency-current density (b) characteristics of the resulting tandem OLEDs based on C60:ZnPcbulk organic heterojunction as CGL. The maximum current efficiency reaches36.5 cd/A, which is 2.4 times higher than that of the single-unit device (15.2 cd/A).In addition to the enhancement in current efficiency, the power efficiency is alsosignificantly improved to 21 lm/W, which is 2 times that of the single-unit device(10.1 lm/W). The other bulk organic heterojunctions as CGLs (e.g., CuPc:C60 andH2Pc:C60) are also used to further demonstrate the effectiveness of the design conceptof bulk organic heterojunctions as CGLs. It is found that a similar power efficiencyimprovement can also be achieved. Table 4.4 summarizes the EL performance of theresulting tandem OLEDs based on different organic bulk heterojunctions as CGLs.These results indicate that bulk organic heterojunctions as CGLs are also a viableoption for the fabrication of high-performance tandem OLEDs.

However, for the bulk heterojunctions as CGLs, the blended two molecules cansufficiently contact each other, and thus the charge transfer occurs in the whole bulkheterojunction, which is different from the bilayer organic heterojunctions.

Fig. 4.27 a Current efficiency-current density and b power efficiency-current density character-istics of the resulting tandem OLEDs based on C60:ZnPc bulk organic heterojunction as CGL.Reprinted from [41]

Table 4.4 Summary of the EL performance of the resulting tandem OLEDs based on differentbulk organic heterojunctions as CGLs

CGL Vt (V) L (cd/m2) ηCE, max (cd/A) ηCE (cd/A) ηPE, max (lm/W) ηPE [lm/W]

C60:ZnPc 4.9 7020 36.1 36 21 6.8

C60:CuPc 5.1 6445 33.4 31.5 22.4 5.6

C60:H2Pc 5.5 6623 34.8 34.4 16 5.8

Vt turn-on voltage examined at the brightness of 1 cd/m2. L luminance, ηCE current efficiency, andηPE power efficiency obtained at the current density of 20 mA/cm2. ηCE,max maximum value ofcurrent efficiency. ηPE,max maximum value of power efficiency. Reprinted from [41]


Therefore, the charge transfer in bulk heterojunctions should occur more easily thanthat in bilayer organic heterojunctions, thus generating more charges. This is provenby the achievement of higher current and power efficiencies in tandem OLEDsbased on ZnPc:C60 bulk organic heterojunction as CGL versus in tandem OLEDswith ZnPc/C60 bilayer organic heterojunction as CGL.

So, the charge carrier mobility then begins to play a major role in bulk organicheterojunctions as CGLs. As we know, the mobility of electrons and holes in bulkorganic heterojunctions will be lower than those in the pure materials used inbilayer organic heterojunctions [41]. If the mobility is not sufficiently high, thecharge carriers will not be transported rapidly by external electric field. As a result,the excess charges will recombine again or remain in the heterojunction as unde-sirable space charges that oppose the drift of new charges. The latter problem canoccur if the electron and hole mobilities are highly imbalanced such that one speciesis much more mobile than the other. In this case, the device performance will beworse. Therefore, it is very necessary to choice high mobility organic semicon-ductors in order to construct highly efficient bulk heterojunction as CGLs.

It is well-known that ZnPc and C60 blend is a rich charge generation system inorganic solar cells due to the effective electron transfer from ZnPc to C60 [42]. Inreality, several groups have successfully described the electronic structures of theZnPc/C60 heterojunction by in situ ultraviolet photoelectron spectroscopy [43, 44].It has been demonstrated that ZnPc:C60 blend facilitates the charge transfer betweenthe donor and acceptor. Also recent photophysical and theoretical investigations onZnPc/C60 heterojunction show that the ground-state electronic interaction betweenZnPc and C60 has been evidenced from the observation of well-defined chargetransfer absorption bands in the visible region [45]. Taking great benefit from thecharge redistribution, in this case, the accumulation of free electrons and holes inthe bulk heterojunction can be formed by the charge transfer processes from theHOMO of ZnPc to the LUMO of C60 in the ZnPc:C60 bulk heterojunction, asshown in Fig. 4.28. Therefore, under the application of an electric field, thebounded electron–hole pairs can be separated into free electrons and holes, and thegenerated electrons and holes effectively transport along the LUMO of C60 and theHOMO of ZnPc, and finally effectively inject into the corresponding EL units.

Fig. 4.28 Operationprinciple of charge transferbetween ZnPc and C60, andthe charge transport in ZnPc:C60 bulk heterojunction asCGL induced by an electricalfield. Reprinted from [41]

4.4 N:P Bulk Heterojunction as Charge Generation Layer … 117

4.5 N/N:P/P Composited Heterojunction as ChargeGeneration Layer for High-Efficiency TandemOrganic Light-Emitting Diodes

Taking into account the distinguishing features of N/P bilayer organic hetero-junctions and N:P bulk heterojunctions as CGLs, it is predicted that the N/N:P/Pcomposited organic heterojunctions should also be a very good structure as CGLsto fabricate high-efficiency tandem OLEDs. An advanced example is the HAT-CN/HAT-CN:TAPC/TAPC composited organic heterojunction as CGL, wherehigh-efficiency phosphorescent tandem OLEDs are successfully fabricated [12, 46].Compared with HAT-CN/TAPC bilayer heterojunction, as shown in Fig. 4.29,where the capacitance–voltage characteristics of capacitance devices based onHAT-CN/HAT-CN:TAPC/TAPC composited heterojunction and HAT-CN/TAPCbilayer heterojunction are given, the HAT-CN/HAT-CN:TAPC/TAPC compositedheterojunction products larger capacitance at the same bias voltage, indicating thatthe HAT-CN/HAT-CN:TAPC/TAPC composited heterojunction does generatemore charges, which should be favor for the fabrication of high-efficiencytandem OLEDs.

Figure 4.30 shows the schematic diagram of the resulting red, green, and bluetandem OLEDs based on HAT-CN/HAT-CN:TAPC/TAPC composited hetero-junction as CGL. The EL performance of all tandem OLEDs is depicted inFig. 4.31 and summarized in Table 4.5. The turn-on voltages of red, green, andblue TOLEDs reach 4.6, 4.9, and 5.4 V, respectively, which are about twice of thesingle-unit devices. This means that the voltage drops on the HAT-CN/HAT-CN:TAPC/TAPC composited heterojunction CGL, and its interface barriers withcharge-transporting layers are pretty small. It can be seen that the current efficiencyof the fabricated tandem OLEDs is greatly enhanced by more than doubled at thewhole brightness range compared with the corresponding single-unit devices.

0 1 2 3

1

1.2

1.4

1.6 A(bilayer) B(BHJ)

C/C

0

Voltage (V)

Fig. 4.29 C-V characteristics of capacitance devices based on HAT-CN/HAT-CN:TAPC/TAPCcomposited heterojunction and HAT-CN/TAPC bilayer heterojunction, where the device structuresare ITO/LiF (50 nm)/HAT-CN (10 nm)/HAT-CN:TAPC (1:1, 40 nm)/TAPC (10 nm)/LiF(50 nm)/Al and ITO/LiF (50 nm)/HAT-CN (30 nm)/TAPC (30 nm)/LiF (50 nm)/Al, respectively.Reprinted from [46]


Taken the green device for example, it can be seen that the brightness is indeedenhanced by more than two times, from 848 cd/m2 of the single-unit device to2050 cd/m2 of the tandem device at 1 mA/cm2, although the voltage is doubled atthe same time. Similar results are also obtained in red and blue TOLEDs. Asexpected, therefore, the power efficiencies of all the tandem devices are greatlyimproved from the single-unit devices that exhibited a max power efficiency of48.1, 103.8, and 49.9 lm/W, respectively, reaching 57.5, 126.8, and 52.7 lm/W atmaximum values for red, green, and blue color, and only slightly reduced to 50.0,121.7, and 47.1 lm/W at the brightness of 1000 cd/m2. This also proves directly thevalidity of N/N:P/P composited heterojunctions as CGLs.

It is found that HAT-CN/HAT-CN:TAPC/TAPC composited heterojunction ishighly conductive. In order to prove this point, a set of hole-only devices withstructures of ITO/HAT-CN (10 nm)/HAT-CN:TAPC (30 wt%) (X nm)/TAPC(Y nm)/Al (device A-C) was fabricated. Figure 4.32 shows the J-V properties ofthese devices. It can be seen that they exhibit almost identical current propertieswith thickness, indicting high conductivity. The slope of approximate 2 means that

Fig. 4.30 Schematic diagram of the resulting red, green, and blue tandem OLEDs based onHAT-CN/HAT-CN:TAPC/TAPC composited heterojunction as CGL. Reprinted from [46]

4.5 N/N:P/P Composited Heterojunction as Charge Generation Layer … 119

the conductance in these devices is space charge limited, and the J-V characteristicsfollow Child’s law [47]

J ¼ 98ere0lp

V2

d3ð4:7Þ

where e0 is the permittivity of free space, er is the relative permittivity, lp is thezero-field hole mobility, and d is the thickness of the space charge layer. From the

Fig. 4.31 EL performance of the resulting red, green, and blue tandem OLEDs based onHAT-CN/HAT-CN:TAPC/TAPC composited heterojunction as CGL. Here, a and b for reddevice, c and d for green device, and e and f for blue device. Reprinted from [46]


slope of the log(J)–log(V) plot, a clear quadratic dependence of current on appliedvoltage is observed, which can be well fitted by Eq. (4.7). This property is mainlydominated by the intrinsic TAPC.

The high conductivity is further proven by capacitance–frequency(C-F) spectroscopy in these devices, shown in Fig. 4.33. The capacitance is flatunder 10 kHz, which is a strong intimation of trap-free transportation, or someshallow hole traps exist, but are filled by the free carriers in organic heterojunctionsystems [48]. So in the following calculation, the trap contributions to the capac-itance are not taken into account. It can be seen that the variation of blending layerthickness in heterojunction does not change the capacitance of the devices, whichmeans that the composited heterojunction indeed possesses rather high conductivity[49]. In contrast, the thickness of intrinsic TAPC layer affects significantly thedevice capacitance, from 40.6 to 20.3 nF at 5 kHz when the intrinsic TAPCthickness varies from 10 to 20 nm, exactly showing the same trend as an ideal

Fig. 4.32 J-V properties ofITO/HAT-CN (10 nm)/HAT-CN:TAPC (30 wt%)(X nm)/TAPC (Y nm)/Aldevices. The red solid line isfitted by Eq. (4.7). Reprintedfrom [46]

Table 4.5 Summary of EL performance of the resulting red, green, and blue tandem OLEDs basedon HAT-CN/HAT-CN:TAPC/TAPC composited heterojunction as CGL. Reprinted from [46]

Device V on

(V)100cd/m2 1000 cd/m2 5000 cd/m2

PE(lm/w)

LE(cd/A)

EQE(%)

PE(lm/w)

LE(cd/A)

EQE(%)

PE(lm/w)

LE(cd/A)

EQE(%)

Red unit 2.3 47.5 37.0 22.5 43.0 35.8 20.1 32.1 29.6 17.3

Redtandem

4.6 56.3 89.0 55.3 50.0 85.8 52.4 40.2 77.1 45.8

Greenunit

2.5 103.4 85.9 21.9 92.7 84.1 22.2 65.6 77.4 20.7

Greentandem

4.9 124.1 201.8 51 6 121.7 210.1 53.7 100.8 192.8 49.4

Blue unit 2.8 48.9 48.6 21.7 42.7 45.6 20.1 34.2 39.6 17.3

Bluetandem

5.4 52.5 99.0 44.1 47.1 96.2 42.9 39.8 87.7 39.4


parallel plate capacitor. Therefore, it is assumed that the intrinsic TAPC layer isfully depleted in these devices, and the intrinsic layer is the main contributor to thecapacitance. Using the capacitance equation of a parallel plate capacitor:

C0 ¼ ere0A=w ð4:8Þ

C0 is the capacitance at zero bias (at a fixed frequency of 5 kHz), w is the widthof depletion layer, A is the effective area of the device, e0 is the permittivity of freespace, er is the relative dielectric constant, commonly an er of 3–4 is assumed formost organic semiconductors. Brought the capacitance of the devices into Eq. (4.8),the corresponding depletion layer width of about 10.3 and 20.6 nm is calculated,which is identical to the assumption that the intrinsic TAPC layer is fully depleted.

Furthermore, according to the classical evaluation of the Mott–Schottky relationfor such devices [50], the C-V relation can be written by,

1C2 ¼

2 Vb � Vð ÞeNAere0A2 þ d2i

ðere0AÞ2ð4:9Þ

where e is the charge of one electron, di the thickness of intrinsic layer, NA thedensity of ionized free charges. Thus, NA is estimated to be at the range of 1019/cm3

in the composited heterojunction from Fig. 4.34 by Eq. (4.9), which is several ordershigher than that in a typical intrinsic organic semiconductor (usually less than 1015/cm3) [51]. This high free charge density confirms that the HAT-CN/HAT-CN:TAPC/TAPC composited heterojunction could generate free charges efficiently,which makes it highly conductive and preferable for hole transport.

In order to quantitatively reveal the electron injection process from HAT-CN toadjacent ETL in above devices, ITO/HAT-CN (20 nm)/Cs:BPhen (100 nm)/Aldevice E has been fabricated. As seen, under reverse bias, the electrons are injectedfrom HAT-CN to N-doped BPhen, simulating the real sequence in tandem OLEDs.In this device, both ITO/HAT-CN and Cs:BPhen/Al contact can be treated as ohmic,meanwhile Cs:BPhen is also highly conductive. Therefore, the C–F property would

Fig. 4.33 C-F characteristicsof ITO/HAT-CN (10 nm)/HAT-CN:TAPC (30 wt%)(X nm)/TAPC (Y nm)/Alhole-only devices A-D at zerobias. Reprinted from [46]


be determined by the thickness of the depletion layer (wd) in the N-doped BPhenlayer. When the reverse bias increases, the depletion region becomes wider, thus thecapacitance will be smaller, as shown in Fig. 4.35a, and as the reverse bias increasesfrom 0 to −1 V, the device capacitance drops correspondingly. It can be clearly seenthat the capacitance decreases slowly below 5 kHz, which could be attributed to therole of traps in BPhen. Above 10 kHz, the capacitance drops drastically because inthis case, the bulk relaxation can no longer follow the high frequency [52]. From thecapacitance under zero bias at 5 kHz, then the depletion layer can estimate to be4.5 nm in Cs:BPhen by Eq. (4.8). The typical Mott–Schottky plot is plotted in theinset of Fig. 4.35a, from which the built-in potential Vb can be obtained to be about0.3 V. Furthermore, it is found that the J-V characteristics of device E under reversebias can be well described by an electron tunneling model [53]

Jr ¼ �e2aNth�1V

� �exp U Vb � Vð Þ�1

2

h ið4:10Þ

here Jr is the density of reverse tunneling current, V is the applied voltage, Nt is thestate density involved in the tunneling process, h is the Planck constant, a is thetunneling distance (presume it to be 1 nm here), U is a constant related to thematerial property, Vb is the built-in potential, which can be obtained by the Mott–Schottky plot above. Using Eq. (4.10) to fit the J-V plot of device E, as shown inFig. 4.35b, a perfect agreement is obtained. From the fitting, Nt is calculated to be atthe range of 1016, which is several magnitudes smaller than the typical density ofLUMO states in organic semiconductors, so the tunneling process may not takeplace at the LUMO level. In organic semiconductors, the distribution of band statesis considered to be Gaussian type [54], which could extend much deeper into theband gaps compared to inorganic ones. Therefore, we proposed that the tunneling atthe HAT-CN/N-BPhen interface is via gap states rather than the LUMO states.

Fig. 4.34 Mott–Schottkyplot of ITO/HAT-CN(10 nm)/HAT-CN:TAPC (30wt%) (100 nm)/TAPC(10 nm)/Al hole-only deviceB at 5 kHz. Reprinted from[46]


With the depletion width Wd and built-in potential Vb extracted from thecapacitance spectroscopy, a detailed energy diagram can be drawn in Fig. 4.36. Thefree electrons and holes are generated in the HAT-CN:TAPC layer and separatedunder certain voltage. Afterward the electrons transport through an intrinsicHAT-CN and then inject into N-doped BPhen layer. The depletion region at theN-doped side is very thin, which make the electron tunneling rather easy.

Fig. 4.36 Schematic diagram of working principles of the designed composited heterojunction.The dotted line in transverse direction is the quasi-Fermi level, vacuum-level pinning D iscalculated using D = LUMOHAT-CN − LUMOBPhen − eVb. Reprinted from [46]

Fig. 4.35 a C-F plot of ITO/HAT-CN (20 nm)/Cs:BPhen (100 nm)/Al device E at differentreverse biases. The Mott–Schottky plot of device E at 1 kHz is given in inset. b J-V plot of deviceE under reverse bias (voltage in absolute value). The solid line is the fitting result from Eq. (4.10).Reprinted from [46]


References

1. Y.H. Chen, J.S. Chen, D.G. Ma, D.H. Yan, L.X. Wang, Appl. Phys. Lett. 99, 103304 (2011)2. J. Kido, T. Matsumoto, T. Nakada, J. Endo, K. Mori, N. Kawamura, A. Yokoi, SID. Int.

Symp. Dig. Tech. 34, 964 (2003)3. H.D. Sun, Y.H. Chen, J.S. Chen, D.G. Ma, IEEE J. Sel. Top. Quantum Electron. 22, 7800110

(2016)4. H. Sasabe, K. Minamoto, Y.J. Pu, M. Hirasawa, J. Kido, Org. Electron. 13, 2615 (2012)5. Y.H. Lee, M.W. Lin, T.C. Wen, T.F. Guo, J. Appl. Phys. 114, 154512 (2013)6. Q.Y. Bao, J.P. Yang, Y.Q. Li, J.X. Tang, Appl. Phys. Lett. 97, 063303 (2010)7. M.K. Fung, Y.Q. Li, L.S. Liao, Adv. Mater. 47, 10391 (2016)8. Q.Y. Bao, J.P. Yang, Y. Xiao, Y.H. Deng, S.T. Lee, Y.Q. Li, J.X. Tang, J. Mater. Chem. 21,

17476 (2011)9. L.S. Liao, K.P. Klubek, C.W. Tang, Appl. Phys. Lett. 84, 167 (2004)

10. Y.H. Chen, Q. Wang, J.S. Chen, D.G. Ma, D.H. Yan, L.X. Wang, Org. Electron. 13, 1121(2012)

11. Y.H. Chen, J.S. Chen, D.G. Ma, D.H. Yan, L.X. Wang, J. Appl. Phys. 110, 074504 (2011)12. Y.F. Dai, H.M. Zhang, Z.Q. Zhang, Y.P. Liu, J.S. Chen, D.G. Ma, J. Mater. Chem. C 3, 6809

(2015)13. M. Kroger, S. Hamwi, J. Meyer, T. Dobbertin, T. Riedl, W. Kowalsky, H.H. Johannes, Phys.

Rev. B 75, 235321 (2007)14. Y.H. Lee, M.W. Lin, T.C. Wen, T.F. Guo, J. App. Phys. 114, 154512 (2013)15. S.H. Lee, J.H. Lee, J.H. Lee, J.J. Kim, Adv. Funct. Mater. 22, 855 (2012)16. L.S. Liao, W.K. Slusarek, T.K. Hatwar, M.L. Ricks, D.L. Comfort, Adv. Mater. 20, 324

(2008)17. Y.H. Chen, D.G. Ma, J. Mater. Chem. 22, 18718 (2012)18. S.L. Lai, M.Y. Chan, M.K. Fung, C.S. Lee, S.T. Lee, J. Appl. Phys. 101, 014509 (2007)19. L.S. Liao, K.P. Klubek, Appl. Phys. Lett. 92, 223311 (2008)20. D.H. Yan, H.B. Wang, B.X. Du, Introduction to Organic Semiconductor Heterojunctions

(Wiley, 2010)21. B.L. Sharma, R.K. Purohit, Semiconductor Heterojunctions (Pergamon Press, Oxford, 1974)22. X.L. Liu, C.G. Wang, C.C. Wang, I. Irfan, Y.L. Gao, Org. Electron. 17, 325 (2015)23. Y.H. Chen, H.K. Tian, Y.H. Geng, J.S. Chen, D.G. Ma, D.H. Yan, L.X. Wang, J. Mater.

Chem. 21, 15332 (2011)24. H.K. Tian, J.W. Shi, B. He, N.H. Hu, S.Q. Dong, D.H. Yan, J.P. Zhang, Y.H. Geng, F.S.

Wang, Adv. Funct. Mater. 17, 1940 (2007)25. Q.X. Guo, H.D. Sun, J.X. Wang, D.Z. Yang, J.S. Chen, D.G. Ma, J. Mater. Chem. C 4, 376

(2016)26. I.D. Parker, J. Appl. Phys. 75, 1656 (1994)27. R.H. Fowler, D.L. Nordhe, Proc. R. Soc. Lond. 119, 173–181 (1928)28. A.J. Heeger, I.D. Parker, Y. Yang, Synth. Met. 67, 23 (1994)29. M. Koehler, I.A. Hummelgen, Appl. Phys. Lett. 70, 3254 (1997)30. E. Oha, S. Park, J. Jeong, S.J. Kang, H. Lee, Y. Yi, Chem. Phys. Lett. 668, 64 (2017)31. T. Chiba, Y.J. Pu, R. Miyazaki, K. Nakayama, H. Sasabe, J. Kido, Org. Electron. 12, 710

(2011)32. H.D. Sun, Q.X. Guo, D.Z. Yang, Y.H. Chen, J.S. Chen, D.G. Ma, ACS Photon. 2, 271–279

(2015)33. C.L. Lin, T.Y. Cho, C.H. Chang, C.C. Wu, Appl. Phys. Lett. 88, 111106 (2006)34. R. Meerheim, M. Furno, S. Hofmann, B. Luessem, K. Leo, Appl. Phys. Lett. 97, 253305

(2010)35. A.R. Riben, D.L. Feucht, Inter. J. Electron. 20, 583–599 (1966)36. C. Zener, Proc. R. Soc. A Math. Phys. Eng. Sci. 145, 523 (1934)37. K. McAfee, E. Ryder, W. Shockley, M. Sparks, Phys. Rev. 83, 650 (1951)

References 125

38. E.O. Kane, J. Phys. Chem. Solids 12, 181 (1960)39. G. Yu, J. Gao, J.C. Hummelen, F. Wudl, A.J. Heeger, Science 270, 1789 (1995)40. S.H. Park, A. Roy, S. Beaupre, S. Cho, N. Coates, J.S. Moon, D. Moses, M. Leclerc, K. Lee,

A.J. Heeger, Nat. Photon. 3, 297 (2009)41. Y.H. Chen, J.S. Chen, D.G. Ma, D.H. Yan, L.X. Wang, F.R. Zhu, Appl. Phys. Lett. 98,

243309 (2011)42. W.J. Zeng, K.S. Yong, Z.M. Kam, F.R. Zhu, Y.N. Li, Appl. Phys. Lett. 97, 133304 (2010)43. S.H. Park, J.G. Jeong, H.-J. Kim, S.-H. Park, M.-H. Cho, S.W. Cho, Y. Yi, M.Y. Heo, H.

Sohn, Appl. Phys. Lett. 96, 013302 (2010)44. C. Hein, E. Mankel, T. Mayer, W. Jaegermann, Sol. Energy Mater. Sol. Cells 94, 662 (2010)45. A. Ray, D. Goswami, S. Chattopadhyay, S. Bhattacharya, J. Phys. Chem. A 112, 11627

(2008)46. H.D. Sun, Y.H. Chen, L.P. Zhu, Q.X. Guo, D.Z. Yang, J.S. Chen, D.G. Ma, Adv. Electron.

Mater. 1, 1500176 (2015)47. M.A. Lampert, P. Mark, Current Injection in Solids (Academic Press Inc., New York, 1970)48. Y. Zhang, B. de Boer, P.W.M. Blom, Phys. Rev. B 81, 085201 (2010)49. P. Pahner, H. Kleemann, L. Burtone, M.L. Tietze, J. Fischer, K. Leo, B. Lussem, Phys. Rev.

B 88, 195205 (2013)50. J. Drechsel, M. Pfeiffer, X. Zhou, A. Nollau, K. Leo, Synth. Met. 127, 201 (2002)51. K. Walzer, B. Maennig, M. Pfeiffer, K. Leo, Chem. Rev. 107, 1233 (2007)52. L. Burtone, J. Fischer, K. Leo, M. Riede, Phys. Rev. B 87, 045432 (2013)53. A.R. Riben, D.L. Feucht, Int. J. Electron. 20, 583 (1966)54. M.L. Tietze, L. Burtone, M. Riede, B. Luessem, K. Leo, Phys. Rev. B 86, 035320 (2012)


Chapter 5Tandem White Organic Light-EmittingDiodes Based on Organic SemiconductorHeterojunctions

5.1 Basic Structures of Tandem White OrganicLight-Emitting Diodes

As we know, tandem OLEDs (TOLEDs) are technologically interesting because notonly the luminance and current efficiency can be improved linearly with the numberof electroluminescent (EL) units in the TOLEDs, but also leakage current andbreakdown of the electric field can be avoided due to the higher luminance at a lowcurrent density and the thicker organic films, resulting in a long lifetime.Importantly, the state-of-the-art TOLEDs are very easy to vertically stack eitherindividual red, green, and blue emission units or multiple white emission units inseries via charge generation layers (CGLs) to achieve white emission [1].

There are two kinds of approaches to construct the TOLEDs to realize white lightemission by voltage control. Figure 5.1 shows their schematic structure diagrams.The one is the TOLED structures with two (or more) independently addressableunits emitting light of different colors [2, 3]. This approach provides the advantageof much greater control over the emission spectrum, thus the easy change of colortemperature (CT). However, to individually address each unit, an additional elec-trode must be added into the device stack as a connection between two adjacentunits. This intermediate electrode must be transparent and therefore must be madeeither of a thin metal layer (* 15 nm) or from indium tin oxide. As we know, thedeposition of indium tin oxide can be problematic due to sputter-induced damage tothe organic material underneath [4], whereas metal films absorb a significant amountof light and introduce additional microcavity effects [5]. As a result, the indepen-dently addressable TOLEDs that have been reported thus far demonstrate relativelymodest efficiencies (<10% external quantum efficiency (EQE), <10 lm/W), despitethe use of phosphorescent emitter systems [6]. Therefore, the development of highlyefficient color-tunable TOLEDs by independently addressing remains experimen-tally challenging.


127

The alternative is the TOLED structures with two (or more) units emitting lightof different colors in series by CGLs as intermediate connector [7], which has beenreported to be the best structures due to its high efficiency, good color stability, andlong lifetime. It can be seen that this kind of TOLEDs is driven by a single powersource. The CGLs that play a role of similar metal electrode is floating between twoemitting units, which is completely different from the case of metal electrode as theinterconnector that needs to be connected to the external power supply. The elec-trons and holes injected into respective emitting units from the intermediate con-nector are generated in the floated CGLs under external electric field. This is themost important feature of CGLs work; therefore, the CGLs play an important role inthe achievement of high-efficiency white TOLEDs. Obviously, the luminance andcurrent efficiency of this kind of TOLEDs would be the sum of each EL unit, butthe power efficiency cannot be enhanced due to the increase of working voltage,especially when an undesirable CGL is used. Therefore, the development of effi-cient CGLs has become an important research topic.

The white emission of TOLEDs with CGLs can be well realized by effectivecombination of emitting units with different colors. The reported structures includethe types of “W + W” [8], “B + Y” [9], “B + GR” [10], “B + YR” [11], “B + G+ R” [12], “B + GY + R” [13], and “Y + B + Y” [14]. The schematic structurediagrams are summarized in Fig. 5.2. They all show good white emission with highefficiency. In these white TOLEDs, the first consideration is the design of CGLs,mainly concluding MoO3/NPB, HAT-CN/NPB, HAT-CN/TAPC, NDN-1:NET-5/NDP-2:NHT-5, C60/pentacene, etc. It can be seen that organic heterojunc-tion types become important structures of CGLs.

As lighting sources, white OLEDs must simultaneously have high efficiency andhigh color rendering index (CRI) as well as enough luminance. However, CRI andluminous efficacy are usually in trade-off relation in actual white OLEDs [15].Therefore, it is necessary for us to theoretically and experimentally design andoptimize the structure parameters of white OLEDs. Cho et al. [16] simulated

Fig. 5.1 Schematic structure diagrams of two kinds of white TOLEDs that emit white light. Here,R: red, G: green, B: blue, ETL: electron-transporting layer, HTL: hole-transporting layer, EML:emitting layer

128 5 Tandem White Organic Light-Emitting Diodes Based on Organic …

three-color white TOLEDs with two emitting units to simultaneously obtain highefficiency and high CRI. They first optimized CRI by comparing two different blueemitters, namely bis[2-(4,6-difluorophenyl)pyridinato-N,C2′](picolinato)iridium(III) (FIrpic) and tris[2-(4-fluorophenyl)-1-(5′-isopropyl-(1,1′:3′,1″-terphenyl)-2′-yl)-1H-imidazole] iridium(III) (Ir(itpim)3). Figure 5.3a shows the emissionspectra of yellow-green, red, and blue emitters that are used to obtain white spectra.Thus, various white spectra can be constructed by varying the portions of yellow-green (yG) and red (xR) emission as follows,

SW ¼ xRSR þ yGSG þ SB ð5:1Þ

where SW, SR, SG, and SB are the emission spectra of white, red, yellow-green, andblue, respectively. For example, with a ratio of yellow-green to red as xR:yG = 2:3,an SW is obtained, as shown in Fig. 5.3b. Because Ir(itpim)3 has a wide spectrum atshorter wavelength than that of FIrpic, SW with Ir(itpim)3 shows a wider spectralregion.

W+W B+Y B+GR

B+G+R B+GY+R Y+B+Y

Fig. 5.2 Schematic structure diagrams of white TOLEDs with different emitting units

5.1 Basic Structures of Tandem White Organic Light-Emitting Diodes 129

Figure 5.3c, d give the CRI contour plots of arbitrary SW with Ir(itpim)3 andFIrpic, and various relative ratios of red and yellow-green. It can be seen thatwide-ranging CRI values can be obtained by varying the relative augmentation ratioof yellow-green and red. Both blue dopants can achieve very high CRI (>90) if redand yellow-green portions are well controlled. However, Ir(itpim)3 offers a widerlatitude in achieving high CRI, and Ir(itpim)3 also yields a larger CRI than FIrpicfor a given pair of xR and yG, strongly indicating the importance of the presence ofshorter wavelength emission in achieving high CRI white TOLEDs.

However, the blue emission has a limited impact on the luminous efficacy.Generally, the luminous efficacy (ηP) can be calculated from the luminous flux(F) and electrical power (P) as [17]

gp ¼FP¼ 683:0

REEl kð ÞgocVðkÞ dkR

EEl kð Þ 1gint

qVkhc dk

ð5:2Þ

where EEl(k) and ηoc are the normalized radiant flux and out-coupling efficiency ofOLEDs, ηint is the internal quantum efficiency, i.e., the amount of excitons neededto create one photon, V(k) is the photopic response curve, h is the Planck constant,

Fig. 5.3 a EL spectra of FIrpic, Ir(itpim)3, red and yellow-green. b An example of arbitrary whitespectrum constructed by the relative portion of blue: yellow-green: red spectrum = 1:2:3. CRI ofarbitrary white spectra with c Ir(itpim)3 and d FIrpic dopant according to the portion of red (x-axis)and yellow-green (y-axis). Reprinted from [16]


c is the light speed, V is the operating voltage, k is the wavelength, and hc/qVkrepresents the relation between the optical power of the photon and energy forcreating the exciton. Figure 5.4a, b show the simulation of luminous efficacy ofarbitrary white spectrum SW with Ir(itpim)3 and FIrpic by Eq. (5.2). It can be seenthat the calculated luminous efficacies of two blue dopants are more or less similar.The luminous efficacy is dominantly influenced by the emission ratios ofyellow-green. As seen, when the portion of yellow-green emission increases, theluminous efficacy of the WOLEDs increases proportionally. This means that a highEL emission at a short wavelength does not necessarily yield a high luminousefficacy.

Thus, a final overall performance good white TOLED can be determined byoptimizing the thickness of ETL and HTL. Figure 5.5a, b shows the EQE contourplots of blue and red emissions as functions of HTL and ETL thicknesses by optical

Fig. 5.4 Luminous efficacy of arbitrary white spectra with a Ir(itpim)3 and b FIrpic dopantaccording to the red (x-axis) and yellow-green (y-axis) portions. Reprinted from [16]

Fig. 5.5 EQE contour plots of a blue and b red emissions as functions of the HTL and ETLthicknesses. The simulation parameters, including the thicknesses of the HTL and ETL, range from50 nm to 350 nm in a device structure made of Glass/indium tin oxide (ITO) (70 nm)/HTL (x nm)/EML (10 nm)/ETL (y nm)/Al (100 nm). Reprinted from [16]

5.1 Basic Structures of Tandem White Organic Light-Emitting Diodes 131

simulation. Obviously, the achievement of high-efficiency white TOLEDs mustsimultaneously chooses the resonance thickness of ETL and HTL. Because the ETLthickness determines the distance between the EML and reflective cathode, so theeffect of ETL thickness is more significant than the HTL thickness.

By simulating and optimizing, the actually obtained structure of white TOLEDsis Glass/ITO (70 nm)/HTL (190 nm)/blue:yellow-green EML (10 nm)/ETL(20 nm)/CGL (50 nm)/HTL (45 nm)/yellow-green:red EML (10 nm)/ETL (70 nm)/LiF/Al, whose cross-sectional schematics are shown in Fig. 5.6a. Figure 5.6b showsthe EL spectra of white devices with Ir(itpim)3 and FIrpic. Measured efficiencies andCRI are summarized in Table 5.1. It can be seen that the white TOLEDs with Ir(itpim)3 exhibit higher CRI than those with FIrpic. The stronger emission in shorterwavelength makes it possible to achieve a CRI higher than 90. The slightly higherefficiency of white TOLEDs with Ir(itpim)3 than that with FIrpic is because the blueunit itself has different efficiency for two dopants.

5.2 Fluorescence Tandem White Organic Light-EmittingDiodes

Due to the advantages of long lifetime and low cost of fluorescence materials usedin the fabrication of white OLEDs, people are always making an effort to furtherenhance the efficiency of fluorescence white OLEDs. However, the limitation of

Fig. 5.6 a Cross-sectional schematic diagram and b EL spectra of the fabricated white TOLEDswith Ir(itpim)3 and FIrpic. Reprinted from [16]

Table 5.1 Efficiency, CRI,and CIE of the fabricatedwhite TOLEDs with Ir(itpim)3and FIrpic

EQE(%)

PE(lm/W)

CRI Colorcoordinate

Ir(itpim)3

39.86 38.34 90.94 (0.452, 0.439)

FIrpic 36.58 35.63 88.76 (0.475, 0.449)


intrinsic quantum efficiency due to 25% singlet exciton emission hinders theirapplication in lighting. Tandem structures become one of main methods to furtherimprove the efficiency of fluorescence white OLEDs [18, 19]. Figure 5.7a showsthe structure of a two-stack tandem fluorescence white OLED [20]. In this device, a“R/Y/G” emissive unit and a “Y/B” emissive unit are connected by a CGL con-sisted of doped-N/doped-P heterojunction. The EL spectrum is shown in Fig. 5.7b,where a broader spectrum meets the color requirements for lighting. This devicehad a CCT of 3500 K and a color within the Energy Star specifications for SSLwith a CRI of 75. This device also showed high efficiency and long lifetime.A current efficiency of 38 cd/A and a power efficiency of nearly 20 lm/W at adriven voltage of 6.0 V were achieved. The external quantum efficiency reached13.9%. The lifetime was estimated to be 140,000 h at an initial luminance of1000 cd/m2. If considering the light extraction technique, the power efficiency ofthe same tandem white OLEDs is estimated to be over 40 lm/W. The exceptionalefficiency and lifetime achieved demonstrates that tandem white OLEDs based onfluorescent emitters are a compelling choice for first generation OLED solid-statelighting products having a long lifetime as the most critical requirement.

However, in terms of efficiency, it does not still meet the requirement of lightingapplications. It is expected for us to make a breakthrough in fluorescence materials,thus further enhancing the efficiency of the resulting fluorescence white TOLEDs.An alternative route is using thermally activated delayed fluorescence (TADF)materials, which are successfully developed by Adachi et al. [21]. In the TADFprocess, light emission can be extracted as delayed fluorescence after intersystemcrossing (ISC) from T1 to S1 states in TADF molecules, resulting in efficientradiative decay from S1 state. Therefore, TADF materials can realize fluorescenceOLEDs with nearly 100% excitons emission [22, 23], also including efficientfluorescence white OLEDs [24, 25]. For example, Adachi et al. [23] demonstratedhighly efficient fluorescence OLEDs with EQE as high as 13.5, 15.8, 18, and 17.5%

Fig. 5.7 a Schematic structure diagram and b EL spectrum of two-stack tandem fluorescencewhite OLEDs. Reprinted from [20]

5.2 Fluorescence Tandem White Organic Light-Emitting Diodes 133

for blue, green, yellow, and red colors, respectively. The devices are realized byutilization of TADF molecules as assistant dopants that permit efficient transfer ofall electrically generated singlet and triplet excitons from the assistant dopants tothe fluorescent emitters. The schematic illustration of proposed energy transfermechanism in the emitter with assistant dopant under electrical excitation andchemical structures of the assistant dopants used in this study are shown in Fig. 5.8.The device EL performance of the four color OLEDs with assistant dopants fortriplet harvesting is then summarized in Table 5.2. It can be seen that the powerefficiencies of blue, green, yellow, and red OLEDs also reach 18, 47, 58, and28 lm/W and remain at 7, 30, 33, and 10 lm/W at 1000 cd/m2 luminance,respectively, which are much more higher than that of pure fluorescence-basedOLEDs. Moreover, high-efficiency fluorescence white OLEDs with maximumefficiency of 18.2% and 44.6 lm/W have also been realized by strategic manage-ment of singlet and triplet excitons within an efficient emissive zone. Figure 5.9gives the device structure with energy levels and exciton processes of the designedfluorescence white OLEDs and the molecular structures of the used organic ma-terials in this device. The EML consisted of two separated red/green sub-EMLs

Fig. 5.8 Schematic illustration of proposed energy transfer mechanism in the emitter withassistant dopant under electrical excitation and chemical structures of the assistant dopants used inthis study. S and T represent singlet and triplet states, respectively. A, E, and H express assistantdopant, emitter, and host, respectively. Reprinted from [23]


Table 5.2 Summary of EL performance of the four color OLEDs with assistant dopants for tripletharvesting. Reprinted from [23]

Device Turn-onvoltage(V)

MaxEQE(%)

MaxCE(cd/A)

MaxPE(Im/W)

CIE Performance at 1,000 cd/m2

Voltage(V)

EQE(%)

CE(cd/A)

PE(Im/W)

Blue 4.7 13.4 27 18 (0.17,0.30) 7.8 8.7 18 7

Green 3.0 15.8 45 47 (0.29,0.59) 4.1 11.7 38 30

Yellow 3.2 18.0 60 58 (0.45,0.53) 5.2 17.2 56 33

Red 3.0 17.5 25 28 (0.61,0.39) 6.4 10.9 20 10

OLED organic light-emitting diode; CE current efficiency; CIE Commission Internationale deI’Eclairage; EQE external electroluminescence quantum efficiency; PE power efficiency

Fig. 5.9 Device structure with energy levels and exciton processes of the designed fluorescencewhite OLEDs and the molecular structures of the used organic materials in this device. Reprintedfrom [25]

5.2 Fluorescence Tandem White Organic Light-Emitting Diodes 135

with ultralow doping concentration and a sandwiched sub-EML doped with red andgreen fluorescent dyes at a relatively high concentration was designed to harness allelectrogenerated excitons and reduce the energy loss to the utmost extent. Thedevices yet remained the efficiencies as high as 16.2% and 27.2 lm/W at1000 cd/m2 luminance, which will satisfy the requirement of practical lighting withthe help of out-coupling technology. The EL performance is summarized inTable 5.3. Therefore, it is believed that the fluorescence white TOLEDs thatachieve the efficiency of fluorescent lamp must be fabricated out by the utilizationof new high-efficiency fluorescence materials and the strategic design of highlyefficient device structures.

5.3 Phosphorescence Tandem White OrganicLight-Emitting Diodes

It is well-known that phosphorescence organic molecules as emitters in OLEDs canrealize 100% internal quantum efficiency due to the harvest of all singlet excitonsand triplet excitons [26]. Therefore, using phosphorescence organic molecules asemitters should be the best way to fabricate white OLEDs in order to obtain highefficiency even though phosphorescence white OLEDs still exist the problems ofthe short lifetime of blue phosphors and the efficiency roll-off at high luminance[27]. It is encouraging that tandem structures can resolve the above problems to acertain extent because TOLEDs work at relatively lower bias voltage, and it may bemore effective for warm white OLEDs due to small amount of blue light emission.So far, the best white OLED is still made of phosphorescent materials and tandemstructures [28].

It has been shown experimentally that N/P organic heterojunctions as CGLs canenhance the power efficiency of the fabricated TOLEDs [29]. Accordingly, thewhite TOLEDs based on N/P organic heterojunctions as CGLs have been suc-cessfully fabricated [30–32] and found that the power efficiency was indeedenhanced and the efficiency roll-off also was greatly reduced compared to theconventional CGL-based white TOLEDs. Figure 5.10 shows the device structuresof the fabricated three-color phosphorescence white TOLEDs based on N/P organic

Table 5.3 Summary of EL performance of the designed fluorescence white OLEDs. Reprintedfrom [25]

Von

(V)EQE/CE/PE (%/cd/A/lm/W) CIE(x,y) CRI

Maximum At 500 cd/m2 At 500 cd/m2

DeviceW1

2.8 13.6/33.5/33.6 13.0/32.2/24.7 12.2/30.1/21.0 (0.306,0443) 72

DeviceW2

2.8 14.2/32.8/34.7 13.9/32.0/25.1 12.9/29.9/20.8 (0.314, 0434) 76

Device W 2.8 18.2/40.9/44.6 17.2/38.9/32.2 16.2/36.4/27.2 (0.318,0.390)

82


heterojunctions as CGLs, including undoped C60/pentacene (Fig. 5.10a) and dopedTPBi:Li2CO3/TCTA:MoO3 heterojunctions (Fig. 5.10b). In the devices, the EMLsare stacked by a green–red phosphorescence unit and a blue phosphorescence unitvia CGL.

Figure 5.11 shows the EL characteristics of both white TOLEDs based on C60/pentacene and TPBi:Li2CO3/TCTA: MoO3 organic heterojunctions as CGLs. Asshown in the inset of Fig. 5.11a, device (b) has a turn-on voltage of 5.5 V and theoperational voltages of 6.7 V at 100 cd/m2 and 8.3 V at 1000 cd/m2. However, theturn-on voltage is reduced to 5.1 V and the operational voltage of 6.0 V at100 cd/m2 and 6.9 V at 1000 cd/m2 as using C60/pentacene organic heterojunctionas CGL in device (a). This indicates that C60/pentacene organic heterojunction CGLshould have better electrical property than TPBi:Li2CO3/TCTA:MoO3 organicheterojunction CGL. As a result, the maximum CE, EQE, and PE of device(a) reach 101.5 cd/A, 45.7%, and 53.8 lm/W, respectively. As seen, although theTPBi: Li2CO3/TCTA: MoO3 as CGL indeed also leads to the high CE and EQE,

Fig. 5.10 Schematic diagrams of the fabricated white TOLEDs based on N/P organic hetero-junctions as CGLs. a Undoped C60/pentacene organic heterojunction and b doped TPBi:Li2CO3/TCTA:MoO3 organic heterojunction. Reprinted from [30]

Fig. 5.11 a Power efficiency versus luminance characteristics. Inset: J-V-L characteristics.b Current and external quantum efficiency versus luminance characteristics. Inset: EL spectradetected at the luminance of 100 and 1000 cd/m2. Reprinted from [30]

5.3 Phosphorescence Tandem White Organic Light-Emitting Diodes 137

which reach 98.6 cd/A and 44.9%, respectively, indicating an effective CGL aspreviously reported, but the maximum PE of 48.7 lm/W is low due to the highoperational voltage. More importantly, it can be seen that the utilization of C60/pentacene organic heterojunction as the CGL significantly improves the efficiencyroll-off of the fabricated white TOLEDs. The efficiency of device (a) is slightlyreduced to 101 cd/A, 45.5%, and 53 lm/W and 99.9 cd/A, 45%, and 45 lm/W at100 cd/m2 and 1000 cd/m2, respectively. The power efficiency roll-off values areonly 1.5% at 100 cd/m2 and 16% at 1000 cd/m2. Comparatively, the serious effi-ciency roll-off is given in device (b) based on TPBi: Li2CO3/TCTA: MoO3 CGL, asseen, the power efficiency is decreased to 34 lm/W at 1000 cd/m2 with the roll-offvalue of 24%. As demonstrated [29], the effective charge transfer from pentaceneto C60 will result in the accumulation of holes on the P-side and electrons on theN-side in the C60/pentacene organic heterojunction CGL. The formation of theaccumulation-type space charge region not only supplies the amounts of chargesused to recombine, but also leads to the formation of a high conductance region,thus greatly reducing the voltage drop in the CGL. The generation of amounts ofcharges in C60/pentacene organic heterojunction CGL should also be favor toimproving the charge balance, thus reducing the efficiency roll-off. However, thedoped TPBi: Li2CO3/TCTA: MoO3 organic heterojunction does not have theseproperties as C60/pentacene organic heterojunction, the formation of depletionjunction , and high resistance region at TPBi: Li2CO3/TCTA: MoO3 interface leadsto efficiency reduction and severe roll-off. The inset of Fig. 5.11b shows the ELspectra of two white TOLEDs at 100 and 1000 cd/m2. They all show better whiteemission, but the similar EL spectra indicate that the utilization of C60/pentaceneorganic heterojunction does not lead to additional optical interference and micro-cavity effect.

HAT-CN/TAPC organic heterojunction has proven to also be the function ofexcellent CGLs. Figure 5.12 shows the schematic structure diagram of the proposedphosphorescent three-color two-stack white TOLEDs based on N/P organic

Fig. 5.12 Schematic structure diagram of the proposed phosphorescent three-color two-stackwhite TOLEDs based on HAT-CN/TAPC organic heterojunction as CGL. Reprinted from [11]


heterojunction as CGL and the material structures that were used [11]. In the whiteTOLEDs, the white emission was realized by the stack of a blue emission unit andyellow–red emission unit connected by HAT-CN/TAPC organic heterojunctionCGL. The yellow–red unit was designed as a double EML structure to effectivelyconfine the excitons, and the blue emission unit was located far from the cathodedue to the higher light out-coupling for all the three colors. The structure of arubidium carbonate (Rb2CO3) doped 1,3,5-tri[(3-pyridyl)-phen-3-yl] benzene(TmPyPB) as ETL and a thin metal layer (Al) near HAT-CN layer is to enhance theelectron injection from CGL. For tandem device structures, as we know, by locatingEML units at their antinode position in terms of radiance intensity, where theradiance intensity has the highest value, the device efficiency could be maximized.According to this design principle, thus the device 3 with optimized structure ofITO/TAPC (45 nm)/DACTA:6% FIrpic (15 nm)/TmPyPB (15 nm)/TmPyPB:30%Rb2CO3 (25 nm)/Al (1 nm)/HAT-CN (15 nm)/TAPC (45 nm)/Bepp2:20% Ir(tptpy)2(acac) (5 nm)/Bepp2:5% Ir(mphmq)2(tmd) (10 nm)/TmPyPB (65 nm)/LiF(1.5 nm)/Al (100 nm) was fabricated. Figure 5.13 shows the normalized EL spectraof the fabricated Device 3. It can be seen that the Device 3 emits almost the sameEL spectra as that in the calculated one. Two peaks, respectively, at 472 nm and612 nm are clearly observed, forming a good white emission with CRI of 71. Asshown in Fig. 5.14, the white TOLEDs show a maximum power efficiency of

Fig. 5.13 Normalized ELspectra of the fabricatedDevice 3, 4, and 5. Thecalculated spectrum is alsogiven. Reprinted from [11]

Fig. 5.14 PE-L characteristics of thefabricated Device 3, 4 and 5.The corresponding L-V characteristics are shown inthe inset. The correspondingcalculation is also given.Reprinted from [11]


42.3 lm/W and remain at 38.2 lm/W at 1000 cd/m2 luminance. Of course, the ELspectra, thus CRI, could be changed as locating the EML units slightly far fromtheir antinode position in terms of photonic mode density, such as Device 4 with120 nm TAPC layer and Device 5 with 120 nm TAPC layer and 15 nm yellow/redEML, where their blue emission intensity is greatly reduced, thus the CRI isincreased to 82 and 85, respectively. However, there is some efficiency loss from38.6 to 32.5 lm/W (Device 4) and 29.4 lm/W (Device 5) due to the inevitableelectrical and optical losses.

Figure 5.15 shows a schematic structure diagram of the fabricated phospho-rescence white TOLEDs based on HAT-CN/TAPC organic heterojunction as CGLdeveloped by Kim et al. [10]. The white emission is obtained by the combination ofan orange emission unit and a blue emission unit. The orange unit is formed bycodoped red Ir(mphmq)2(tmd) and green Ir(ppy)2(tmd) phosphorescent dyes withhorizontally oriented transition dipoles into an exciplex-forming mCP:B3PYMPMcohost, whereas the blue unit is the exciplex-forming cohost of mCP andB3PYMPM doped with FIrpic. Thus, a high-efficiency white TOELD with amaximum EQE of 54.3% without any out-coupling enhancement structures and anEQE of 90.6% at 1000 cd/m2 by attaching an index-matched glass half sphere ontothe glass substrate was well fabricated by optimizing the device structure using anoptical simulation to maximize the out-coupling. In the devices, the position oforange and blue units, and the total layer thickness were determined by using thesimulation based on the classical dipole model. Finally, white TOLED I with 20 nm

Fig. 5.15 Schematicstructure diagram of thefabricated phosphorescencewhite TOLEDs based onHAT-CN/TAPC organicheterojunction as CGLdeveloped by Kim et al.Reprinted from [10]


HAT-CN and 40 nm HTL thicknesses and white TOLED II with 40 nm HAT-CNand 80 nm HTL thicknesses were obtained well.

Figure 5.16a shows the EL efficiencies of the resulting white TOLED I and II,showing maximum EQEs (PEs) of 46% (49 lm/W) and 54.3% (63 lm/W),respectively. It can be seen that the efficiency roll-offs were also low with EQEs(PEs) of 43.7% (41 lm/W) and 52.6% (52 lm/W) at 1000 cd/m2 and 32.9 and 42%at 10,000 cd/m2 for white TOLED I and II, respectively. An EQE of 90.6% and PEof 86 lm/W at 1000 cd/m2 were achieved by attaching an index-matched glass halfsphere onto the glass substrate. The efficiency value of white TOLED II with threeprimary emission colors without light extraction structure should be the highest onereported so far. The normalized EL spectra of the white TOLED I and II at a currentdensity of 1 mA/cm2 are shown in Fig. 5.16b. The CIE coordinates were (0.327,0.391) and (0.359, 0.500) for the white TOLED I and II, respectively. The corre-sponding correlated color temperatures (CCT) and CRI were 5685 K and 71 forwhite TOLED I, and 4989 K and 63 for white TOLED II. It can be seen that thedevice performance are well changed by varying the device thickness parameters.

It can be seen that the electron injection exists larger barrier from the CGL into theETL of emission unit when using HAC-CN as N-type layer in HAT-CN-basedorganic heterojunction CGL due to its extremely low LUMO [33]. Liao et al.designed a new three-layer CGL consisted of tetrafluoro-tetracyanoquinodimethane(F4-TCNQ)/HAT-CN/NPB and used it to successfully fabricate high-efficiencyphosphorescence white TOLEDs [9]. Figure 5.17 shows the schematic energy leveldiagram of F4-TCNQ/HAT-CN/NPB heterojunction CGL. It can be seen that theformation of an efficient gradient barrier will greatly enhance the electron injection.The fabricated white TOLEDs is connected by the two same white light units ofITO/HAT-CN (10 nm)/TAPC (45 nm)/SSTF:FIrpic 15 vol.% (19 nm)/SSTF:PO-019 vol.% (1 nm)/TmPyPB (40 nm)/Liq (2 nm)/Al via the F4-TCNQ/HAT-CN/NPBheterojunction CGL. Figure 5.18 shows the current and power efficiencies. It can beseen that the tandem device achieves a current efficiency of 152.3 cd/A and a powerefficiency of 61.3 lm/W at 1000 cd/m2, which was much higher than that of the

Fig. 5.16 a EL efficiencies and b EL spectra at 1 mA/cm2 current density of the resulting whiteTOLED I and II. Reprinted from [10]


single-unit device (65 cd/A, and 52.3 lm/W), not only is the current efficiency overtwo time, but also is the power efficiency improved. This indicates that the F4-TCNQ/HAT-CN/NPB heterojunction possesses an excellent charge injection andtransport capability without generating additional barriers.

5.4 Fluorescence/Phosphorescence Hybrid Tandem WhiteOrganic Light-Emitting Diodes

As we know, although the phosphorescence white TOLEDs have the advantages ofhigh efficiency, the short lifetime of blue phosphorescence emitters greatly still limitstheir applications. At present, a relatively better method is the fluorescence/phosphorescence hybrid tandem structures, from which the white emission can be

Fig. 5.17 Schematic energylevel diagram of F4-TCNQ/HAT-CN/NPBorganic heterojunction CGL.Reprinted from [9]

Fig. 5.18 Current and powerefficiency characteristics ofthe fabricated white TOLEDsbased on F4-TCNQ/HAT-CN/NPBorganic heterojunction asCGL. Reprinted from [9]


realized by stacking a green/red (or orange) phosphorescence unit and a bluefluorescence unit. Since the approach avoids the utilization of a relatively unstableblue phosphor, yet still realizes the harvesting of both singlet and triplet excitons,thus high efficiency becomes possible [34]. Obviously, the hybrid tandem devicesshould be intrinsically superior to the phosphorescence and fluorescence tandemdevices when the efficiency and lifetime are simultaneously required in lightingapplications.

The basic condition of a hybrid white TOLED is that one of the EL units is madeup of fluorescence blue emitter. Therefore, the fluorescent blue EL unit is critical indetermining color, efficiency, and lifetime of the hybrid white devices. Emphasiswould be still needed to be put on the advanced research of fluorescentblue-emitting materials in order to further improve the performance of the hybridwhite TOLEDs. Figure 5.19 shows the schematic structure diagram of a hybridwhite TOLED [7]. Wherein, the yellow EL unit (LEL-2) is a phosphor incorporatedwith both a green phosphorescent dopant and an orange phosphorescent dopant,while the blue EL unit (LEL-1) is a fluorophor consisted of a blue fluorescentdopant doped fluorescent host. The blue fluorescent EL unit LEL-1 is formed closeto the ITO anode side because the plasmon-quenching effect from the Al cathode onthe blue emission could be avoided with this arrangement in order to obtain a

Fig. 5.19 Schematic structure diagram of a hybrid white TOLED. a EL spectra and b operationalstabilities of the hybrid device. The inset shows the current efficiency–current densitycharacteristics. Reprinted from [7]

5.4 Fluorescence/Phosphorescence Hybrid Tandem White … 143

maximal blue emission. The intermediate connector then used an HAT-CN/P organic heterojunction as CGL. In this study, two different fluorescent bluedopants, BDM1 (Devices A-1) and EK-BD9 (Devices A-1), were used in order toinvestigate the effect of the blue EL unit on the overall EL performance of thewhole device. As shown in Fig. 5.19a, b, they emit different blue spectra, effi-ciencies, and lifetimes. Figure 5.20 shows the EL performance of the fabricatedhybrid white TOLEDs based on two different blue fluorophors. It can be seen thatwhen the first EL Unit-BDM1 is used, Device C-1 can have a relatively higher blueemission in the white spectrum, resulting in CIE (0.33, 0.35). However, when thefirst EL Unit-EK-BD9 is used, Device C-2 is lack of enough blue emission in thespectrum, resulting in CIE (0.34, 0.40). And Device C-2 exhibits lower drivevoltage, higher luminous efficiency, higher power efficiency, and longer operationallifetime than Device C-1, confirming the importance of efficient blue fluorophorselection for the construction of high-performance hybrid white TOLEDs.

Based on the basic structure of the above hybrid white TOLEDs, by introducingan internal extraction structure (IES), the efficiency of the hybrid white TOLEDshas been further improved [1]. Figure 5.21 depicts the schematic diagrams of thefabricated hybrid white TOLEDs with and without extraction structure. In thedevices, the CGL still utilized N/P organic heterojunction. The fluorescent blue ELunit had a single-doped light-emitting layer, whereas the phosphorescent EL unithad two light-emitting layers, where the first one was doped with a green-emitting

Fig. 5.20 a EL spectra, b current density–voltage characteristics, c current efficiency–currentdensity characteristics, and d operational stabilities of the hybrid white TOLEDs containingdifferent fluorescent blue EL units. Reprinted from [7]


dopant and the second one with a red-emitting dopant. Thus, the desired white lightwas achieved by properly optimizing the concentration and thickness of thelight-emitting layers. Figure 5.22 shows the EL spectra of three white devices at1 mA/cm2 current density. The corresponding performance parameters are sum-marized in Table 5.4. The power efficiency was enhanced to 56.0 lm/W at1000 cd/m2 luminance due to IES, and CRI is as high as 83.6. The color of the IESdevice at (0.387, 0.389) was well within the Energy Star quadrangle for 4000 K. Itcan be seen that the emission from the blue fluorophor is observed below 500 nmand that of the green and red phosphors peak at 545 nm and 590 nm, respectively.

Fig. 5.21 Schematic diagrams of the fabricated hybrid white TOLEDs with and withoutextraction structure. Reprinted from [1]

Fig. 5.22 EL spectra of threehybrid white TOLEDs withand without extractionstructure at 1 mA/cm2 currentdensity. Reprinted from [1]


The enhancement of light over the entire wavelength range is due to the extractionrole of EES and IES, slightly changing the emission color, thus CCT and CRI.

Figure 5.23 depicts the schematic structure diagram of hybrid white TOLEDswith three-emission units designed and fabricated by LG [34]. In the hybrid tandemdevices, one fluorescent blue EL unit is sandwiched between two phosphorescentorange EL units. Between them are the N/P organic heterojunction CGLs, whereLG101 is used as the P-type layer material and also key component for devices. Toimprove the efficiency, both the external light extraction film at the glass/airinterface and the internal light extraction layer at the glass/TCO interfaces are used.The internal light extraction layer of nanosized particles is used to destruct thewaveguide mode of the emitted light and thus enhances the light output fromorganic layer to the glass. The device size was 100 � 100 mm2, and thelight-emitting area was 90 � 90 mm2. Figure 5.24a, b shows the EL spectra andlifetime of the fabricated hybrid white TOLEDs with three different internal lightextraction materials, respectively. The EL performance is summarized in Table 5.5.

Table 5.4 Comparison of performance parameters of fabricated hybrid white TOLEDs with andwithout extraction structure. Reprinted from [1]

mA/cm 2 cd/m 2 EQE% cd/A CIE-x CIE-y V Im/W CCT CRI EQE/NES

NES 1 453 21.5 45.3 0.380 0.392 5.7 24.8

EES 1 795 37.0 79.5 0.383 0.378 5.7 43.9 3865 81.1 1.72

IES 1 1022 49.2 102.2 0.387 0.389 5.7 56.0 3836 83.6 2.29

Fig. 5.23 Schematic diagramof the hybrid white TOLEDswith three-emission unitsfabricated by LG. Reprintedfrom [34]


The operating voltage is 8.1 V, and the current density is 1.2 mA/cm2. The powerefficiency reaches 105.7 lm/W at CCT around 2800 K and 3000 cd/cm2 luminance.The devices also show very long lifetime, LT70 about 40,000 h, meeting theapplication requirements.

5.5 Applications of Tandem White OrganicLight-Emitting Diodes in Display and Lighting

Due to the advantages of high efficiency (over 100 lm/W) and long lifetime(40,000 h), white TOLEDs have exhibited great applications in the fields of displayand lighting. As the mainstream application in display, LG Display has developedover 77-in. OLED TV using their advanced hybrid white TOLEDs technology.Figure 5.25 shows the 77-in. OLED TV developed by LG Display. The large sizeOLED TV is the pinnacle of technological achievement and a new paradigm thatwill change the dynamics of the next generation TV market.

The method which produces colors through the use of RGB color filters is called“color filter method” (Fig. 5.26). With this method, the white OLEDs includingRGB color elements are then filtered to obtain the desired colors. Obviously, themethod is very simple. However, because the filters used with this method absorbmost of the emitted light energy, it is required that the white light has to be

Fig. 5.24 a EL spectra and b lifetime of the fabricated hybrid white TOLEDs with three differentinternal light extraction materials. Reprinted from [34]

Table 5.5 Summary of ELperformance of the hybridwhite TOLEDs withthree-emission units.Reprinted from [34]

IELstructure

EQE(%)

P.E.(lm/W)

CCT Duv CRI

IEL-1 114 97.3 2828 0.0020 85

ILE-2 117 99.9 2785 0.0021 85

ILE-3 123 105.7 2823 0.0024 85


relatively strong. Thus, there exist some disadvantages of low color purity andcontrast; meanwhile, the power consumption is also higher.

As lighting sources, white OLEDs have the features of green, environmentalprotection, and energy conservation and also have the advantages of flat light, softlight, no glare, no shadow, no ultraviolet and blue light damage, color temperatureadjustable, thin, large area, flexible, and transparent. It can be seen that whiteOLEDs are really green and healthy lighting sources, which will be to becomemainstream products in the future. It is foreseeable that white OLEDs will bewidely used in interior lighting, decorative lighting, medical lighting, automotivelighting, museum lighting, and other fields, creating enormous economic values. Atpresent, LG Display has developed different size OLED lighting panels (Fig. 5.27)based on their advanced hybrid white TOLEDs technology and used them tofabricate protected eye lamps and library’s reading lights (Fig. 5.28) [35].

Fig. 5.25 77-in. OLED TVdeveloped by LG display

Fig. 5.26 Color filter methodto obtain color display bywhite light


And in the future, the ability to fabricate OLED devices on flexible substrates,particularly plastic, will soon lead to a whole new range of highly innovativeproducts. For displays, as shown in Fig. 5.29, this will lead to bendable, foldable,

Fig. 5.27 OLED lighting panels with differeent sizes developed by LG Display

Fig. 5.28 OLED-protected eye lamps and reading lights fabricated by LG Display

5.5 Applications of Tandem White Organic Light-Emitting … 149

and rollable displays over the next few years, resulting in products that could not bemade or conceived using any other display technology. Highly differentiatedlighting products will also be enabled by plastic OLEDs. An example of new typeof flexible lighting is shown in Fig. 5.29. Flexible surfaces will emit gentle light, asopposed to high intensity bulbs for traditional lighting. Conventional lightingrequires diffuser and lamp shade to hide the glare of the bulb, while the flexibleOLEDs themselves can become the lamp.

References

1. Y.S. Tyan, Y.Q. Rao, X.F. Ren, R. Kesel, T.R. Cushman, W.J. Begley, in SID 09 DIGEST,p. 895 (2009)

2. Z. Shen, P.E. Burrows, V. Bulovic, S.R. Forrest, M.E. Thompson, Science 276, 2009 (1997)3. M. Fröbel, T. Schwab, M. Kliem, S. Hofmann, K. Leo, M.C. Gather, Light Sci. Appl. 4, e247

(2015)4. H.K. Kim, D.G. Kim, K.S. Lee, M.S. Huh, S.H. Jeong, K.I. Kim, Appl. Phys. Lett. 86,

183503 (2005)5. H. Becker, S.E. Burns, N. Tessler, R.H. Friend, J. Appl. Phys. 81, 2825 (1997)6. C.W. Joo, J. Moon, J.H. Han, J.W. Huh, J. Lee, N.S. Cho, J. Hwang, H.Y. Chu, J.I. Lee, Org.

Electron. 15, 189 (2014)7. L.S. Liao, X.F. Ren, W.J. Begley, Y.S. Tyan, C.A. Pellow, in SID Symposium Digest, vol. 39,

p. 818 (2008)8. Q. Wang, J.Q. Ding, Z.Q. Zhang, D.G. Ma, Y.X. Cheng, L.X. Wang, F.S. Wang, J. Appl.

Phys. 105, 076101 (2009)9. L. Ding, Y.Q. Sun, H. Chen, F.S. Zu, Z.K. Wang, L.S. Liao, J. Mater. Chem. C 2, 10403

(2014)10. S.H. Lee, H. Shin, J.J. Kim, Adv. Mater. 26, 5864 (2014)11. Y.H. Son, S.K. Kim, J.H. Kwon, J. Inf. Display 15, 185 (2014)12. G.F. He, C. Rothe, S. Murano, A. Werner, O. Zeika, J. Birnstock, J. SID 17, 159 (2009)13. Y.H. Son, J.M. Lee, B.Y. Kang, J.H. Kwon, in SID 2015 DIGEST, vol. 38, p. 561 (2015)14. S. Jang, Y. Lee, M. Park, in SID 2015 DIGEST, vol. 46, p. 661 (2015)15. Q. Wang, J.Q. Ding, D.G. Ma, Y.X. Cheng, L.X. Wang, F.S. Wang, Adv. Mater. 21, 2397

(2009)16. H. Cho, C.W. Joo, J. Lee, H. Lee, J. Moon, J.I. Lee, J.Y. Lee, Y. Kang, N.S. Cho, Opt.

Exp. 24, 24162 (2016)

Fig. 5.29 Examples of novel display and lighting concept enabled by flexible white OLEDs


17. S. Chhajed, Y. Xi, Y.-L. Li, Th Gessmann, E.F. Schubert, J. Appl. Phys. 97, 054506 (2005)18. P.C. Yeh, S.W. Hwang, J.W. Ma, M.H. Ho, S.F. Hsu, C.H. Chen, in SID 07 DIGEST, p. 871

(2007)19. T.K. Hatwar, J.P. Spindler, in SID 08 DIGEST, p. 814 (2008)20. J.P. Spindler, T.K. Hatwar, J. SID 17, 861 (2009)21. H. Uoyama, K. Goushi, K. Shizu, H. Nomura, C. Adachi, Nature 492, 234 (2012)22. D.D. Zhang, L. Duan, C. Li, Y.L. Li, H.Y. Li, D.Q. Zhang, Y. Qiu, Adv. Mater. 26, 5050

(2014)23. H. Nakanotani, T. Higuchi, T. Furukawa, K. Masui, K. Morimoto, M. Numata, H. Tanaka, Y.

Sagara, T. Yasuda, C. Adachi, Nat. Commun. 5, 4016 (2014)24. T. Higuchi, H. Nakanotani, C. Adachi, Adv. Mater. 27, 2019 (2015)25. Z.B. Wu, Q. Wang, L.L. Yu, J.S. Chen, X.F. Qiao, T. Ahamad, S.M. Alshehri, C.L. Yang, D.

G. Ma, ACS Appl. Mater. Interfaces 8, 28780 (2016)26. H. Yersin, Top. Curr. Chem. 241, 1 (2004)27. C. Murawski, K. Leo, M.C. Gather, Adv. Mater. 25, 6801 (2013)28. V.I. Adamovich, P.A. Levermore, X.Xu, A.B. Dyatkin, Z. Elshenawy, M.S. Weaver, J.

J. Brown, J. Photon. Energy 2, 021202-1 (2012)29. Y.H. Chen, Q. Wang, J.S. Chen, D.G. Ma, D.H. Yan, L.X. Wang, Org. Electron. 13, 1121

(2012)30. Y.H. Chen, J.S. Chen, D.G. Ma, D.H. Yan, L.X. Wang, Appl. Phys. Lett. 99, 103304 (2011)31. Y.H. Chen, H.K. Tian, J.S. Chen, Y.H. Geng, D.H. Yan, L.X. Wang, D.G. Ma, J. Mater.

Chem. 22, 8492 (2012)32. Q. Wang, Y.H. Chen, J.S. Chen, D.G. Ma, Appl. Phys. Lett. 101, 133302 (2012)33. E. Oha, S. Park, J. Jeong, S.J. Kang, H. Lee, Y. Yi, Chem. Phys. Lett. 668, 64 (2017)34. S. Jang, Y. Lee, M. Park, in SID 2015 DIGEST, vol. 46, pp. 661 (2015)35. www.lgoledlight.com

References 151

Subject Index

AAccumulation junction, 3, 21, 27,

28, 30, 43, 104

CCapacitance-voltage characteristics, 46, 47,

118Charge Generation Layer (CGL), 30, 33, 34,

48, 55, 60, 64Charge transport, 30, 32, 51–53, 69, 78, 79,

103, 106, 117Current-voltage characteristics, 9–11, 37

DDepletion junction, 3, 98, 104, 138Display, 31, 63, 74, 78, 82, 147–149

EElectron injector, 60, 61, 70–72, 74, 75,

77, 79, 80, 85, 86Energy band profiles, 4, 13, 20Exciton, 30, 32–34, 37, 38, 86, 103, 130,

133–136, 143

FFluorescence, 33, 61, 67, 132, 133,

135, 136, 143FN tunnelling, 10–13, 35, 44–46, 51,

59, 72, 82

HHeterojunction, 1, 3, 4, 7–13, 20, 22–27,

29–31, 34, 35, 37–41, 43–48, 53–56,59–64, 68, 70–72, 74–80, 82, 85, 86,89, 92, 94, 95, 97–101, 103–111, 113,115–122, 128, 133, 136–142, 144

Hole injector, 61, 63, 64, 67, 70, 77Hybrid, 143–148

LLighting, 31, 33, 128, 133, 136, 143, 147–150

OOrganic, 1–3, 20–35, 37, 39, 40, 42, 43, 45,

47–49, 51–56, 59–64, 67–75, 77–80,82, 84–86, 89, 92, 97, 98, 100,103–106, 108, 110, 113, 115–118,121–123, 127, 134–138, 140, 146, 147

Organic Light-Emitting Diodes (OLEDs),30–35, 48, 55, 59–64, 67, 70–80, 85,86, 92, 97, 115, 118, 132, 136, 142, 147

PPhosphorescence, 33, 61, 77, 78, 80, 82, 136,

137, 140–143

SSemiconductor, 1–6, 8, 9, 11, 14, 20, 22, 23,

26, 31, 34, 35, 37, 40, 42–44, 46, 49, 51,53–56, 59–61, 64, 69, 70, 74, 85, 98, 99,105, 106, 113, 115, 117, 122, 123, 127

Singlet, 33, 133, 134, 143Space Charge Limited Current (SCLC), 53, 65,

66

TTandem, 30, 33, 34, 48, 60, 64, 89–91, 94–97,

100, 101, 105–107, 112, 119, 133, 136,139, 141, 142, 146

Triplet, 33, 134, 136, 143

© Springer-Verlag GmbH Germany 2017D. Ma and Y. Chen, Organic Semiconductor Heterojunctions and Its Application inOrganic Light-Emitting Diodes, Springer Series in Materials Science 250,https://doi.org/10.1007/978-3-662-53695-7

153

Documents

Organic semiconductor heterojunctions and its application in organic light-emitting diodes