ZnS LUMOCEN (luminescence from molecular centers) films typically 2000-Å thick doped with rare-earth and transition-metal fluorides have been fabricated by vacuum coevaporation. Details of the fabrication techniques are presented for the 12 fluorides studied: PrF3, NdF3, SmF3, EuF3, TbF3, DyF3, HoF3, ErF3, TmF3, YbF3, CrF3, and MnF2. Concentration dependence of the brightness has been investigated for the ZnS/TbF3 device, the maximum occurring at 1.8 at.% Tb. Studies of brightness as a function of input power indicate that the ZnS/TbF3 device is the brightest with 50 foot-lamberts readily obtained. The power efficiency for ZnS:TbF3 is about 10−4. Emission spectra of the devices have been measured from 0.35 to 1.1 μ and related to known energy level schemes where possible. The intensity distributions of the spectra generally differ from those obtained by doping ZnS with the bare ions; the ZnS:CrF3 device yields a spectrum radically different from any Cr3+ luminescence previously investigated.

In a series of experiments we have found that the field in the ZnS layer of an ac-operated thin-film high-field electroluminescent device becomes clamped at values ranging from 2.1×106 to 3×106 V/cm. The device is fabricated as a sandwich with the following layers from top to bottom: thin metal; ZnS:Tb3+; Ta1.98Mo0.02O5 and Ta0.99Mo0.01. Since the electroluminescence is expected to be very strongly field dependent because the excitation proceeds via direct impact of the Tb3+ ions by hot carriers, the observed field clamping may impose a serious limitation on the efficiency that one can obtain from such structures. Evidence for field clamping is obtained by measuring as a function of peak-to-peak applied voltage the following: (i) Tb3+ emission intensity, (ii) the ratio of emissions originating in two different Tb3+ levels and (iii) the magnitude of a pulsed photocurrent. In addition, the clamped field may be determined quantitatively through the use of a novel technique which permits the measurements of the current-field relationship in the active layer. The value of the clamped field depends only slightly on the method of preparation of the ZnS layer and is insensitive to the type of metal contact used, to the Tb concentration, and to temperature. Several mechanisms which could give rise to the steep current-field relationship responsible for field clamping are considered. One of these, multiplication by impact ionization, is ruled out by the pulsed photocurrent measurements; Frenkel-Poole emission and Schottky emission are too temperature dependent and do not exhibit sufficiently steep current-field characteristics. It is concluded that field emission at the contacts must be responsible although it is not entirely clear why it is so insensitive to the contact materials.

It is experimentally demonstrated in a new and direct manner that electroluminescence in thin films of ZnS: Tb3+ is excited by the direct impact of hot electrons on the Tb3+ ions. The scheme consists of monitoring as a function of applied voltage across the structure studied-a ZnS: Tb3+, Ta2O5 sandwich-the ratio of fluorescence intensities I(5D3)/I(5D4) originating from the 5D3 and 5D4 levels of Tb3+ at 3.25 and 2.54 eV, respectively. As the electron energy distribution is shifted to higher energies by increasing the voltage across the structure, I(5D3)/I(5D4) grows. The Tb3+ ion thus acts as a probe for the electron distribution in these films. By comparing the results obtained to simple theory, it is concluded that in considering low-energy losses of the electrons in the active layer, inelastic scattering by low-lying 7F levels of Tb3+ can not be the sole energy-loss mechanism at the Tb concentration of interest. The optical phonons of ZnS must also be taken into account.

It has been shown recently that efficient electroluminescence may be obtained from thin films formed by the. coevaporation of ZnS and rare-earth fluorides. In order to assess the importance of the fluorine ions in such systems, films have been produced by coevaporating ZnS and rare-earth metals (Tb, Sm, Nd). It is shown that large concentrations of well-dispersed rare-earth ions may also be obtained by this method. For ZnS: (Tb metal) the concentration which optimizes the electroluminescence power efficiency has about the same value as that for ZnS: TbF 3 (2 at. % Tb), and the peak electroluminescence power efficiency is slightly lower than that obtained for ZnS: TbF 3 • In general, the metal-doped systems yield electroluminescence spectra which differ from their fluoride-doped counterparts at a similar concentration. This is taken as evidence that the fluorine ions remain in the vicinity of the rare-earth ion. In the case of ZnS: EuF 3 , it is further suggested that the fluorine ions may help to control the valence of the rare-earth ion to yield Eu 3 + since independent evidence indicates that in ZnS crystals only Eu 2 + is stable. Absolute photoluminescent power efficiency measurements on films of ZnS : (Tb metal) are used to show that a mechanism for electroluminescence based on electron-hole pair formation with subsequent transfer of this energy to the Tb3+ ion may be effectively ruled out; indirect evidence for direct impact excitation of the active ion is thus obtained. For purposes of application, it appears that systems in which the rare-earth ions are introduced by evaporation of the fluorides are superior, since the evaporation of these is simpler and the electroluminescence spectra of such devices are more reproducible.

No abstract