BANDGAP ENGINEERING OF TFEL DEVICES

Abstract

We introduce an alternative concept to increase the efficiency and brightness of thin-film electroluminescent (TFEL) devices. The method utilizes bandgap engineering of the active layer of the device. First steps of our work using a ZnSxSe1−x alloy are also presented to demonstrate workability of the method. The related obstacles and future potentials of the bandgap engineering for monochrome and color TFEL devices, are discussed.

Description

[0001] The operational efficiency of today's alternating-current thin-film electroluminescent (ACTFEL) display depends on three main factors, which are the radiative efficiency, the outcoupling efficiency of the emitted photons, and the excitation efficiency of dopant centers.1 These three components of the total efficiency must be increased to meet the brightness requirements for full-color applications and to improve the operational parameters of a monochrome (typically, ZnS:Mn-based) display. The radiative efficiency is determined mainly by the recombination cross-section of the luminescent centers. A straightforward increase of the dopant concentration results in higher concentration quenching of luminescence.2 The improvement of the photon outcoupling efficiency is mainly associated with the optical matching of the individual layers of the device. Difficulties in obtaining a significant increase in the excitation efficiency of luminescent centers are generally attributed3,4 to the low efficiency of the tunneled electrons, caused partially by space charge formation and a lack of electron acceleration due to short electron mean-free path (the distances required for electrons to gain enough energy in ZnS to impact excite Mn centers and to ionize the lattice are 16 and 40 nm, respectively). This leads to a decreasing impact ionization probability with the distance from the insulator-semiconductor interface (ISI) to the middle of the phosphor layer. As a result, only the phosphor region adjacent to the cathodic ISI is efficiently involved in light emission.5 All attempts to increase the impact ionization probability of the kinetic electrons through the phosphor bandgap Eg reduction have so far shown little success, mainly due to a decreased energy of the tunneled electrons or temperature quenching problems.6

[0002] In this article we report on use of bandgap engineering to improve the ACTFEL device efficiency through increasing the impact ionization probability of kinetic electrons.

[0003] The device preparation was described in detail elsewhere.7 Briefly, two stacked dielectrics (normally, SiON as a barrier and Y2O3 or Al2O3-based insulator as an “injector”) where used to confine current in semiconductor active layer (ZnS doped with Mn in conventional devices). Opaque Al and transparent polycrystalline indium-tin-oxide cladding electrodes were used to ensure matrix addressing of the device. Upon application of the external bipolar trapezoidal waveform, the trapped electrons start tunneling from the momentary cathodic ISI into the bulk of the active layer, causing excitation of the luminescent impurity centers and ionization of the crystal lattice.

[0004] To demonstrate the bandgap engineering concept, a ZnSxSe1−x alloy doped with Mn has been chosen for the active layer of the proposed device. The layer was engineered through automated thermal co-evaporation in such a way as to provide a monotonic decrease of the bandgap from both interfaces to its middle part with precise control of the dopant concentration, which was maintained at ˜0.7%. The presence of wide bandgap ZnS at both ISIs ensures high energy of the tunneled electrons. At the same time, the impact ionization threshold energy of the semiconductor lattice is expected to decrease with bandgap from the momentary cathodic interface to the middle part of the active layer. In fact, the ionization rate &agr;i is defined as the number of electron hole pairs generated by an electron per unit time and is a function of the electric field and the ionization threshold energy Ei:8 1 α 1 = ( q ⁢ F F i ) ⁢ exp ⁡ [ - ( F 0 F ) 2 ] , ( 1 )

[0005] where F=ƒ(Ei2) and F0 is a characteristic electric field. Thus, &agr;i is expected to increase dramatically with decreasing Ei, where Ei is proportional to the bandgap energy.

[0006] ZnS (room temperature bandgap Eg=3.7 eV) and ZnSe (Eg=2.68 eV) are known to ideally match in both single crystalline and polycrystalline phases and to produce an amenable ZnSxSex−1 compound with linear dependence of almost all its properties on the alloy composition (FIG. 1a). ZnS, ZnSe, and their alloys have also very similar electronic structures.9 Mn emission curve peaks at 585 nm for any ZnSxSex−1 composition and corresponds to a 4 T1-6 A1 transition within Mn 3d5 configuration.

[0007] X-ray diffraction measurements were used to calibrate the alloy composition. Diffraction peak shifts linearly from 28.6° for ZnS(111) to 27.3° for ZnSe(111) (FIG. 1b). For quick estimation of the ZnSxSex−1 composition, threshold voltage of the flat-band device can also be used. To verify the band profile of the active layer, we used the secondary electron emission (SEE) contrast profile technique developed by Krasnov.10

[0008] It is known that for some compounds of the II-VI family, including ZnS, ZnSe, and their alloys, the energy level position of particular native or impurity defect remains roughly the same in respect to the vacuum level.11 Changes in the alloy composition, however, determine the position of the energy level in respect to the conduction and valence bands. This in turn determines the space charge distribution and the dynamics of carrier transfer inside the active layer, favoring positive space charge formation closer to the ISIs and not in the bulk of the active layer.

[0009] It should be noted at this point that the ZnSxSe1−x:Mn alloy is used in this study only in order to demonstrate the concept and may not be the best choice for a bandgap engineered TFEL device. The reasons for that are as follows. The impact cross-section &thgr; for Mn is 10−16 cm−2 in ZnS compared to 5×10−17 cm−2 in ZnSe; the reduction in &thgr; decreases the excitation efficiency of the impurity center. The luminescence decay time of Mn centers has been found to be 1 and 0.1 &mgr;s for ZnS and ZnSe, respectively, which significantly decreases the integral brightness of the device; the negative impact of this effect is less pronounced if the device is driven at high frequencies. Finally, the transmittance of the alloy decreases linearly from 0.9 to 0.7 with the Se content increase from 0 to 1. Therefore, in parallel with the ionization rate increase, the absorption coefficient increase also takes place. This is taken into account during the active layer design. FIG. 2 shows schematically the band profile of the proposed device, which is expected to enhance carrier multiplication and increase total efficiency of the device. It should be pointed out that the classical understanding of the multiplication factor12 is applicable in this particular case only to a half of the device between momentary cathodic interface and the region of the active layer with minimum bandgap. Also demonstrated in FIG. 2 is an example of SEE contrast profile of the active layer. Taking into account that the refractive index is 2.89 for ZnSe and 2.35 for ZnS, the region with the minimum bandgap was placed closer to the top (Al-side) interface of the structure, thus providing a better outcoupling of the emitted photons. The optimal alloy composition of the middle part of the active layer was found to be ZnS0.6Se0.4 with the bandgap of ˜3.3 eV.

[0010] Despite of all the listed drawbacks of the ZnSxSe1−x:Mn alloy, a ˜2 times brightness increase of the new device compared to a non-bandgap-engineered structure has been achieved. The explanation is evident from FIG. 3, which demonstrates the results of the probe layer analysis, indicating much less significant brightness decrease of the probe layer (a thin Mn-doped layer in otherwise undoped sample) with distance from the momentary cathodic interface of the new device. This in turn suggests higher electron excitation efficiencies; more luminescent centers can be excited by the increased number of electrons at the same amount of the delivered energy. Poor Mn emission in ZnSe, therefore, is compensated with higher excitation efficiency of the kinetic electrons.

[0011] Theoretical modeling and preliminary experiments showed the combination of other materials such as CaS, SrS, MgS, BaS, and ZnO along with ZnS may be preferable to maximize the “bandgap” effect. For these materials a steeper variation of the ionization rate with the electric field (and, in turn, with Ei) is expected.13 Also, in addition to Mn, experiments to use bandgap engineering for the efficiency increase of alternate dopants, such as Ho, Ce, Pr, etc., were started. The goal of these experiments is to achieve acceptable brightness and efficiency of the device for white and full-color applications. The related results will be discussed in separate reports.

[0012] In conclusion, a concept of TFEL device utilizing the bandgap engineering concept has been demonstrated. The suggested active layer bandgap profile allows to significantly increasing the number of electrons hot enough to sustain carrier multiplication in the bulk of the phosphor layer. Despite of some inherent drawbacks of the ZnSxSe1−x:Mn system, a ˜2 times of brightness increase has been achieved in the proposed device. The proposed concept is likely to have far-reaching technological and commercial effects for both monochrome and full-color TFEL devices.

BRIEF DESCRIPTION OF THE DRAWINGS

[0013] FIG. 1. Dependence of parameters on ZnSxSe1−x alloy composition.

[0014] FIG. 2. Band profile of the bandgap engineered device (a) and its SEE contrast profile pattern (b).

[0015] FIG. 3. Luminance intensities of the conventional and bandgap engineered devices depending on the probe layer position.

Claims

1. An improved alternating current thin-film electroluminescent display having two stacked dielectrics, a semiconductor active layer therebetween, and metallic cladding electrodes at each side thereof;

wherein the semiconductor layer is developed by automated thermal co-evaporation so as to provide a monotonic decrease of the band gap thereof from the respective interfaces with said stacked dielectrics to the middle of said semiconductor active layer so that the dopant concentration thereof is maintained at about 0.7%

2. The improved alternating current thin-film electroluminescent display of claim 1, wherein said two stacked dielectrics are SiON as a barrier and either a Y2O3 or Al2O3-based insulator as an injector;

wherein said metallic cladding electrodes are opaque Al and transparent indium-tin-oxide, respectively; and

wherein said semiconductor active layer is a ZnSxSe1−x alloy doped with Mn.