Open problem originating from Fornell, Grimmeiss, Mach and Müller (Semicond. Sci. Technol. 3, 511 (1988)). Steady-state photoionisation measurements on ZnS single crystals showed that doping with about 2-5% manganese almost completely suppresses the 0.9 eV photoionisation peak of the self-activated (SA) centre, a zinc-vacancy/co-activator acceptor complex, and revealed no Mn-specific photoionisation feature in the accessible energy window, implying that any Mn-induced occupied gap states have binding energies below about 0.2 eV or above about 2 eV at concentrations of order 1016 cm−310^{16}\ \mathrm{cm}^{-3} or higher. The linked unresolved questions are: (1) whether manganese, isolated or clustered, introduces localized occupied ground states in the ZnS band gap and at what energies; (2) by what microscopic mechanism manganese suppresses the SA centre - reduced zinc-vacancy concentration, preferential Mn occupation of vacancy sites, or another structural/chemical modification; (3) which defect species provides the deep hole traps required by models of the electroluminescent memory effect (brightness-voltage hysteresis) of ZnS:Mn, given that the SA acceptor is depleted at device-relevant Mn levels. Later work (Bacherikov et al. 2017, 2018) has proposed mechanisms and indirect estimates placing the substitutional-Mn ground level about 0.35-0.6 eV above the valence band - inside the window where the 1988 measurement found no signal - but explicitly leaves the exact level position, the decomposition of the quenching mechanism, and the memory-trap identification unresolved.

Background

Zinc sulfide (ZnS) is a II-VI compound semiconductor (zinc from group II, sulfur from group VI) with a band gap of about 3.7 eV. When doped with manganese, Mn2+^{2+} ions substitute for Zn2+^{2+} and give the classic yellow-orange phosphor emission near 585 nm, arising from the spin-forbidden intra-3d-shell transition from the excited state to the ground state of the Mn2+^{2+} ion. This emission is the basis of alternating-current thin-film electroluminescent (ACTFEL) display panels, in which a polycrystalline ZnS:Mn layer is sandwiched between insulating films and driven by high alternating fields. A technologically important property of ZnS:Mn ACTFEL devices is the electroluminescent memory effect: a bistability or hysteresis in the brightness-voltage characteristic that allows passive-matrix addressing (driving a display from row and column electrodes without a switching transistor at each pixel). Marrello, Ruhle and Onton (1977) showed that the memory resides in localized bright filamentary regions of the film, and Marrello and Onton (1980) showed it depends on how the manganese is distributed in the layer. Howard, Sahni and Alt (1982) modelled the hysteresis as requiring deep hole traps - localized states that capture and store holes - at concentrations above roughly 1016 cm−310^{16}\ \mathrm{cm}^{-3}, whose stored space charge shifts the internal electric field from one voltage cycle to the next. Which microscopic defect plays that role was not known.

A second key ingredient is the self-activated (SA) centre of ZnS. In ZnS that is not deliberately doped with a metal activator but contains a co-activator such as a halogen, broad blue-green luminescence bands appear together with a characteristic deep-acceptor optical signature. Optically detected magnetic resonance (ODMR), a technique in which an electron-paramagnetic-resonance signal is detected through a change in luminescence intensity and therefore selects only the centres participating in a given emission band, assigned the SA emission to a complex of a zinc vacancy with a nearest-neighbour co-activator, for example a zinc vacancy paired with iodine on a sulphur site (Nicholls, Davies, Poolton, Mach and Muller, 1985). Lee, O'Donnell and Watkins (1982) had earlier demonstrated ODMR detection of the isolated zinc vacancy in ZnS.

The experimental method behind the source paper is deep-level photoionisation spectroscopy in the photocapacitance configuration: monochromatic sub-band-gap light is shone on a junction or high-resistivity sample held at constant capacitance, and the steady-state photocapacitance records the spectral distribution of the photoionisation cross section of occupied deep levels (hole emission from acceptor-like states to the valence band). Grimmeiss, Meijer, Mach and Muller (1984) applied such methods to ZnSx_xSe1−x_{1-x} alloys, where the SA centre produces a well-resolved photoionisation peak at about 0.9 eV above the valence-band edge.

Fornell, Grimmeiss, Mach and Muller (1988) then measured the spectral distributions of the photoionisation cross section in an undoped ZnS reference crystal and in three crystals doped with roughly 2%, 3% and 5% manganese. The 0.9 eV SA peak, prominent in the undoped reference, is almost completely suppressed in the Mn-doped crystals while the higher-energy part of the spectrum keeps its shape, so Mn doping strongly reduces the SA acceptor population. No additional photoionisation feature attributable to a Mn-specific centre appears in the measured window (photon energies up to about 2 eV, probing occupied states in the lower half of the gap), so any Mn-induced occupied levels must be very shallow (binding energies below about 0.2 eV) or deeper than about 2 eV (the upper limit of the photon-energy window) to have escaped detection. Because Mn depletes the SA acceptor, the SA centre is unlikely to supply the deep hole traps demanded by the Howard-Sahni-Alt memory model at device-relevant Mn contents, and the identity of the actual traps becomes an open question. Electron paramagnetic resonance (EPR), which detects paramagnetic defects through microwave absorption between spin levels, could not resolve the vacancy question in these samples because the intense multi-line Mn2+^{2+} resonance masks the vacancy signals. The mechanism of the suppression itself - fewer zinc vacancies, manganese sitting on the vacancy-related sites, or some other structural or chemical modification such as Mn clustering - was left undetermined.

Problem Statement

Determine the deep-level defect physics that manganese introduces into ZnS, at Mn contents relevant to electroluminescent devices, by answering the following linked questions posed by the photoionisation measurements (deep-level optical spectroscopy in which sub-band-gap light ejects carriers from occupied gap states) of Fornell, Grimmeiss, Mach and Muller (1988):

(1) Mn-induced localized states. Does manganese incorporated in ZnS (as isolated substitutional Mn or as clusters such as nearest-neighbour pairs and triplets) create localized occupied ground states in the band gap, and if so at what binding energies and concentrations? The 1988 steady-state photoionisation measurements detected no Mn-specific feature, constraining any such states to binding energies below about 0.2 eV (very shallow) or above about 2 eV relative to the valence-band edge at the stated detection sensitivity of order 1016 cm−310^{16}\ \mathrm{cm}^{-3}; later indirect estimates place the substitutional-Mn ground level at about 0.35-0.6 eV above the valence band, inside the excluded window, so the substitutional-Mn ground level remains experimentally unverified at these Mn contents.

(2) Mechanism of suppression of the self-activated (SA) centre. By what microscopic mechanism does manganese doping at a few percent almost completely suppress the 0.9 eV photoionisation peak and hence the population of the self-activated acceptor centre, the complex of a zinc vacancy with a nearest-neighbour co-activator impurity (for example a halogen such as iodine on a sulphur site)? Candidate mechanisms to discriminate: a reduction of the thermodynamic zinc-vacancy concentration; preferential incorporation or passivation of Mn at zinc-vacancy-related sites; or another structural/chemical modification (for example Mn clustering or formation of Mn-SA complexes that remove the unperturbed SA signature).

(3) Memory-trap identification. Which defect species are the deep hole traps - localized states in the band gap that capture and store holes - at concentrations of roughly 1016 cm−310^{16}\ \mathrm{cm}^{-3} or more, that the Howard-Sahni-Alt model (in which hole charge stored in such traps shifts the internal electric field between successive voltage cycles) requires to produce the brightness-voltage hysteresis (electroluminescent memory) of ZnS:Mn thin-film devices? Since Mn doping depletes the SA acceptor, the SA centre is unlikely to be that trap; the alternatives (Mn clusters, other intrinsic complexes, extrinsic impurities, interface-related states) have not been spectroscopically identified; trap-distribution measurements on ACTFEL device structures (Krasnov 1999) locate the space-charge traps spatially but do not assign their microscopic species.

These three questions share one scientific objective: a microscopic, spectroscopically grounded account of which localized states exist in Mn-doped ZnS, what Mn does to the pre-existing vacancy-related SA centres, and which of these states supplies the hole-trapping that produces the electroluminescent memory effect.

References

  1. J.-O. Fornell, H. G. Grimmeiss, R. Mach, G. O. Muller, 'The influence of manganese doping on the self-activated centre in ZnS', Semiconductor Science and Technology 3, 511-513 (1988). DOI: 10.1088/0268-1242/3/5/016, https://doi.org/10.1088/0268-1242/3/5/016
  2. H. G. Grimmeiss, E. Meijer, R. Mach, G. O. Muller, 'Deep energy levels in ZnSxSe1-x alloys', Journal of Applied Physics 56, 2768-2777 (1984). DOI: 10.1063/1.333808, https://doi.org/10.1063/1.333808
  3. J. E. Nicholls, J. J. Davies, N. R. J. Poolton, R. Mach, G. O. Muller, 'ODMR study of self-activated emission bands from ZnSxSe1-x mixed crystals', Journal of Physics C: Solid State Physics 18, 455-466 (1985). DOI: 10.1088/0022-3719/18/2/022, https://doi.org/10.1088/0022-3719/18/2/022
  4. K. M. Lee, K. P. O'Donnell, G. D. Watkins, 'Optically detected magnetic resonance of the zinc vacancy in ZnS', Solid State Communications 41, 881-883 (1982). DOI: 10.1016/0038-1098(82)91228-5, https://doi.org/10.1016/0038-1098(82)91228-5
  5. V. Marrello, W. Ruhle, A. Onton, 'The memory effect of ZnS:Mn ac thin-film electroluminescence', Applied Physics Letters 31, 452-454 (1977). DOI: 10.1063/1.89738, https://doi.org/10.1063/1.89738
  6. V. Marrello, A. Onton, 'The Dependence of the Memory Effect in ZnS:Mn A-C Thin Film Electroluminescence on Mn Distribution', Journal of the Electrochemical Society 127, 2220-2222 (1980). DOI: 10.1149/1.2129378, https://doi.org/10.1149/1.2129378
  7. W. E. Howard, O. Sahni, P. M. Alt, 'A simple model for the hysteretic behavior of ZnS:Mn thin film electroluminescent devices', Journal of Applied Physics 53, 639-647 (1982). DOI: 10.1063/1.329971, https://doi.org/10.1063/1.329971
  8. A. N. Krasnov, 'Direct observation of traps responsible for positive space charge in alternating-current thin-film electroluminescent devices', Applied Physics Letters 74, 1120-1122 (1999). DOI: 10.1063/1.123462, https://doi.org/10.1063/1.123462
  9. M. Tanaka, J. Qi, Y. Masumoto, 'Comparison of energy levels of Mn2+ in nanosized- and bulk-ZnS crystals', Journal of Luminescence 87-89, 472-474 (2000). DOI: 10.1016/S0022-2313(99)00212-4, https://doi.org/10.1016/S0022-2313(99)00212-4
  10. Yu. Yu. Bacherikov, I. Vorona, A. Zhuk, A. V. Gilchuk, N. Korsunska, I. Markevich, 'New insight on the interaction of self-activated and Mn-related emission centers in ZnS', Semiconductor Science and Technology 32, 025006 (2017). DOI: 10.1088/1361-6641/32/2/025006, https://doi.org/10.1088/1361-6641/32/2/025006
  11. Yu. Yu. Bacherikov, I. P. Vorona, I. V. Markevich, N. O. Korsunska, R. V. Kurichka, 'Competition of the self-activated and Mn-related luminescence in ZnS single crystals', Solid State Communications 274, 31-35 (2018). DOI: 10.1016/j.ssc.2018.02.017, https://doi.org/10.1016/j.ssc.2018.02.017