Metal-insulator Interface Research Articles

Origin of the dielectric dead layer in nanoscale capacitors

Capacitors are a mainstay of electronic integrated circuits and devices, where they perform essential functions such as storing electrical charge, and blocking direct current while allowing alternating currents to propagate. Because they are often the largest components in circuits, extensive efforts are directed at reducing their size through the use of high-permittivity insulators such as perovskite-structure SrTiO3 (refs 1, 2), which should provide more capacitance per unit area of device. Unfortunately, most experiments on thin-film SrTiO3 capacitors have yielded capacitance values that are orders of magnitude smaller than expected. The microscopic origin of this reduced capacitance, which is often discussed in terms of a low-permittivity interfacial 'dead layer', is not well understood. Whether such a dead layer exists at all, and if so, whether it is an intrinsic property of an ideal metal-insulator interface or a result of processing issues such as defects and strains, are controversial questions. Here we present fully ab initio calculations of the dielectric properties of realistic SrRuO3/SrTiO3/SrRuO3 nanocapacitors, and show that the observed dramatic capacitance reduction is indeed an intrinsic effect. We demonstrate the existence of a dielectric dead layer by calculating the dielectric profile across the interface and analyse its origin by extracting the ionic and electronic contributions to the electrostatic screening. We establish a correspondence between the dead layer and the hardening of the collective SrTiO3 zone-centre polar modes, and determine the influence of the electrode by repeating our calculations for Pt/SrTiO3/Pt capacitors. Our results provide practical guidelines for minimizing the deleterious effects of the dielectric dead layer in nanoscale devices.

RF MEMS dielectric sensitivity to electromagnetic radiation

SiO 2 and Si 3N 4 as well as other insulators are used in MEMS. However, their tendency for electrostatic charging diminishes the device reliability. The charging effect becomes significant when these devices are subjected to ionizing radiation. The irradiation induced charging depends on the nature of irradiation, the vicinal metal layers, and the metal–insulator interface properties. The sensitivity of RF MEMS to electromagnetic radiation is presented, taking into account the simulation of charge generation to device's structure with different insulating layer composition and thickness.

Characterization of interfacial properties in magnetic tunnel junctions by bias-dependent complex impedance spectroscopy

The bias-dependent complex impedance spectra of ion-beam sputtered CoFe-AlO/sub x/-CoFe tunnel junctions have been systematically investigated. An equivalent circuit model, composed of a resistance component and two sets of parallel resistance (R) and capacitance (C) components in series, has been utilized to describe the individual impedance contribution from lead of cross patterns, CoFe-AlO /sub x/ interfaces and bulk AlO/sub x/ in CoFe-AlO/sub x/-CoFe tunnel junctions. The fitting results reveal that the RC component characterizing the CoFe-AlO/sub x/ interface can be explained by a Schottky barrier with imperfectly blocking characteristic, which is attributed to the existence of traps in the depletion region. The tendency of interfacial resistance, R/sub i/, and interfacial capacitance, C/sub i/, as a function of dc bias, V/sub dc/, is expected for the space charge limited current model with exponential trap distribution, which coincides with the analyzing results of dc four-probe current-voltage curves. On the other hand, the R and C components of bulk AlO /sub x/ almost remain constant when V/sub dc/ increases, in contrast to the electrical behavior of CoFe-AlO/sub x/ interfaces. The results suggest the metal-insulator interface, instead of bulk insulator, dominates the electrical transport process of MTJs.

Bonding of Pd, Ag, and Au atoms on MgO(100) surfaces andMgO∕Mo(100)ultra-thin films: A comparative DFT study

We have studied by means of density functional theory (DFT) plane-wave calculations the adsorption properties of ultrathin MgO films grown on Mo(100) as function of the thickness of the oxide film, and compared the results with those of unsupported slabs representing a single-crystal MgO(100) surface. The presence of a metal substrate at the interface with the MgO film results in the charging of adsorbed atoms with high electron affinity, like Ag and Au, while Pd is basically unperturbed. As a consequence, while the properties of Pd are the same on MgO(100) or $\mathrm{Mg}\mathrm{O}∕\mathrm{Mo}(100)$, for Ag and Au substantial differences are found. On $\mathrm{Mg}\mathrm{O}∕\mathrm{Mo}(100)$ films the $\mathrm{Ag}(5s)$ and the $\mathrm{Au}(6s)$ levels fall below the Fermi level of Mo, leading to negative charging of the metal atom and to a different bonding mechanism; the $5s$ level of Pd, on the contrary, is just above the Fermi level of Mo, and no charge transfer occurs. The effect tends to decrease for thicker MgO films and is closely related to the change in work function of the Mo(100) surface when a thin MgO film is deposited on it and to the appearance of metal-induced gap states at the metal-insulator interface. The results are of relevance to understand the properties of ultrathin oxide films. The charging effect is absent on single-crystal or polycrystalline MgO surfaces.

The influence of the insulator surface properties on the hydrogen response of field-effect gas sensors

The hydrogen response of gas-sensitive field-effect devices is mainly due to trapping of atomic hydrogen on the insulator side of the metal-insulator interface of the metal-insulator-semiconductor (MIS) structure. Therefore an influence of the choice of insulator on the hydrogen response properties is expected. We have investigated this influence by producing MIS capacitors with four different insulators; SiO2, Al2O3, Si3N4, and Ta2O5. The results show that the choice of insulator influences the detection limit, the saturation concentration, and the saturation response. Furthermore, there is a strong correlation between the observed saturation response and the oxygen concentration of the insulator surface, as measured by Auger electron spectroscopy, which indicates that the trapping of hydrogen at the interface occurs at the oxygen atoms of the insulator surface. Finally, if the metal film is porous a catalytic oxidation of the insulator surface appears to be facilitated, which can increase the hydrogen response.

Hydrogen interaction with platinum and palladium metal–insulator–semiconductor devices

Hydrogen-sensitive Pd–SiO2–Si and Pt–SiO2–Si metal–insulator–semiconductor (MIS) devices have been studied in ultrahigh vacuum in the temperature range of 223–523K. Adsorption/absorption of hydrogen occurs at the metal surface, in the metal bulk, and at the metal–insulator interface. The sensor signal, caused by hydrogen adsorption at the interface, shows a logarithmic dependence on the applied hydrogen pressure. The Pt-MIS device, which is fully functional at atmospheric pressures, is sensitive to changes in hydrogen pressure down to the 10−12-Torr scale. We propose that the interface adsorption follows a so-called Temkin isotherm with an interface heat of adsorption that varies with hydrogen coverage as ΔHi0(1−aθ). The initial heat of adsorption ΔHi0 is determined to 0.78eV∕hydrogen atom. The adsorption potential at the external Pt surface is found to be 0.45eV∕hydrogen atom. These values were obtained by modeling the hydrogen interaction with the MIS devices and fitting the model to a number of experimental results. Also studies of Pd-based devices were performed and compared with Pt. The hydrogen adsorption on the metal surface, previously treated as a first-order process on Pd, is shown to follow a second-order process. Qualitatively the results from the Pd- and Pt-MIS devices agree. Quantitatively there are differences. The hydrogen sensitivity of the Pt-MIS device is only approximately one-third compared to that of the Pd-MIS structure. This agrees with the result that the concentration of available hydrogen adsorption sites at the Pt–SiO2 interface is approximately 7×1017m−2 whereas the concentrations of sites at the Pd–SiO2 interface is roughly three times larger (2×1018m−2). An estimate of the size of the dipole moments (0.6–0.7D) implies that the interface hydrogen atoms are strongly polarized. Differences are also observed in the microstructure of the metal films. Atomic force microscopy results show that the Pd surface reconstructs during H2–O2 exposures, while the Pt surface shows no such change at these temperatures.

RF MEMS sensitivity to electromagnetic radiation

Silicon dioxide and silicon nitride as well as other insulating materials are used in micro-electro-mechanical systems. However, their tendency for electrostatic charging diminishes the device reliability. The charging effect becomes significant when these devices are subjected to ionizing radiation. The irradiation induced charging depends on the nature of irradiation, the underlying metal layers and the metal-insulator interface properties. The sensitivity of RF micro-electromechanical systems to electromagnetic ionizing radiation is presented, taking into account the simulation of charge generation and the device structure.

Complex capacitance spectroscopy as a probe for oxidation process of AlOx-based magnetic tunnel junctions

Proper as well as under- and over-oxided CoFe−AlOx−CoFe magnetic tunnel junctions (MTJs) have been systematically investigated in a frequency range from 102 to 108Hz by complex capacitance spectroscopy. The dielectric relaxation behavior of the MTJs remarkably disobeys the typical Cole–Cole arc law probably due to the existence of imperfectly blocked Schottky barrier in the metal-insulator interface. The dielectric relaxation response can be successfully modeled on the basis of Debye relaxation by incorporating an interfacial dielectric contribution. In addition, complex capacitance spectroscopy demonstrates significant sensitivity to the oxidation process of metallic Al layers, i.e., almost a fingerprint of under, proper, and over oxidation. This technique provides a fast and simple method to inspect the AlOx barrier quality of MTJs.

Ab initio calculations of interface effects in tunnelling through MgO barriers onFe(100)

The tunnel-magneto-resistance (TMR) effect has been extensively studied in thelast couple of years. While some experiments showed a strong dependence onthe metal–insulator interface in these systems, most theoretical work gave littleemphasis to the complicated interplay between electronic structure, atomic structureand the tunnelling process. We present calculations of the atomic structure ofFe/MgO/Fe(001) interfaces and its influence on the electronic structure. The tunnel currentis calculated using the Landauer approach to describe the electron transport.

The conduction current in a metal-insulator-metal structure

The flow of the conduction current in a metal-insulator-metal structure under X-ray and optical excitation is considered. Accumulation of positive charge at the negative electrode in the KBr and CsI crystals is studied. The method of discharge currents after the X-ray and optical excitation is suggested for estimation of the near-surface charge. It is shown that the values of charge transported by the conduction current and those determined from the measured discharge-current signal are in satisfactory agreement. The lifetimes of near-surface charges of holes and anionic vacancies and the values of the electric-field strength at the metal-insulator interface are estimated. It is assumed that electroneutrality of the sample is established owing to the motion of electrons from the surface into the bulk over dislocations.

Surface morphological influence on charging at metal–insulator interface in XPS depth profiling

Charging in depth profiling in X-ray photoelectron spectroscopy (XPS) for metal layers on insulated substrates (Au/Cr/SiO 2 (quartz)) was studied both with and without conduction between a specimen and a spectrometer (i.e., the conducting mode and the floating mode). Using a neutralizer, differential charging and electrical interface “charge jump” were simultaneously observed at the Cr/SiO 2 interface in the conducting mode. In contrast, in the floating mode differential charging was avoided after the charge jump. By AFM performed in the tapping mode for phase ex situ observation, the layer Cr distribution before charge jump was apparently changed to the island distribution after the charge jump. This surface morphological change due to a heavy sputtering process is considered to be the origin of the observed differential charging and the charge jump. Thus, it is concluded that charging is strongly related to the surface morphology in XPS depth profiling for metal layers on insulated substrates. Hence, in XPS depth profiling for a sample with a metal–insulator interface, the floating mode performance, which is attained by insulating a sample from the spectrometer, is desired to keep the surface potential of the sample uniform.

Morphology and transport properties of nanostructural gold on silicon

Nanometer sized Au clusters deposited on a silicon substrate forming Au–SiO2–Si structure are important for the development of contacts in nanotechnology. Systematic x-ray reflectivity, scanning probe microscopy, and scanning tunneling spectroscopy measurements were done to understand the relationship between morphology and electrical transport properties of this nanostructural metal–insulator–semiconductor system. The presence of an interfacial layer at the metal–insulator interface dictates the tunneling current through this structure and exhibits a gap leading to a suppression of current. Local density of states and electron density/thickness of the interfacial layer have been extracted from the measurements to understand the evolution of metallicity of this Au–SiO2–Si structure.

Electrical breakdown in ionic crystals exposed to nanosecond pulses

Electrical breakdown in alkali halide crystals subjected to ∼10-ns-long electrical pulses is studied. Two, primary and basic, channels of the anodic discharge are noted. In the presence of the primary channel, the other arises at static breakdown voltages. Otherwise, the basic channel forms at voltages exceeding the static breakdown voltage by more than four times. The basic channel is assumed to form via a cascade of Auger transitions. The generation and migration of linear defects seem to play a significant role in the basic channel formation. The enhancement of the pulsed dielectric strength of the crystals is related to conditions of current passage through the metal-insulator interface.

Characterization of the metal–insulator interface of field-effect chemical sensors

The metal–insulator interface of hydrogen-sensitive metal–insulator–semiconductor capacitors, with SiO2 as the insulator and Pt as the metal contact, has been studied. The thin Pt films were prepared in ultrahigh vacuum by electron beam evaporation and dc magnetron sputtering. Deposition parameters were substrate temperature and sputtering pressure. The hydrogen responses of the differently prepared devices were measured in a semi-inert ambient (and used as a measure of the concentration of available adsorption sites for hydrogen at the interface). A large variation of responses was found for differently prepared sensors, and the magnitude of the response was found to increase for decreasing bonding strength between the Pt film and the SiO2 substrate, as determined by scratch adhesion measurements. The bonding strength was controlled via the energetics of the Pt deposition flux. The largest interfacial roughness, from cavities between noncoalesced metal grains, and the poorest adhesion, was obtained by a reduced surface diffusion during growth and incomplete coalescence of the metal grains on the oxide surface as studied by transmission electron microscopy and atomic force microscopy. From x-ray photoelectron spectroscopy studies it was concluded that no chemical bonds were formed between Pt and SiO2. It is inferred that the hydrogen active in the sensor response is adsorbed on the oxide side of the interface in a spillover process. The difference in hydrogen response between differently prepared devices can be explained by a difference in concentration of available adsorption sites giving rise to a detectable dipole moment, on the oxide due to a blocking by Pt atoms in contact with the oxide. Thus, the concentration of Pt atoms in contact with the oxide affects both the hydrogen response and the metal–oxide adhesion.

Temperature dependence of the dielectric response of anodized Al–Al2O3–metal capacitors

The temperature dependence of capacitance, CM, and conductance, GM, of Al–Al2O3–metal capacitors with Cu, Ag, and Au electrodes has been measured between 100 and 340 K at seven frequencies between 10 kHz and 1 MHz. Al2O3 films between 15 and 64 nm thick were formed by anodizing evaporated Al films in borate-glycol or borate-H2O electrolyte. The interface capacitance at the Al2O3–metal interface, CI, which is in series with the capacitance CD due to the Al2O3 dielectric, is determined from plots of 1/CM versus insulator thickness. CI is not fixed for a given metal–insulator interface but depends on the vacuum system used to deposit the metal electrode. CI is nearly temperature independent. When CI is taken into account the dielectric constant of Al2O3 determined from capacitance measurements is ∼8.3 at 295 K. The dielectric constant does not depend on anodizing electrolyte, insulator thickness, metal electrode, deposition conditions for the metal electrode or measurement frequency. By contrast, GM of Al–Al2O3–metal capacitors depends on both the deposition conditions of the metal and on the metal. For Al–Al2O3–Cu capacitors, GM is larger for capacitors with large values of 1/CI that result when Cu is evaporated in an oil-pumped vacuum system. For Al–Al2O3–Ag capacitors, GM does not depend on the Ag deposition conditions.

Voltage-dependent dielectric breakdown and voltage-controlled negative resistance in anodized Al–Al2O3–Au diodes

Theories of dielectric breakdown in insulating films normally assume that dielectric breakdown depends on the electric field in the sample; that is, the thicker the film the higher the breakdown voltage. Contrary to theoretical expectations, voltage-dependent dielectric breakdown is observed in Al–Al2O3–Au diodes where Al2O3 is made by anodizing in different electrolytes. The breakdown voltage is ∼4.5 V, independent of Al2O3 thickness and anodizing electrolyte. Voltage-controlled negative resistance (VCNR) develops in the current–voltage (I–V) characteristics of Al–Al2O3–Au diodes after voltage-dependent breakdown. Electron emission into vacuum accompanies the formation of VCNR in the I–V characteristics. Detailed studies of the development of VCNR show that the maximum current, the voltage for maximum current, and the voltage threshold for electron emission depend on the maximum voltage applied to the sample. A large current increase occurs for maximum applied voltage between 5 and 7 V. A fully developed VCNR characteristic has an ohmic contact suggesting that the development of an ohmic contact at a metal–insulator interface initiates breakdown.

Correlation between charge transfer and stick-slip friction at a metal-insulator interface

Techniques have been developed that facilitate the measurement and imaging of the charge exchanged between metal-insulator surfaces in relative motion. In the regime where the forces of friction lead to stick-slip motion, we find that the charge transfer accompanying the slip events is proportional to the force jumps and is bunched at the stick locations. The constant of proportionality is measured in electron volts per angstrom and has a small variance over a large range of slip sizes, suggesting that in these experiments macroscopic friction originates from and scales to the intrinsic electronic interactions that form between metal and insulator surfaces.

Interference-induced blockage of direct tunneling current in metal–insulator–semiconductor structure

Current density–field (J–E) characteristics of a metal–insulator–heavily doped semiconductor structure were computed numerically by solving the time-dependent Schrödinger equation for insulators (barriers) of different potentials and widths. It is shown that the (J–E) profile exhibits three regions of different J(E). The region of high E is found to consist of a slowly increasing part followed by a fast increasing one. The slow region part is attributed to quantum mechanical interference effects at the metal–insulator interface. The presence of this region in experimental results should be used as an evidence of tunneling-dominated conduction. The whole computed J–E curves are supported by experimental results published on thin films of semiconductor oxides and varistors. The effects of barrier width and height on the J–E characteristics are also discussed.

Probing conducting particles buried in a Nix(SiO2)1−x composite by conducting atomic force microscopy

In this article, we present an experimental study on probing conducting particles buried in a Nix(SiO2)1−x composite with x around the percolation threshold xc by conducting atomic force microscopy (C-AFM). The buried conducting particles were “observed” via the electric current image of C-AFM at constant bias. The current from buried conducting particles originates from field assisted tunneling through the insulating layer. Examples of measuring the thickness of the insulating layer will be given. The analysis shows that it is possible to probe the buried metal particles as deep as several nanometers underneath the surface. By correlating the surface topographic and the current image, the profile of the metal-insulator interface can be measured. General issues on spatial resolution will also be discussed in this article.

Photoelectron spectroscopy of thin discontinuous metal films deposited onto a sapphire substrate

The charging of small metallic clusters forming discontinuous gold and gallium films, condensed in vacuum onto an insulating substrate (sapphire), was experimentally studied by the method of ultraviolet photoelectron spectroscopy (UPS) for a photon energy of 5.4 eV. It seems that the charging is due to electron emission as well as to electron transfer from metal islands to the empty electron states in the metal-insulator interface.