1. TFEL Phosphor Host / Luminescent Impurity Doping Trends J. F. Wager, J.C. Hitt, B.A. Baukol, J.P. Bender, and D. A. Keszler Abstract The performance of inorganic thin-film electroluminescent displays depends critically on the phosphor employed. Three of the best-studied EL systems, ZnS:Mn, SrS:Ce, and SrS:Cu, are compared and contrasted in order to elucidate device implications inherent in selecting different types of phosphor host / luminescent impurity systems. Introduction The development of full-color thin-film electroluminescent (TFEL) displays has motivated researchers to explore a variety of phosphor host/luminescent impurity combinations. Three TFEL phosphor host/luminescent impurity combinations that have elicited the most research and development attention are ZnS:Mn, SrS:Ce, and SrS:Cu. All three of these luminescent impurities are substitutional dopants on cation sites. However, the nature of these three dopants differ dramatically: Mn is isovalent, having the same valence as the host cation, i.e., Mn 2+ on a Zn 2+ site; Ce is a donor, having a valence greater than the host cation, i.e., Ce 3+ on a Sr 2+ site; and Cu is an acceptor, having a valence less than the host cation, i.e., Cu 1+ on a Sr 2+ site. The focus of this work is to explore the implications and ramifications of employing different types of substitutional dopants as luminescent impurities in TFEL phosphors. It is shown that the dopant nature of the luminescent impurity can affect many aspects of TFEL behavior such as charge neutrality, self-compensation, defect creation and association, static/dynamic space charge creation and annihilation, trailing-edge emission, color tuning, and electroluminescence (EL) thermal quenching. [1] Figure 1. Approximate ionization energy locations for Mn in ZnS, Ce in SrS, and Cu in SrS. The charge states shown indicate Mn Zn to be a single donor, Ce Sr to be a double donor (with the second ionization state corresponding to ionization of the Ce luminescent impurity), and Cu Sr to be a single acceptor. Table 2. Luminescent impurity dopant trend summary (tbii = trap-to-band impact ionization, bbii = band-to- band impact ionization). Phosphor / iuminescent impurity Dopant type Compensating defect Associated defect complex / binding energy Dynamic space charge creation mechanism Other trends ZnS:MnIsovalentNone Static space charge SrS:CeDonor (double) V Sr (Ce Sr -V Sr ) - 0.7 eV tbiiDynamic space charge Trailing edge emission Charge collapse SrS:CuAcceptor (single) VSVS (Cu Sr -V S ) + 1.0 eV bbii + hole trappingDynamic space charge Trailing edge emission Color tuning EL thermal quenching Defect typeDefect dopant natureZnS (eV)SrS (eV) Cation vacancyDouble acceptor2.94.1 Anion vacancyDouble donor2.52.9 Cation interstitialDouble donor11.89.6 Anion interstitialDouble acceptor19.614.2 Anion-on-cation antisiteQuadruple donor9.318.5 Cation-on-anion antisiteQuadruple acceptor9.514.9 Figure 2. Association of Ce Sr and V Sr isolated point defects to form a Ce Sr -V Sr defect complex. Expected association-induced defect ionization trends are illustrated. Figure 5. An important aspect of TFEL device behavior involves positive space charge. This is intimately related to the dopant nature of the luminescent impurity. A complete description of TFEL positive space charge requires specifying both creation and annihilation mechanisms. (a) An energy band diagram of a TFEL device under bias which illustrates the three possible positive space charge creation mechanisms: field-emission, trap-to-band impact ionization, and band-to-band impact ionization with hole trapping. (b) An energy band diagram of a TFEL device near zero bias which illustrates the two possible mechanisms of positive space charge annihilation: electron emission/recombination, and hole emission. Table 1. Estimated intrinsic point defect energies for ZnS and SrS. The defect dopant nature is assessed using Kröger’s rules. Conclusions The dopant properties of ZnS:Mn, SrS:Ce, and SrS:Cu TFEL phosphors are discussed. Mn is isovalent, Ce is a double donor, and Cu is a single acceptor. The differing dopant nature of these three luminescent impurities leads to radically different defect and device physics trends. The isovalent nature of ZnS:Mn results in relatively ideal ZnS:Mn TFEL device operation, except for the presence of static space charge. In contrast, the donor (acceptor) nature of Ce (Cu) in SrS results in many different kinds of non-ideal TFEL device behaviors, i.e., dynamic space charge, trailing-edge emission, and charge collapse for SrS:Ce; dynamic space charge, trailing-edge emission, color tuning, and EL thermal quenching for SrS:Cu. The driving force for this non-ideal TFEL behavior is the high luminescent impurity doping concentration which leads to the introduction of new electronic charge states into the bandgap. References [1] J. F. Wager, J. C. Hitt, B. A. Baukol, J. P. Bender, and D. A. Keszler, J. Lumin. 97, 68 (2002). Acknowledgements This work was supported by the National Science Foundation under Contract No. DMR-0071899 and by the Defense Advanced Research Projects Agency under the Phosphor Technology Center of Excellence, Grant No. MDA 972-93-1-0030. Luminescent Impurity Doping Trends Figure 3. Cu-doped SrS as an example of vacancy- mediated self-compensation. E F crit is the critical Fermi-level position beyond which it is more energetically favorable to create self-compensating sulfur vacancies than to modulate the Fermi-level closer to the valence band maximum. (a) (b) Figure 4. Possible equilibrium energy band diagrams for a SrS:Cu phosphor. (a) Flat-band situation with the Fermi level positioned below midgap due to Cu doping. This implies that a SrS:Cu device should operate via hole injection and transport, since hole injection barriers are much smaller than corresponding electron barriers. (b) Fermi-level pinning positions the interface Fermi-level in the upper portion of the bandgap, as required for electron injection, and yet allows the bulk Fermi level to be located below midgap, due to acceptor doping. Conclusion from Table 1: Vacancies are the most likely native point defects.