Chapter 2 of this dissertation starts from the process on dark current induced by photo-excited below bandgap. Auger exciton dissociation through deep level defects was proposed as the possible mechanism. The inverse process of Auger process, impact ionization, was calculated also. A feasible electrolu-minescence for unipolar light-emitting diode based on the impact ionization process is presented[60].
Chapter 3 presents a method to improve electroluminescence efficiency of conjugated polymer LED theoretically. Doping magnetic complexes into polymer LED can turn on the transition channel between radiative singlet exciton and nonradiative triplet exciton. With suitable concentration for the doped magnetic complexes, it is possible for the singlet exciton ratio being up to 80% higher than 25% for the spin independent value[61].
Chapter 4 shows device simulation for multilayer organic light-emitting diode. Some structures are shown to have possibility to tune the lumines-cence color from red, green, then to blue as bias voltage increasing. These structures provide some possible solutions for full-color display with organic material other than the common technology by ink-jet printing method[62].
1.4. THE STRUCTURE OF THIS DISSERTATION 19
Table 1.1: Device specifications for the luminescence materials of OLED[8].
20 CHAPTER 1. INTRODUCTION
Table 1.2: Device specifications for the phosphorescence materials of OLED[8].
1.4. THE STRUCTURE OF THIS DISSERTATION 21
B CIE (0.16,0.14) (0.22,0.27) (0.15,0.12)
10,000 hrs
Table 1.3: Device specifications for the luminescence materials of PLED[8].
22 CHAPTER 1. INTRODUCTION
Chapter 2
Defect Auger exciton dissociation and impact ionization in conjugated polymers
23
24CHAPTER 2. EXCITON DISSOCIATION AND PHOTOCONDUCTIVITY
2.1 Exciton dissociation and generation by deep level defects
The past ten years has witnessed a tremendous progress in both the science and technologies of light-emitting devices based on conjugated polymers[1].
Yet many fundamental questions regarding the two single important prop-erties, electroluminescence (EL) and photoconductivity (PC), remain unan-swered. The defects in the polymer chain, either structural or chemical, are believed to play an important role in both EL and PC. The deep electronic levels associated with the defects provide a convenient way to facilitate the dissociation of the exciton, and limit the luminescence quantum yield in EL. On the other hand, excitons must be dissociated in order to produce charge carriers for PC for excitation below the continuum threshold[63]. Even though the enhancement of PC and reduction of EL by oxidation, presum-ably due to exciton dissociation at the carbonyl defects, has been reported experimentally[24], the microscopic mechanism which controls the dissocia-tion rate is not well understood.
A new exciton dissociation mechanism, the defect Auger process, is stud-ied in this work. In this process the electron (hole) in the exciton drops into the empty (occupied) deep level while the hole (electron) is released by Coulomb scattering and becomes a free charge carrier with high kinetic en-ergy as required by enen-ergy conservation. The corresponding Coulomb matrix element is shown in Fig. 2.1(a) and Fig. 2.1(b). The defect Auger process for exciton is in sharp contrast with the usual free carrier Auger process, which occurs only at high carrier concentrations because the relaxation energy of one free carrier is carried away by the kinetic energy of another nearby free carrier. Therefore the Auger rate usually depends strongly on the free car-rier density and consequently the excitation level. On the other hand, in conjugated polymers the electron-hole pair remains bound to form exciton even at room temperature. So when one of the carrier relaxes there is always another oppositely charged carrier nearby to carry away the relaxation en-ergy. In other words, each exicton can act alone and the dissociation rate is independent of the exciton density. This unique mechanism is expected to be quite efficient because the effective carrier distance, the exciton Bohr ra-dius, is very small compared with the mean distance among the excited free carriers. If we use the material parameters suitable for poly(para-phenylene vinylene)(PPV) and assume, as in the case of inorganic semiconductor, that each exciton samples the average defect density by interacting with many defects within its lifetime (the volume dissociation regime), our calculation shows that the rate is of the order of 1016s−1 times the number of defect per
2.1. EXCITON DISSOCIATION 25
Figure 2.1: (a)Diagram for the direct Coulomb scattering term in which one conduction electron (c, ke, s) is captured by defect (d), while one free valence electron (v, −kf h, s0) is scattered to (v, −kh, s0). k is the wave number, and s is the spin index. (b)Diagram for the exchange Coulomb scattering term in which one valence electron (v, −kf h, s) is captured by defect, while one con-duction electron (c, ke, s0) is scattered to the valence band state (v, −kh, s0).
repeat unit, which is expected to be no less than 10−3. Such a high rate is three orders of magnitude faster than the more common multi-phonon emission process[63]. Moreover, it can happen even at zero temperature because no energy barrier is present, consistent with the sweep-out regime experiment[64]. The defect Auger process is therefore identified as the pri-mary microscopic origin for the photocarrier generation and luminescence quenching in conjugated polymers. The calculated dissociation rate can not, however, be used naively to obtain the PL and PC yield quantitatively. For example, the dissociation rate is in the order of 1013s−1 with defect density equal to one per 400 repeat unit. The corresponding non-radiative lifetime would be around 0.1 picosecond (ps). This value is four order of magnitude shorter than the radiative lifetime of the excitons, and implies that the light emission would be completely quenched if the decay were in the volume dis-sociation regime. This is, however, inconsistent with the experiment that the PL yield is reduced to only half at such defect density.[65] The reason is that the exciton dissociation process is not in the volume capture regime, in
26CHAPTER 2. EXCITON DISSOCIATION AND PHOTOCONDUCTIVITY which each exciton encounters many defects before decay and a uniform exci-ton density is maintained throughout the system volume. Instead, the decay is in the diffusion regime[65], in which the excitons do not have the chance to sample the average defect density but are immediately quenched by the first defect they hit along the path of their diffusive motion in the chain. In this case, the deep levels act as a black hole and no exciton can pass through it. Unlike the volume dissociation regime, in the diffusion regime the steady state exciton density is not uniform along the chain but vanishes at the defect positions. The decay dynamics of the total number of excitons, controlled not only by the the transition matrix element but also the diffusion coeffi-cient of the excitons, is therefore not a simple exponential. We confirm this picture by calculating the dissociation probability of one single passage of the exciton through the defect with arbitrary incident velocity. The result is indeed close to one for excitons with thermal velocity.
In addition to the Auger process, we also study the rate of its reverse process, the defect impact ionization, by slightly modifying the calculations.
Interestingly, in defect impact ionization the incident hot hole can kick out the electron in the deep level and form a neutral exicton with itself when the incident kinetic energy reaches the threshold. The number of charge carriers is reduced from one to zero, in sharp contrast with the usual impact ionization for which the number of carriers multiplies and causes avalanche breakdown eventually. If the kinetic energy of the incident hot hole is increased further, it becomes possible to create a free electron-hole pair and the number of carriers multiplies as usual. In this circumstance the channel for carrier decrease (exicton production) and increase (free pair production) compete.
Impact ionization coefficient to neutral exciton is found to be around 108/cm times the number of defect per repeat unit when holes are driven by the electric field around 105 V/cm. Exicton production by impact ionization opens the possibility of light emission under unipolar charge injection.
In section 2.2, the defect Auger dissociation rate for exciton as a func-tion of the incident exciton momentum is calculated. The matrix element is derived in Appendix A. In section 2.3, the rate for defect impact ionization as a function of the incident hot hole momentum is calculated. Two possible final states, the exciton (2.3.1) and the free electron-hole pair (2.3.2), with different impact thresholds are considered. Averaged impact ionization coef-ficient for holes under high electric field is calculated in 2.3.3 We discuss and conclude in section 2.4 and 2.5, respectively.