On-state behaviour and degradation effects in thin-film amorphous switching devices

The on-state behaviour of thin-film chalcogenide threshold switches was studied with special reference to the influence of electrode materials and the progressive effects of device degradation. For operations during the forming period, the on-state current-voltage characteristics remain identical for both rising and falling current. The on-state minimum voltage is found to be closely proportional to the square root of the electrode thermal conductivity and, in addition, shows some polarity dependence with dissimilar electrode combinations. This behaviour is interpreted in terms of a model in which the voltage drop across the on-state current channel is dominated by constrictions and associated hot spots near the channel/electrode boundaries. After forming and after a further number of operations depending both on electrode material and ambient temperature, one or more stable intermediate steps or “branches” develop in the I–V characteristic of the transition from off-state to on-state. The exact pattern shows some polarity dependence. Branching is identified with the progressive precipitation of tellurium-rich “islands” in the glass as the number of device operations increases and as a result of which the growth of the on-state channel proceeds via a number of “island-hopping” stages.     

Electrical switching in chalcogenide liquid semiconductors

Reversible electrical switching from a higher resistance off-state to a lower resistance on-state has been observed in several chalcogenide semiconductor alloys both in the solid and liquid phases. Experiments conducted between 100 and 500°C showed that the threshold switching voltage decreases with increasing ambient temperature, and that the on-state resistance remains relatively constant up to the melting region but generally increases thereafter. The alloy Se_0.77Te_0.13S_0.10 shows a preserved on-state which can be maintained using 60 Hz ac in both the solid and liquid phases. The on-state can be reversed to the off-state by decreasing the applied field, and hence is not a conventional memory state. Below the melting region (250°C) the on-state I–V curve has a distinct knee or barrier voltage above which the current increases steeply. Above the melting range the knee becomes less clearly defined and occurs at a lower voltage. Below 345°C the above alloy has an activation energy of 0.37 eV. At this temperature an abrupt slope change or a discontuinity exists in the temperature dependence of off-state resistance, threshold field, threshold power and on-state resistance.     

Épitaxie en phase liquide des composés III–V sur substrat InP

The surface preparation and characterization of InP substrates have been studied. It is shown that a 1–3% bromine-methanol etching solution gives a smooth surface if 10 μm or more has been removed. Auger spectrometry detected some contamination of carbon and oxygen by the atmosphere during substrate preparation. The concentration of carbon impurities is less than in the case of GaAs. Liquid phase epitaxy of InP and Ga_xIn_1?xAs_1?yP_y was achieved by a slider technique. Equilibrium cooling, supercooling and step-cooling processes have been tried. It is demonstrated that misorientation from a (111) plane gives a “scale” structure, the orientation of which is exactly (111). Smooth LPE surfaces have been obtained with the supercooling and the step-cooling processes.     

Control of holding currents in amorphous threshold switches

It is usually assumed that the holding current of an amorphous semiconductor threshold switch is not strongly dependent on the circuit surrounding the switch. This paper shows that the holding current is in fact controlled by both the series resistance and the parallel capacitance, according to two simple inequalities. In order to obtain low holding currents a large series resistor and minimum parallel capacitance are required, and in this way a holding current of 10 μA has been observed. Data are also presented on the effects of device dimensions and temperature on the on-state I-V curves and holding current.     

Strukturelle Umklapp-Vorgänge in metastabilen β1-Kupfer-Zink-Mischkristallen bei tiefen und höheren Temperaturen (III). Zur Gitterkinetik der isothermen Phasen-Umwandlungen

Whereas part (I) of the present series treated chiefly lattice-dynamical aspects, being important for the initiation of the β₁/α-transformation processes, and the preceding part (II) discussed explicitely the phase transformation of β₁Cu/Zn-crystals, the submitted third part, (III), handles of a lattice kinetic detailing of the results, exclusively received by X-ray measurements (transformation isotherm 360°C). Keeping the subdivision of the transformation processes into the 3 main periods (A) PREDIFFUSION, (B) “UM-.KLAPPING”, (C) POSTDIFFUSION there are the following results: (A) The quickly running heterogenization of the initial β₁-matrix lattice, β₁ ⟹ β₁ + β₁ takes place with an activation energy Q_w for the wandering of vacancies, frozen- in by quenching: “Primary vacancies”, which will disappear in sinks already after about 2 sec. (B)The first structural actions of “Umklapping”,β₁⟹β₁⟹β₁, occurring in Zn-poorβ₁-districts, will be caused dynamically by longwavy thermo-acoustic sher-vibrations – “soft modes” of the type ({110)}, 〈110〉 with wavelengths of about 86 Ä – and succeeding dissociations of dislocations in the momentary structure β₁. The activation energy of this transformation stage, being isolated from the increase of the lattice period aβ₁, is QII= 0.8 eV> QI; therefore the energy difference QII- QI⋍ 5.10⁻² eV can be interpreted as energy barrier for the structural transitions β₁⟹ β₁⟹ β₂ – As the “mean atomic volume” of the diffusionless period is smaller than that of β₁, “secondary vacancies” will be formed during this period, and so the following diffusion processes become accelerated for some time. In this subperiod, in which the first α-crystals appear and begin to grow, the activation energy is 1.0 eV. the α-crystals are formed with 12-fold crystallographical strong defined orientations relative to the β₁-matrix. (C) During the postdiffusion period, which will begin with the disappearance of the transition lattice β₂ and which is ruled by growth of the original (coherent) fcc α-crystallets, there are only diffusive direct transitions of (Zn-rich) β₁+− and of residual (Zn-poor but not more “umklapping”) β₁-districts; β₁+ α and β₁ α respectively. The growth of the α-crystallets takes place with the activation energy QIV = 1.45 eV. This energy is the sum of the activation energies for the formation and wandering of a vacancy, Q_IV=Q_B + QW, because in the meantime the excessive “secondary vacancies” are annihilated and the “mean atomic volume” of thc reacting associations is increasing; consequently QB = 0.7 eV. Caused by the increase of the “mean atomic volume” and exhibited by X-ray diffraction photographs, near the end of the isotherm reaction the (at least partially) coherent β₁/α-association suffers a self-deformation. This deformation, only taking place in the α-members of the association, will be signified already in much former reaction stages by weak satellite reflections of some α-crystal interferences (to this s. part (II)). A texture-like diffuse prolongation of several a-reflexions shows the β₁/β₂/α-lattice coherence to be removed partially in some micro-regions (very weak X-ray intensities? ) For this there are a few reasons: (1) In the moment of “Umklapping” the concerned micro-lattice regions are superplastic so that the other strong crystallographical Pβ₁α-orientation relationship can easy get lost, at least partially. (2) It is possible that also the so called “massive transformation” contributes to this effect, since this transformation operates without lattice coherency-the rather the lower the Zn-content of thc transforming lattice regions, as it is in fact realized for the Zn-poorest β₁-districts in the moment of “Umk1apping”.     

Effective convection coefficient for porous interface and solute segregation

Convective heat and mass transfer coefficients are used to calculate the rate at which convection sweeps heat and mass away from the interface. Convection in melt growth is driven by various forces, and the resulting convoluted flows are laminar or turbulent. Furthermore, cross-flow through the “porous” interface has a profound effect on convection. Thus, a general effective coefficient (h_eff ), which accounts for: (i) uniform flow “suction” through the porous interface, (ii) forced and/or natural convection, (iii) laminar or turbulent flow, and (iv) finite Schmidt numbers, is derived. Focusing next on solute segregation, mass conservation is used to derive a simple equation for k_eff (effective segregation coefficient) as a function of h_eff. Here, h_eff is an input, which provides the rate at which convection sweeps the rejected solute away from the interface.From the general expression, even simpler expressions are developed for restricted range of conditions, e.g., Czochralski growth under forced laminar convection (no natural convection or turbulence). h_eff utilizes numerous established correlations, all developed for impermeable solids.     

Untersuchungen zur Wechselwirkung von Ar-, O2- und H2-Niederdruckplasmen mit Metallschichten

By the interaction of the positiv column of a glow discharge (plasma gases: Ar, O₂, H₂) with the surface of metal films (Au, Ni, Sn) sputtering processes are observed. The sputter yields are larger than will be assumed from known yields of low energetic ions which are dominant in the investigated surface layer. Impulse interaction increases the sputter yields. This effect is reduced to a destruction of metal lattice caused by reconstruction of the surface layer and the inertion of gases during the intervals between the impulses. The sputter yields strongly depend on the plasma gases and the current density directed to the wall (i_W ∼10⁻⁵ A. cm⁻²) it is possible by decreasing the sputter yield to observer reactive processes and gas insertion in consequence of the plasma-wall interaction. Sustained upon electron diffraction of Ni- and Au-films and kinetic analysis of sputter yields a model is developed assuming a critical surface layer of 30–50 Å thickness during the plasma influence in “equilibrium” with a metal vapour-plasma gas phase above it.     

Crystal structure of modular sodium-rich and low-iron eudialyte from Lovozero alkaline massif

The structure of the sodium-rich representative of the eudialyte group found by A.P. Khomyakov at the Lovozero massif (Kola Peninsula) is studied by X-ray diffraction. The trigonal cell parameters are: a = 14.2032(1) and c = 60.612(1) Å, V = 10589.13 Å3, space group R3m. The structure is refined to the final R = 5.0% in the anisotropic approximation of atomic displacement parameters using 3742|F| > 3σ(F). The idealized formula (Z = 3) is Na₃₇Ca₁₀Mn₂FeZr₆Si₅₀(Ti, Nb)₂O₁₄₄(OH)₅Cl₃ · H₂O. Like other 24-layer minerals of the eudialyte group, this mineral has a modular structure. Its structure contains two modules, namely, “alluaivite” (with an admixture of “eudialyte”) and “kentbrooksite,” called according to the main structural fragments of alluaivite, eudialyte, and kentbrooksite. The mineral found at the Lovozero alkaline massif shows some chemical and symmetry-structural distinctions from the close-in-composition labyrinthite modular mineral from the Khibiny massif. The difference between the minerals stems from different geochemical conditions of mineral formation in the two regions.     

X-ray diffraction study of an as-grown crystal of a two-phase sample of Lu0.65Ca0.35F2.65

The X-ray diffraction study of an “as-grown” crystal of Lu_0.65Ca_0.35F_2.65 revealed for the first time two phases with a regular mutual orientation. The c axis of the hexagonal phase (a = 3.88 Å, c = 6.89 Å) is oriented along the body diagonal to the cubic phase (a = 5.50 Å). The hexagonal phase has a tysonite-type structure with a “small” unit cell (P6₃/mmc, Z = 2). The cubic phase has a fluorite structure ( \(Fm\bar 3m\) , Z = 4).     

Dielectric relaxation in ferroelectric TlInS2 layered crystals within metastable chaotic state

The results of investigations low frequency dielectric relaxation in layered ferroelectric TlInS₂ crystals are presented. The measurements were performed in the temperature range of 180‐230 K and in the frequency range of 5 kHz–1 MHz. Two different relaxation processes were observed in mentioned temperature interval. The crystal has “slow” and “fast” relaxation mechanisms in low and high frequency region, respectively. The presence of two different relaxation mechanisms in TlInS₂ is discussed. (© 2007 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Mössbauer spectra in the xNa2O·(1 − x − y)B2O3·yFe2O3 glass system

A Mössbauer spectroscopic study is made on xNa₂O·(1 ? x ? 7)B₂O₃·yFe₂O₃; 0.045 ? x ? 0.405 and y = 0.05, 0.10 and 0.15 samples quenched from the melt. The presence of two quadrupole doublets indicates that an amphoteric cation like Fe³⁺ occupies both the tetrahedral “network forming” and the octahedral “network modifying” sites in the glass structure. The observed variations of isomer shift and quadrupole splitting with the values of x are similar to the ones found earlier in lead borates but different from those observed earlier in soda borates and other alkali alumino borates. In samples with smaller values of x, an additional devitrified magnetic phase of Fe₂O₃ is seen. The distribution of Fe³⁺ ions in different sites and phases is also discussed.     

Refinement of the crystal structure of sanidine-like feldspar

The crystal structure of sanidine-like feldspar of the composition KAlSi₃O₈ from the Khibiny alkaline massif (the Kola Peninsula) has been refined (X-ray diffraction analysis; automated Syntex $P\bar 1$ diffractometer; 2θ: θ scanning technique; 2320 reflections; R _(hkl) = 0.0409; anisotropic refinement; AREN program package). The data obtained for KAlSi₃O₈ are: a = 8.615(9), b = 13.030(7), c = 7.200(5) Å, α = 89.99(5)°, β = 116.01(6)°, γ = 89.98(7)°, Z = 4, sp. gr. $C\bar 1$ . Microtwinning revealed in the crystal structure of the mineral explains the simultaneous existence of two structural-optical types in one sample—“high” and “low” sanidines.     

Crystal structure of the cubic modification of Cu2FeSnS4

The structure of Cu_{2 ? x} Fe_{1 ? x} SnS₄, the last member of the isomorphous series Cu₃FeS₄-Cu₂FeSnS₄, was synthesized for the first time and solved by the Rietveld method (the WYRIET 3.3 version). The compound occurs as a cubic phase (a ₀ = 5.4179(5) Å) crystallizing in the sp. gr. $P\bar 43m$ . One-half of the tetrahedral cavities in the close packing of sulfur atoms [4e (xxx), x = 0.2583(5)] are statistically filled with metal atoms (Sn + Fe) and (Cu + Fe + Sn) [1a(0 0 0) and 3c (0 1/2 1/2), respectively]. The results of X-ray study (R _p = 2.33, R _wp = 3.00, s = 1.20, D = 1.53, R _B = 2.52, and R _F = 2.00) confirm the existence of the stannite “prototype” with a “small” cubic unit cell.     

Hydrolytic polycondensation of Si(OC2H5)4 and effect of reaction parameters

Hydrolysis of tetraalkoxysilanes leads to the formation of organosiloxane polymers by polycondensation reactions. This process is strongly affected by certain parameters: molecular separation of species, their ratios, the reaction medium, the concentration of catalyst, the reaction lenth, and the temperature. The general effects of these parameters on the polymeric make-up and molecular size distribution during the hydrolytic polycondensation of Si(OC₂H₅)₄ were investigated by using high resolution nuclear magnetic resonance (NMR) and size-exclusion liquid chromatography (SEC). The results show that each of these parameters, particularly the molecular separation during the hydrolytic polycondensation affect the molecular structure in a fundamental way. There is also evidence that the molecular size expansion occurs by two distinct processes: initially by a gradual “growth” process, and later by “recombination” of high-molecular weight species. The latter process often leads to a bi-modal molecular size distribution.     

Structural relaxation of an “unusual” Pd40Cu40P20 metallic glass studied by measurements of the electrical resistance

Precise measurements of the electrical resistance upon thermo-cycling of an “unusual” Pd₄₀Cu₄₀P₂₀ glass in the initial and cold-rolled states have been performed. The uniqueness of this glass is defined by its higher density compared to the reference tetragonal Pd₂Cu₂P crystal. It has been found that this glass displays the relaxation behavior, which is quite similar to that of “usual” Pd₄₀Cu₃₀Ni₁₀P₂₀ glass with the usual relationship between the densities of glassy and crystalline states.     

Analysis of the conformational polymorph crystal structures by means of molecular Voronoi-Dirichlet polyhedra

The nonvalent interactions in the crystal structures of 16 conformational polymorphs of the C _a H _b N _c O _d S _e composition (a, b, c, d, and e are stoichiometric indices), including 5-methyl-2-[(2-nitrophenyl)amino]-3-thiophenecarbonitrile (“ROY”) (which has seven different modifications), have been analyzed using the characteristics of molecular Voronoi-Dirichlet polyhedra. It is established that a unique combination of implementable types of intra- and intermolecular nonvalent contacts corresponds to each conformational polymorph.     

Magnetic susceptibility of sodium rare-earth fluorites Na0.5 − x R 0.5 + x F2 + 2x (R = Dy, Ho, Er, Tm, Yb) and some ordered phases

Magnetic properties of the nonstoichiometric phases with the fluorite-type “defect” structure Na_0.4 R _0.6F_2.2 (R = Ho, Er, Tm, Yb) and Na_0.35Dy_0.65F_2.3 have been studied as well as some ordered phases with different distortions of the initial fluorite lattice (NaHo₂F₇, Na₇Tm₁₃F₄₆, NaYb₂F₇, and Na₇Yb₁₃F₄₆) and Na_1.5Dy_1.5F₆ with the gagarinite-type structure (NaCaYF₆). Magnetic susceptibility χ was measured with the aid of the Faraday balance in the temperature range from 20 to 300 K. The crystals with R= Dy, Ho, Er, Tm are paramagnetic; their temperature dependence χ is described by the Curie-Weiss law. It is shown that in the temperature range studied, magnetic susceptibility does not depend on the degree of order of Dy, Ho, Er, and Tm ions in the structure. Below 80 K, the temperature dependence of magnetic susceptibility of the the Yb³⁺-containing phases differs from the analogous dependence χ of an ideal paramagnetic crystal, which is assumed to be caused by the interaction of the crystalline field with Yb³⁺ ions.     

Dependence of the parameters of equations of viscous flow on chemical composition of silicate melts

The temperature dependence of viscosity of silicate melts is discussed in the framework of the Avramov–Milchev (AM) equation. The composition is described by means of two parameters: the molar fraction, x, and the “lubricant fraction”, l. The molar fraction is the sum of the molar parts x_i of all oxides dissolved in SiO₂, the molar fraction of the latter being 1 ? x. It is shown that, with sufficient precision, two of the parameters of the AM equation can be presented as unique functions of the molar fraction. On the other hand, x is not sufficient to determine properly the reference temperature T_r , at which viscosity is η_r = 10¹³ [dPa.s]. Therefore, additional parameter, “lubricant fraction” l, is introduced. For each of the components, l_i is a product of molar part x_i and a specific dimensionless coefficient 0 ≤ k_i ≤ 1 accounting for the specific contribution of this component to the increased mobility of the system. It is demonstrated that, for l > 0, the reference temperature is related to the “lubricant fraction” l through the reference temperature T_r,SiO2 of pure SiO₂.     

The structure of vapour-deposited arsenic sulphides prepared using a cooled (78 K) substrate

Amorphous molecular As₄S₄ and non-stoichiometric “As₂S₃” evaporated films have been prepared by condensing onto a substrate at liquid nitrogen temperature in situ in the X-ray diffractometer. The structures of these films are shown to be similar, thus confirming the importance of As₄S₄ (realgar-type) molecules in amorphous films prepared by evaporation of “As₂S₃” glass. The local molecular packing in the as-deposited amorphous As₄S₄ film shows a preference for a packing arrangement similar to that found in the α-crystalline phase. Both as-deposited films contain gross structural defects such as molecular clusters and/or voids which irreversibly reduce in importance as the film structures relax on warming to ambient temperatures. Also, the As₄S₄ film partially devitrifies to the high-temperature β-crystalline phase on warming to ambient, but no devitrification occurs in the non-stoichiometeric film due to the presence of more than one molecular type.     

New boron-oxygen layer in the structure of barium hydrodecaborate Ba5[B20O33(OH)4] ⋅ H2O

The crystal structure of the new synthetic compound Ba₅[B₂₀O₃₃(OH)₄] ? H₂O was established by the methods of X-ray diffraction (a Stoe IPDS diffractometer, λMoK _α?, 1860 independent reflections, anisotropic refinement, R = 1.95%, localization of hydrogen atoms): a = 9.495(2) Å, b = 6.713(1) Å, c = 11.709(2) Å, β = 95.09(1)°, sp. gr. P2, Z = 1. The structure is based on double pseudohexagonal layers consisting of BO₄-tetrahedra and BO₃ triangles linked into three-membered rings in two mutually perpendicular directions. The double layers adjacent along the [100] direction are linked together through the Ba-polyhedra and hydrogen bonds with the participation of the OH-groups occupying the “end” vertices of two B-triangles. The interlayer space is also filled with a sheet of Ba-polyhedra. The structure of the compound is compared to the structures of topologically similar Ba and Ca borates and hydroborates.