This PDF file contains the front matter associated with SPIE Proceedings Volume 7025, including the Title Page, Copyright information, Table of Contents, Introduction (if any), and the Conference Committee listing.

The technique for fabrication of thin film filters with high mechanical strength, capable of withstanding the prolonged heat load of 1 W/cm², has been developed. Freestanding multilayer Zr/Si filters of size 20×150 mm² with high transparency of 76% at wavelength λ = 13 nm were manufactured for EUV lithography tool. We have also developed and fabricated various designs of filters (freestanding or mesh supported) with lower transparency of 40-50% for experiments with intensive EUV sources. The tests of differential pressure withstandability and heat-resistance of filter samples were fulfilled. In order to model the influence on the filter of intensive radiation of the lithography source we have tested Zr/Si film samples by the Joule heating in vacuum at residual pressure of 10^-8 Torr. The testing consisted in continuous heating of Zr/Si films at the electrical power per area unit from 0.5 W/cm² to 6 W/cm² during long period of time (up to 2 months). The influence of the long-term heat load on the transparency of samples at λ = 13 nm and within wavelength region 0.3 - 2 μm was investigated.

In the paper first results of study of etching speed and RMS roughnesses depending on an incidence angle and energy of ions for Si and Cr/Sc materials are presented. Possibility of variation in parameters of etching process (the incidence angle, energy of ions and exposition time) allows to pick up the optimum mode providing high speed etching or the minimal development of a roughness of a surface.

In the article a program for simulation of intensity of fluorescence of X-ray tubes (XRT) target materials excited by electrons with the energy less than 15 keV is described. The basis of the program is a modeling of interaction of fast electrons with a matter by the Monte-Carlo method. The paper has a comparison data of theoretical and experimental intensities of XRT radiation at Si Lα characteristic line and bremsstrahlung radiation of W target. Basing on the calculation a new XRT was developed.

Status of activities in field of manufacturing and investigation of objective lens for ultrahigh resolution lithography facilities carrying out in IPM RAS is reported. Some physical aspects of operation of interferometers with diffraction reference wave are considered. A scheme of a point diffraction interferometer developed in IPM is presented. Last experimental data on making high precision spherical substrates with NA=0.25 are presented. A problem of surface shape measurements of aspherical substrates with a help of a point diffraction interferometer is discussed. A scheme of measurements and experimental data of wave deformation of 6-lens objective are given.

A new type of a spherical wave source based on a probe for a near-field microscope for a precision interferometry is reported. Experimental data of wave deformations of the light generated by the probes are given. For a numerical aperture of optics under study NA=0.22 the obtained deformations are low than λ/2000 for a red light of He-Ne laser.

On a base of the Kirchhoff-Helmholtz method and using the Green function for a half-space which is covered by a film of any thickness and with arbitrary optical properties, amplitude-phase characteristics of diffracted on a pin-hole field were calculated. Analysis was carried out for some materials of the film at radiation wave-length λ₀=633 nm using experimental values of the dielectric constants. The diffracted wave front has deformations which dependent on the film thickness and material electrodynamic characteristics, and the observation angle. The deformation determined by the film thickness essentially limits an accuracy provided by point diffraction interferometers.

Dry resistless process was studied of mask image formation by electron beam deposition from hydrocarbon precursor undecane (C₁₁H₂₄) on substrates of SiO₂ (layer 80 nm) on silicon and cupper (layer 430 nm) on silicon. A mask in form of grating of 5-150 nm height strips was created in a cell introduced into the scanning electron microscope CamScan. Strips thickness δ was considerably more than the beam size and depended on substrate material: for SiO₂ δ=0.6 μm, for Cuδ=2 μm. Strong dependence of growth rate V (at Cu) on the line scan time τ was found out. At beam current 1.0 nA varying τ from 20 ms to 13 s led to 7.4 times decreasing V. This effect most likely is caused by significant diffusion delays at τ=13 s in precursor transport into reaction zone during the pixel time. The ion beam etching of substrates through the deposited mask was carried out. SiO₂ substrate was etched by SF₆ ions, Cu substrate was etched by Ar ions. In both cases etching rate of mask material were close to etching rate of substrate. In mask deposited on SiO₂ thin (about 1 nm) intermediate surface layer was found having significantly more (8-10 times) etching resistance than the basic mask material.

A focused ion beam source of a new type is suggested for the first time. The original design of this source is based on the using of both a laser ablation and an advanced ion beam buffer gas cooling technique. Operation of the new ion beam source has been studied by means of computer experiments. For this purpose, detailed gas dynamic simulations based on the solution of a full system of time-dependent Navier-Stokes equations have been performed for both a conical supersonic nozzle having an inner tube on the axis and a novel RF-funnel extraction system. The results of gas dynamic calculations were used for detailed microscopic Monte Carlo ion-beam trajectory simulations under the combined effect of the buffer gas flow and electric fields of the RF-funnel. The obtained results made it apparent that the suggested ion beam source looks very promising for the use in the micro- and nanoelectronics technologies.

The eight-channel thin-film resistive sensor of atomic hydrogen which in the automated mode allows to measure hydrogen atoms flow density in atomic-molecular mix in conditions of the low gas pressure (10^-2-10^-4 Pa) and under action of infra-red and visible radiation noises is described. The sensor can be used for measurement of AH flow density distribution on the large cross-section beam, including measurement during GaAs surface cleaning. The measurement range of atoms flow density is 5×10¹³-10¹⁶ at.cm^-2s^-1 and measurement time is 1-10 minutes. The resistor of the sensor is produced by microelectronics planar technology that provides an opportunity of the high space resolution at beam spatial distributions measurement.

Results of impurity composition investigation of large-area atomic hydrogen beam formed by a low-pressure arc discharge source with self-heated cathode were cited. The study was performed on 18 metallic elements. Method, based on measurement of sheet concentration of metal atoms on Si sample surface before and after exposition in atomic hydrogen flow was used for determination of composition of impurities in atomic hydrogen beam. Measurement of sheet concentration of metal atoms was realized by ToF SIMS method. The method sensitivity was 10⁸ at. cm^-2. Principal reason of parasitic metal particles occurrence in the beam and the methods for reduction of metal impurity concentration in ΑH beam was investigated.

The results of comprehensive study of electron impact ionization processes for neutral boron atoms and BF molecules are presented. In the case of atomic boron, the published ionization cross sections have been analyzed and new cross sections (including ionization from the metastable state) were calculated. The electron impact non-dissociative ionization cross section calculations for BF were carried out using a semi-empirical approach. The Franck-Condon factors involved were calculated basing on the vibrational wave functions which were obtained by solving the vibrational Schrödinger equation for those ground states. The ab initio ground state potential energy curves for BF and BF⁺ were used.

Experimental realization of hysteresis free mode of vanadium reactive magnetron sputtering in Ar/O₂ mixtures has allowed, for the first time, a detailed measurement of discharge current- voltage characteristics (CVC). They appear to be not smooth but with a kink, which was not observed earlier. It is shown, that the existing model of reactive sputtering can describe only part of the observed CVC, the one above a certain ratio of a discharge current to oxygen partial pressure. The experimental data at smaller currents can be interpreted assuming the target oxidation to a depth not smaller than two monolayers, therefore sputtering does not result in the metal base exposure. The mechanisms of oxidation can be recoil implantation of surface oxygen atoms and probably radiation-enhanced thermal diffusion.

An etch-stop phenomenon taking place in Cl₂/O₂/N₂ plasma was investigated by means of the actinometry technique as well as using PlasmaVolt and PlasmaTemp in-situ sensor wafers from K-T promesys of KLA Tencor. The temperature of TCP quartz window was determined to increase from 50°C to more than 100°C during sequential plasma processing that in turn induces a plasma composition change. Based on the optical spectra a 20% increase in both atomic [O] and [Cl] concentrations was determined to take place in Cl₂/O₂/N₂ plasma used for STI etching. Performing experiments on the blanket wafers the dependence of the etch-stop on O₂-flow, pressure, TCP power, wafer temperature and quartz window temperature was determined. The values of RF voltage and temperature were measured locally at different positions on the wafer surface providing additional information about plasma homogeneity. Finally, the results are discussed in terms of the well-known recombination model considering the increase of [O] concentration results in the etch stop in the middle of the wafer.

The purpose of the paper is to investigate the measurable variations in chemistry of SF₆/O₂/Ar plasma due to etching through layers interface of structure poly-Si/SiO₂/Si. The noticeable magnitude and sufficient stability in some single parameters deference make it possible to develop application of Langmuir probe as implementation of a simple end-point-detection technique. The proposed method is based on the established idea that the surface reactions involved to the process of etching lead to dramatic changes in some parameters of the charged plasma species during the process. Particularly it was found that the densities of electrons and ions and the electron temperature are affected. It was shown that effective electron temperature and electron energy distribution function of the reactive gaseous mixture differ greatly from those of Ar plasma under the same excitation conditions. An approach to wafer-surface charging minimization by varying excitation settings and EEDF was proposed.

This work concerns with an industrial method and a device for precise automatic end point detection of ion and ion beam etching of thin films on the surface of heterostructures. The main principle of this method is measuring of the current between a film on a surface of a substrate and the earth (zero potential) during ion or ion beam bombarding. This kind of current, especially if the film is made of dielectric, is called induced current. It has been found that when the film thickness gets value in the range of 8-10 nm, it is registered a jump of induced current. If this jump of the current can be registered, the etching process can be stopped precisely. The main feature of this method and device is using as a probe to measure ion beam parameters. These measuring results can do information about ion beam density, stability, pulses of accelerating voltage etc. Joint measuring of two kinds of current, their comparison, treatment of signals, filtering signals of voltage pulses and other peculiarities make it possible to stop processes with great accuracy.

An adhesion mechanism of thin copper and chromium films fabricated on the surface of dielectric substrates (silicon dioxide, polycrystalline glass) is studied theoretically and experimentally. It is shown that adhesion increasing is provided by organic molecules dissociation in the boundary pollution layer of the metal - substrate system as a result of thermal balance establishment in thin metal film - atomic layer of organic pollution - substrate surface (Me - C_xH_y - Sub) nanosystem. The maximal adhesion value achieved at process time not less than 3 minutes, current value - 80 mA, accelerating voltage value - 4 kV. Ion-electron Me - C_xH_y - Sub structure bombardment increases thin metal films adhesion not less than in 3.8 ... 10 times. It has been shown that the developed method allows to lower surface cleanliness requirements, to reduce process time in 10 times, to achieve adhesion value in 1.5 - 2 times higher than values reached by traditional methods in which substrates with technologically pure surface are used.

Thin (90 nm) a-Si:H films have been crystallized on Corning 7059 glass substrates by 120 fs pulses of Ti:sapphire laser. Initial films were deposited using low-temperature plasma enhanced deposition technique. The pulses with wavelength of 800 nm, pulse energy up to 0.8 mJ, and repetition rate of 1 kHz were employed. The focused to 280 micron laser beam was raster scanned, using x-y sample translation by computer-controlled motors. The structural properties of the films were characterized by the spectroscopy of Raman scattering, excited by the argon laser (line 514,5 nm). The ablation threshold was found to be of about 65 mJ/cm². When pulse energy density was lower than ~30 mJ/cm² no structural changes were observed. In optimal regimes the films were found to be fully crystallized with fine grain structure, according to the Raman scattering data. Numerical estimates show the pulse energy density was lower than the Si melting threshold, so non-thermal "explosive" impacts may play some role. The possibility of the femtosecond laser pulses to crystallize Si films on glass substrates is shown for the first time. The results obtained are of great importance for manufacturing of polysilicon layers on non-refractory substrates for thin film microelectronics.

Periodic micro-scale domain structures were revealed in congruent LiTaO₃ crystals by point electron beam writing in a scanning electron microscope. Peculiarities of the polarization reversal at point e-beam irradiation of the Y-cuts in LiTaO₃ crystals were investigated. Different distances between the points along Z direction on the polar Y-cuts allow revealing the long planar domain structures. Various periods of the structures were formed by set of different point distances in the X direction.

We investigate the intersubband optical transitions in the InAs/GaSb quantum wells using Burt's envelope function theory and the eight-band model. The self-consistent potential and the lattice-mismatched strain are taken into account to study the effects of bulk inversion asymmetry (BIA) and low interface symmetry on optical matrix elements in structures grown on the InAs substrate along the [001] direction. We have found that both BIA and low symmetry interface Hamiltonian (IH) can result in initially forbidden spin-flip optical transitions or initially forbidden spin conserved optical transitions caused by linearly polarized light. For the light polarization in the plane of the structure, the originally forbidden spin-flip processes can be induced if the light polarization is along the quasiparticle wave vector. However, if light polarization is normal to it, then the originally forbidden spin-conserved processes can be induced. If the light is polarized normally to interfaces along the growth direction [001], then the originally forbidden spin-flip transitions are activated, if in-plane wave vector of the initial quasiparticle states is along the [10] direction. We have also found a considerable lateral anisotropy of absorption caused mainly by BIA induced mechanism. The principal point of this mechanism is the interface contribution to the optical matrix elements due to the material-dependent Kane's B-parameter.

The junction termination extension (JTE) avalanche photodiode (APD) with a ring structure around the active area was built with the use of planar technology processing of float zone silicon doped by neutron transmutation. The main junction and JTE structure have been processed simultaneously, by one implantation step followed by annealing. To set a difference in the doping levels of silicon in the main junction and in the ring structure, the single oxide mask was developed with openings gradually patterned on planar surface. To provide a contact with low resistivity the surface of the main junction was covered by a thin silicon p⁺-layer, grown with MBE (molecular beam epitaxy). The thickness of this "dead" layer no more than 100 nm was obtained. As have been found, the dark current is strongly dependent on the implanted dose, decreasing with decreasing dose from 5x10¹² to 2x10¹² cm^-2. The detectors built with the lowest dose only have the level of dark currents suitable for the measurement of the gain resulted from carrier multiplication. The gain up to 25 was obtained for visible light (λ=600 nm) and up to 7×10³ for single 2.5 MeV α-particles. For 4.3 MeV α-particles the best energy resolution of 330 keV FWHM was obtained. For 22.16 keV x-ray from a ¹⁰⁹Cd source the resolution of 4.7 keV FWHM have been measured, which corresponds to 560 rms electrons noise.

Theoretical and device aspects of emerging graphene electronics is discussed. It was found that the gate screening plays crucial role in electrostatics and transport properties near the neutrality point.

The combined electronic devices may be a first step on a path leading to a development of the nanoscale electronic devices. A complexity of a theoretical description of such combined devices is provided by a necessity to take into account the macroscopic properties of the classical electrical circuits and at the same time the quantum peculiarities of the nanoscale elements. In this work we suggest a conception of an effective self-capacitance of an isolated nanoobject (atom, molecule, nanogranule, quantum dot and etc.) on a basis of the functional dependence of it ground state full energy on it full charge. We reveal that electrical self-capacitance of the molecules depends on theirs topology and qualitatively likewise to a classical electrostatic case. The self-capacitance is proportional to the size of the nanoobject within the group of the sufficiently large similar nanoobjects (with the same form and topology). The functional dependence of the self-capacitance on the number of atoms is determined by nanoobject dimension.

Effects of spatial inhomogeneity for the probability current density j_x (x,z) (or a quantum-mechanical current density ej_x (x,z), e is the electron charge) in the semiconductor 1D nanostructures in the form of joints in the direction of propagation of the electron wave (the x-axis) of narrow rectangular and wide parabolic (on the z-axis) quantum wires (QWRs) (z-axis is the axis of the quantization) have been theoretically studied. It is assumed that such structures in the 2D electron system is realized. The inhomogeneous distribution of the j_x (x,z) arises because of the interference of electron waves spreading in the wide QWR simultaneously in different quantum-confined electron subbands. Special attention is given to effects of spatial reproduction for electron waves in such nanostructures. It is shown that transverse distribution j_x (0,z) existing at the entry of the wide QWR is reproduced with some accuracy at a definite distance Χ₁ from the joint. This picture is reproduced periodically in cross-sections Χ_q = qΧ₁ (coefficients q are integers). The results of numerical calculations of this effect in symmetric structure and its modification in asymmetric (on the z-axis) nanostructure are given. It is shown, in particular, that in asymmetric structures the inverse behavior of probability of the detection of a particle in the quantum-confined subband of the wide QWR from the number of the subband is probable.

Method of creation of nanogaps between metallic (Au) thin-film electrodes using additional evaporation of metal film on relatively wide preliminary created gap is elaborated and realized. A technique of electrodes "overhanging" by dry etching of a substrate is suggested and realized. Optimal etching parameters are found as well. A technique of limitation of the region for additional metal deposition by special additional PMMA mask with narrow window along line of gaps is suggested and realized. It allows a minimization of probability of parasitic gap shunting. It is shown that elaborated method permits iterative approach to prescribed size of a gap by both decreasing of the gap width by deposition of additional metal on the walls of a wide gap and reconstruction of a gap by electromigration in the case of closing of the gap due to excessive metal deposition. Samples with gap's width less than 10 nm were obtained by such reconstruction of a gap.

The technique of metal electrodes' production with sub-10 nm gap on the base of the method of thin film breaking by the effect of metal atoms' electromigration is described. The technology of hard suspended mask and the method of two-shady evaporation are used for the creation of the samples of gold electrodes with thin (of about 10 nm) and narrow bridge between them by standard electron-beam lithography. Study of SEM images of implemented by electromigration gaps have revealed a large number of wide gaps (30-50 nm). It was explained by too large bridge thickness. However, the sub-5nm gaps were obtained on the samples with thinner bridge. I-V characteristics of these gaps have a shape typical for tunnel junctions. Information about gap parameters was obtained from the comparison of experimental I-V curves with theoretical ones calculated on the base of Simmons model of classical tunnel junction. It provide the estimation of parameters of real contact: gap width - 2-5 nm, potential barrier - of about 0.65 eV, the area of contacts - 200-300 nm². The obtained results demonstrate a possibility of creation of molecular single-electron transistor on the base of such nanostructures.

Using the pulsed laser deposition (PLD) technique CdTe layers were obtained on various substrates from the target of compound material, as well as by sequential deposition from single sources of Cd and Te. Electron diffraction analyses have shown that layers deposited from single targets on InSb, KBr substrates crystallize in usual cubic zinc-blende structure of CdTe at the growth temperature of ~150 °C, i.e. significantly lower than in other traditional techniques - MBE, MOCVD, PVD. Layers deposited from CdTe compound target on mica substrates crystallize with hexagonal wurtzite structure; and the single-crystalline growth of layers is observed at 300 °C. It was established that significant decrease (down to 170 °C) in monocrystalline growth temperature for CdTe can be achieved in this case by deposition of initial submonolayer bismuth on mica substrate; the subsequent CdTe layer crystallizes in wurtzite structure with the plane lattice parameter close to that of the bismuth (4.546 Å). Lattice-matched multilayer structures CdTe-Bi-CdTe--- were fabricated based on this technique. The observed peculiarities of dependence of layer structure on the intensity of evaporating laser and substrate temperature is related to the energy state of laser-ablated material plasma and its influence on orienting properties of substrate surface.

Magnetic relaxation effects revealed in Moessbauer spectra and magnetization measurements of nanoparticles are discussed in the framework of a general model for magnetic dynamics of ensemble of single-domain particles. The phenomenological model is based on a generalization of the well-known Stoner-Wohlfarth model within more accurate description of relaxation processes and corresponding time-dependent hyperfine interactions in the magnetic system. This model allows one to treat numerically both Moessbauer spectra in radiofrequency magnetic field and magnetization curves in alternative low-frequency magnetic field as well as temperature demagnetization FC and ZFC curves in a self-consistent way within the same set of physical parameters inherent to the system studied. Besides that, a number of qualitative effects can be explained or predicted within the approach, which include interaction effects, relaxation-stimulated resonances in Moessbauer spectra under radiofrequency field excitation, specific shapes of Moessbauer spectra within precession of particle's uniform magnetization, and asymptotic high-temperature magnetization and susceptibility behavior different from the classical Langevin's high-temperature limit for ideal superparamagnetic particles. Corrections to the above-mentioned effects within more general models based on the Landau-Lifshitz-Gilbert or Braun kinetic equations are also discussed.

Since the discovery of giant magnetoresistance a great effort was made for the investigation of field-induced transitions in magnetic multilayers with antiferromagnetic exchange between magnetic layers. Two peculiarities increase the complexity of this problem. At first, the exchange energy, the anisotropy energy, and the Zeeman energy are of the same order of magnitude. At second, the values of critical fields strongly depend on the number of magnetic layers (finite size effect). Despite of a long time history of these investigations the method of analytical calculation of finite size effect is absent. In the present paper the regular method of analytical calculation of critical field values in antiferromagnetic multilayers is developed. The method is based on the using of finite difference technique. Heisenberg and biquadratic exchange interactions and uniaxial anisotropy are taken into account. Using this technique we investigate the field-induced transition from angle phase to the ferromagnetic phase for arbitrary number of magnetic layers. Exact analytical expression of the critical field value is obtained. This expression shows that critical field monotonically increases with increasing layers number. The dependence of critical field value on magnetic field orientation is discussed. Also the investigation of the stability of antiferromagnetic phase under the action of external magnetic field is made. It is shown that in this case the behavior of magnetic superlattice depends on the parity of layers number. For the even layers number the value of critical field does not depend on the layers number. For the odd layers number the value of critical field depends on the orientation of magnetic field with respect to the orientation of magnetization in the boundary layers. If external magnetic field is directed along the direction of magnetization in the boundary layers the value of critical field is decreased with increasing layers number. If external magnetic field is directed in the opposite direction the value of critical field is increased with increasing layers number. To elucidate the physical mechanism giving rise to the difference in critical field values a numerical simulation was made. It shows that difference in critical field values is concerned with nucleation of the surface spin-flop phase near the surface of magnetic superlattice if the external magnetic field has the opposite direction with respect to the orientation of magnetization in the boundary layer. As external magnetic field is increased the surface spin-flop phase spread out over full volume. Thus the surface spin-flop phase serves as mediator between antiferromagnetic and angle phases and it explains the difference in critical field values. The possible use of developed technique for the investigation of field-induced transitions in ferrimagnetic multilayers is discussed.

Iron film growth conditions and films properties on A- and R- sapphire surfaces were investigated. The best growth conditions were achieved at temperatures about 250 - 300°C. A 10 nm Mo seed layer on the A-sapphire surface improves Fe film morphology and film roughness becomes less than 1 nm. As a result epitaxial Fe (011) films with high residual electron mean free paths (about 0.5 mkm) were grown on the A-sapphire surface. These films can be used for ballistic ferromagnetic planar nanostructures fabrication. The magnetic domain configuration of epitaxial iron nanostructures shaped as bridges and crosses depending on the orientation relative to the easy magnetic axis Fe [100] was examined. If the long side of the bridge is directed along the easy magnetic axis, the single domain structure state is easily reached up to maximum structure width about 2 mkm. A stripe domain structure can be observed when the easy magnetic axis is normal to the long rectangular structure axis. The structure orientation at some angle with the easy magnetic axis leads to a magnetic domain configuration along the easy axis is independent of the structure size down to a structure width ~0.5 mkm and depends only on the easy axis direction. The single domain state can be obtained in structures with a width less than 0.5 mkm. The cross-type structures may have only a two-fold symmetry magnetic configuration. Trapeziform domains were found in structures directed along the hard Fe magnetic axis.

By means of magnetoresistance and magnetic force microscopy (MFM) measurements the electrical and magnetic properties of ferromagnetic strips and crosses of hybrid nanostructures Py/Mo (Py - permalloy (Ni 80 Fe 20), Mo - molybdenum) have been investigated. Unusual behaviour of the cross magnetoresistance in these nanostructures was found. The dependences of the cross resistance against external magnetic field have the image of the curves with hysteresis typical for anisotropic magnetoresistance (AMR). However, they demonstrate anomalous magnetoresistance dependence on orientation relatively to the axis of the ferromagnetic arm of the cross. In the transverse magnetic field (perpendicular to the axis of the ferromagnetic arm) magnetoresistance has two minima as for the longitudinal AMR. In the longitudinal magnetic field (parallel to the axis of ferromagnetic arm) there are two maxima, typical for the transverse AMR. The value of the effect is about one percent. The AMR of the strip has the supplementary extremes, indicated on a more complex character of magnetization reversal, depended on the shape of nanostructures. The data of the magnetoresistance measurements are compared with MFM images.

Molecular-beam (MBE) epitaxy of silicon on sapphire (1 1 02) was studied by growing films from 30 nm up to 1 mkm thick in ultra-high vacuum (10^-7 Torr). The substrate temperature during deposition was 450-750°C. Before deposition the high-temperature (~1400°C) substrate annealing procedure was performed. The growth rate was ~2.5-5Å/s. Surface morphology was studied by means of atomic-force microscopy (AFM), structure of the films was controlled by reflection of high energy electrons (RHEED). Formation of the three-dimensional islands (clusters) of two types (square shape and hemispherical) was observed. The "square" clusters appeared mostly under low growth temperatures (450-600°C). The hemispherical clusters had the larger sizes and were observed under long deposition times and high growth temperatures. The minimal thickness of continuous Si film was about 32 with 2.0 nm surface roughness. The electron diffraction patterns contained spots, what proved the single-crystal structure of the Si layer.

The formation of the TiN/CoSi₂ bilayer from the Co/Ti/Si structure in a non-isothermal reactor was analyzed by Auger electron spectrometer. It was concluded in determination of concentration distribution of the Co, Ti and Si atoms in the samples before and after the thermal annealing. The annealing time was equal to 10 sec and 1 min. It is noted, that both the Co and Ti atoms diffuse at opposite directions. The cobalt atoms move through titanium film to silicon surface. On the contrary, the titanium atoms move through cobalt film to surfaces of the sample. After thermal annealing of the samples, located by silicon surface to the heating source, the formation TiN and CoSi₂ phases is fixed. The forming of these phases at the annealing of the samples, located by structure Co/Ti to the source of the heating, does not occur. This is conditioned by brilliant metallic surface reflecting of radiated energy. As a result necessary temperature regime, for the formation of both of these phases TiN and CoSi₂, was not reached.

Zirconium oxide (ZrO₂) films have been deposited on cleaned and heated p-type Si (100) substrates by electron-beam evaporation technique. It is shown that the intermediate SiO₂ layer on ZrO₂/Si interface is absence. The W/YSZ/Si and Mo/YSZ/Si structures with 3÷20-nm-thick dielectric layers were formed by electron-beam evaporation technique. The fixed charge densities in 3-nm-thick YSZ layers are 3x10¹⁰ - 3.7x10¹⁰cm², leakage current density at a voltage -1V achieves ~7,9•10^-7Α/cm².

There are proposed methods of film heterostructures deposition by means of a magnetron and an ion source. They make it possible to regulate compositions of film structures and also other parameters, for example refractive indexes, conductivity. It is possible to form a film with preliminary demanded values of parameters in the wide range. Refractive index of films on the base of Ti was varied in the range of values 1.70 . n . 2.45. Conductivity of film structures changed from metal up to high resistance isolator. Diamond . like carbon , SixCy, SixOy, SixOyNz and other kinds of film structures were made by means of an ion source with a cold cathode directly from an ion beam. Refractive index value changed in the range of 1.40 . n . 2.65. It is modified a laser method of end point detection of transparent film deposition process after reaching of a demanded thickness. It has been modified a laser scanning system for wafers and semiconductor structures surfaces investigation. This system can measure parameters at 10000 points on square up to 100 x 100 mm2. It can be used for computer modelling, solar cells and silicon-oninsulator (SOI) structures investigations.

The article includes the results of the study of image formation in atomic force microscope (AFM). The influence of radius and angle characteristics of cantilever tip as well as the relief of the surface studied on the signal waveform is shown. The authors demonstrate the techniques of AFM calibration and direct measurement of linear sizes of trapezoid structures including the line width with the use of AFM signal and its first derivative. There were obtained the equations establishing relations the sizes of trapezoid structures with the sizes of test segments chosen on AFM signals.

The single elements of relief (protrusions and steps) fabricated by anisotropic etching of the surface of the silicon wafer congruent the crystallographic plane (100) in the scanning electron microscope have been studied. The image registration in the low energy secondary electron collection mode was carried out, and the influence of the probe electron energy and its diameter on the microscope signal formation by relief elements scanning was studied. The electron beam energy varied at the range of 0.3 - 20 keV, the probe diameter changed in the limits of 14 - 500 nm. The widths of upper bases of protrusions varied within 14 - 500 nm. The correlation analysis of experimental results, carried out by the authors, demonstrate high quality of the structures studied.

Using investigation of micromirrors of space light modulator as well as other objects as an example, real capabilities of the domestically produced interference microscope MII-4 equipped with a digital camera are demonstrated. It is shown that a resolution in a vertical direction provided by the microinterferometer lies in a sub nanometer range.

Characterisation of three-layer dielectric embedded into MDS-structure (Metal-Dielectric-Silicon) was provided in the dark and under light illumination. In the dark, increasing of differential capacitance, simultaneously, with variation of differential conductivity of MDS-structures was detected. In the light strong changing of capacitance part of impedance was firstly observed, demonstrating decreasing almost to zero values and restoring up to maximal values in narrow bands of voltage applied. Variation of capacitance exceeds significantly so called dielectric layer capacitance, what interpreted as carriers exchanging between substrate and electronic states in SiN_x probably due to three-layered kind of its nature.

Nondestructive diagnostic method was developed for epitaxial heterostructure quality prediction. Such prediction is very important for production of some types HF FETs. The various heterostructure modifications grown on sapphire substrates have been analyzed. These structures were grown by MOCVD method from the metalorganic compounds with variation of Al mol-content in the AlGaN layers in wide limits (0,05>x>0,42). The C-V measurements were used for electrical analysis Al_xGaN\GaN films with Hg-probe at frequencies 1MHz, 10kHz, 1kHz, 100Hz. Both electrodes (Hg-probe and Common) were placed on the same active surface of the heterostructure. C-V curves looks like a <<step>> with C_max and C_min independent of voltage. The test FETs were produced for comparison with results of C-V measurements. It was found that the voltage of transition from C_max to C_min corresponds with pinch-off voltage (V_p.f.) of test FETs. The values of C _min> 1pF points to the presence of parasitic conducting layer in buffer layer.

The analysis algorithm of quasistatic C-V-characteristics of MIS-structures in the range of the depletion of the semiconductor surface of main carriers of the charge are developed. This algorithm provides the quantitative determination of a concentration of doping impurity, the <<flat bands>> voltage and efficient values of a capacity and a thickness of a gate insulator. On this data, obtained within the framework of the single experiment on the Al-SiO₂-(100) n-Si MOS-structure, dependencies of ψ_s(V_g),Q_s[ψ_s(V_g)] and V_i(V_g) (where ψ_s and Q_s - a surface potential and a density of the surface charge in n-Si, V_g - a gate potential, V_i - a voltage drop on an oxide) are reduced. These dependencies without any a priori information about the state of the electronic gas under the strong accumulation or deep inversion are found. Experimental curves of ψ_s(V_g) and Q_s[ψ_s(V_g)] are possible considered as the criterion of the correct of the theory of the semiconductor space charge region, take into account electron gas degeneration and quantum confinement effects, as observed maximum layered densities of electrons and holes exceed 10¹³cm^-2. The dependency of V_i(V_g) necessary to use under investigations of the conductivity through gate insulators, particularly, in cases their small and ultrasmall of the thickness, when the leakage current is defined basically by the tunnel effect.

Parasitic bipolar effect can significantly decrease SEE tolerance of modern deep submicron bulk and SOI CMOS devices due to amplification of charge collected in interaction between silicon and single ionizing particles. This article is dedicated to studies of bipolar effect phenomenon in modern SOI CMOS technologies and to increasing SEE tolerance without technology modifications.

The sensitivity of sub-100 nm devices to microdose effects, which can be considered as intermediate case between cumulative total dose and single event errors, is investigated. A detailed study of radiation-induced leakage due to stochastic charge trapping in irradiated planar and nonplanar devices is developed. The influence of High-K insulators on nanoscale ICs reliability is discussed. Low critical values of trapped charge demonstrate a high sensitivity to single event effect.

Complex investigations of the epitaxial layers quality in Al_xGa_(1-x)N/GaN heterostructures grown on sapphire substrates by MBE and MOCVD methods were carried out with using of Double Crystal Diffractometry and optical microscope methods. The mobility and surface concentration of charge carriers were analyzed by means of the Hall effect method. It is shown that the blocks sizes of epitaxial layer grown by MBE method were smaller than separate blocks sizes grown by MOCVD method. On the base of Hall-effect measurements and X-diffractometry investigations it is determined band gap changing from Al-mol content into ALGaN layer. It has been received an experimental dependence of charge carries concentration from Al-mol content into analyzed structures.

The process of rapid annealing of implanted silicon wafers under non-isothermal conditions has been studied. It was established, applied temperature gradient ∇T gives rise to relative shift of non-isothermal concentration profiles, corresponding two opposite signs of this gradient. The value of the shift increases with increasing the value of ∇T and nonlinearly depends on the time durations of annealing process. It was established too, the dopant diffusion coefficient increases with increasing the ∇T in the vertical direction in the wafer regardless of ∇T sign. The estimation of thermodiffusion parameters was made. The coefficient of mass transfer of intrinsic non-equilibrium defects was introduced into the phenomenological equations of irreversible thermodynamics to explain anomalous high values of heat of transport. This coefficient allows for influence of generation-recombination processes related to intrinsic non-equilibrium defects on the dopant diffusion. The experimentally observed high values of diffusion coefficients and heat of transport are explained using this coefficient.

In the present work, the matter of stabilization of silicon conductivity versus temperature is discussed for neutron transmutation doped FZ silicon with point radiation defects. It is shown that divacancies introduced by electron irradiation decrease the room-temperature conductivity of the material, making the resistance simultaneously more stable to temperature variations in the temperature range 20 to 160°C. The discrepancy between experimental and simulated data was evaluated and corrected assuming the presence of a deep-level center with energy E_c-0.6 eV in the forbidden gap. As a result of the study, power resistors have been manufactured exhibiting less than 10-% variation of their resistance from nominal value in the indicated temperature range.

The dual collector lateral bipolar magnetotransistor manufactured in the well with an external connection of contacts to the well and substrate has been investigated. Modern methods of device-technological simulation have been used to model the distribution of charge carriers, current densities, and recombination velocity. It is shown that bipolar magnetotransistor in the well have negative relative magnetic sensitivity due to the volumetric recombination mechanism.

The experimental research initial disbalance potential collectors from the scheme of inclusion two-collector lateral bipolar magnetotransistor (BMT) NPN-type, generated in a well is lead. By means of device-technological modelling the mechanism of occurrence initial disbalance is investigated and the way of its reduction initial disbalance is certain at maintenance of preservation of high sensitivity. The choice of the operating mode bipolar magnetotransistor is based on the distributions of the emitter injected electron currents in two symmetrical base electrodes - two contacts to the base-well, in two contacts to a substrate, in two collectors. The mode of magnetotransistor based on the influence of a magnetic field. Reduction of initial disbalance allows to increase relative size of a output valid signal ΔU= U_C1(B) - U_C2(B) - U_C1(0) + U_C2(0).

Self-formed conducting nanostructures are generated in "sandwich"-structures Si-SiO₂-W with an open surface of an insulating slit because of an electroforming and find out effects of switching and memory. Switching is carried out by voltage impulses of certain amplitude and duration. Shortcomings of the "diode" cell of memory based on such structures are discussed and design of the "transistor" cell for improvement of its characteristics is offered. In this design the structure of the bipolar transistor is formed by the introduction of an additional n⁺-layer of silicon (emitter) so the memory element is integrated into the emitter. It has allowed to improve essentially all operational characteristics of a memory matrix: to reduce leakage currents; to increase breakdown voltage and so to increase recording speed due to using of faster switching process which requires higher voltage; to increase dimension of a matrix because of reduction of operating currents for silicon buses in inverse proportion to a transistor current gain; to increase stability of characteristics. Results of experimental samples tests of a memory matrix based on a "transistor" cell are presented.

Electrostatic high energy micromotor based on the ferroelectric films is studied as applied to microelectromechanical devices operating in vibrational mode. It is shown that the micromotor can be efficiently used in high frequency micromechanical vibrators that are used in high energy MEMS devices, such as micropumps, microvalves, microinjectors, adaptive microoptic devices etc.

An improved mathematical model in order to study mechanical behavior of an extensible microplate subjected to nonlinear electrostatic pressure was presented. In this model, the effect of stretching due to fixed boundary conditions and residual stresses because of fabrication process on static instability of the microplate was studied. The derived nonlinear partial integro-differential governing equation considering stretching and residual stresses effects, using of Step-by-Step Linearization Method (SSLM), was linearized. By applying the finite difference method (FDM) to a rectangular mesh, the linearized equation was discretized. By considering the stretching stresses effect, the present mathematical model shows a highly reasonable prediction of divergence instability as compared with previous existing model. The obtained results show that the residual stresses have considerable effects on Pull-in phenomenon. Axial stresses due to stretching and tensile residual stresses increase pull-in voltage and compressive residual stresses decrease it.

The design the new principle of electromechanical energy conversion that allows one to carry the electrome-chanical energy conversion in the nanometer gap, and significantly (up to two orders of magnitude) increase MEMS specific energy output, operation speed and power. The energy conversion takes place in the nanometer gap (5 - 200 nm), when the electric energy accumulated during reversible electrostatic pressing of the free metallic film (moving electrode) to the surface of the thin crystalline dielectric (ferroelectric film, FF) with high dielectric permeability ε (more than 3000-5000) is transformed into mechanical energy . The tension of the metallic film caused by electro-static forces in converted into the mechanical motion of the moving element of the device. With this approach, the specific energy output of 0.3 - 1 10^-6 J/mm² and driving force of 0.01-0.3 N can be achieved starting from the first microseconds of the voltage pulse. An experimental investigation of new electromechanical energy converter is per-formed.

A MEMS capacitive-type sensor is basically an electrostatic transducer that depends on electrical energy in terms of constant voltage (voltage drive) or constant charge storage (current drive) to facilitate monitoring of capacitance change due to an external mechanical excitation, such as force, acoustical pressure or acceleration. Microfabricated cantilever beams are widely used in MEMS capacitive-type sensors as the sensing element1. One such representative of movable microdevices are the surface micromachined mechanical resonators that come in many geometrical configurations, such as laterally movable comb resonators, laterally and vertically movable beam resonators, and torsional resonators. The successful design and fabrication of these devices requires computer-aided design _CAD_ simulation tools capable of accurately simulating the electromechanical performance of MEMS devices. Accurate simulations are critical for the expeditious development of commercial products at reasonable cost. All CAD simulation tools require accurate measurements to verify models and to provide the values of the constants used in the models. In particular, in the case of micromechanical resonators, it is challenging to determine the mechanical properties of both static and dynamic behaviors of such micromechanical resonators2,3. The testing of these movable structures is usually performed using electrostatic excitation and detection by capacitive, piezoresistive, and optical methods4. This paper presents the design, fabrication and testing of capacitive aluminum resonator with various movable membrane geometries, fabricated with surface micromachining.

An analytical two-band k•p model for the conduction band of silicon is compared with the numerical nonlocal empirical pseudo-potential method and the sp³d⁵s* nearest-neighbor tight-binding model. The two-band k•p model gives results consistent with the empirical pseudo-potential method and describes the conduction band structure accurately. The tight-binding model overestimates the gap between the two lowest conduction bands at the valley minima, which results in an underestimation of the non-parabolicity effects. When shear strain is introduced, the two-band k•p model predicts an analytical expression for the strain-dependence of the band structure, which is in good agreement with results of pseudo-potential simulations.

Coupled process and device simulation has been applied to investigate the physical processes which determine the performance and scaling properties of fully depleted thin-silicon-body SOI based (FD-SOI) NMOS transistors at gate lengths of 40 nm and below. A comparison of simulation results with measurements of electrical characteristics of the FD-SOI NMOS transistors showed that the electrical performance of such transistors can only be reproduced by simulation, if contact resistances, ballistic electron transport, quantum mechanical depletion of the electrons near the gate dielectric, and mechanical stress are accounted for. The mechanical stress in thin-silicon-body SOI transistors is simulated as a result of two silicidation processes: the source/drain contact CoSi₂ silicidation and NiSi gate silicidation. The resulting stress in the silicon channel at the end of processing is tensile along the gate length, non-uniform and reaches maximum values in excess of 1GPa. The simulations show that mechanical stress as well as the contact resistances increase for scaled FD-SOI transistors, therefore a careful optimization of mechanical stress and an engineering of contact resistances are mandatory for scaled FD-SOI devices.

The all-quantum program for 3D simulation of an ultra-thin body SOI MOSFET is overviewed. It is based on Landauer-Buttiker approach to calculate current. The necessary transmission coefficients are acquired from the self-consistent solution of Schrodinger equation. The latter is stabilized with the help of expanding the wave function over the modes of transversal quantization inside the transistor channel. The program also contains a domain for one-dimensional classical ballistics intended for calculation of the initial state for subsequent all-quantum simulation. This is a significant point of our approach as the straightforward procedure of the self-consistent solution of Schrodinger equation from the very beginning is diverging or, at least, extremely time-consuming. The main goal of all-quantum simulation is to clarify the impact of interference on charged impurities and quantum reflection from self-consistent potential on I-V curve reproducibility for different randomly doped transistors in a circuit. The 10nm technology node tri-gate (wrapped channel) structure with 2nm silicon body was used in simulation. 20 discrete impurities were dispersed by the source and drain contacts to imitate the same doping. The most important feature we demonstrate is a smoothness of I-V curves in spite of beforehand apprehension. The next peculiarity we came across was that the current spanned within 10% for different discrete impurity realizations. These results manifest that the reproducibility of nanotransistors could be fairly good to make ultra-large integrated circuits still feasible. We have also made a comparison with simulations based on drift-diffusion model.

The Monte Carlo model of electron transport in SOI MOSFETs is proposed. Both 2D and 3D conditions are considered. The Poisson equation and boundary conditions are presented for every case. Fully depleted SOI MOSFETs and partially depleted SOI MOSFETs are contradistinguished. The values of electron current as well as drift velocity in different parts of SOI MOSFETs channel are calculated by means of the Monte Carlo simulation. The SOI MOSFETs with the channel length equal to 0.5, 0.25 and 0.1 μm as well as the channel depth equal to 10, 20, 100, 200, 1000 nm are studied. Drift velocity as a function of the channel depth is obtained. It is shown that the function has a peak at the channel depth equal to 20 nm.

Monte Carlo model of electron transport in short channel MOSFETs at different temperatures ranging from -50° C (223 K) to +50° C (323 K) are proposed. MOSFETs with the channel length equal to 0.5, 0.2 and 0.1 μm are studied by using Monte Carlo simulation. Three mechanisms of temperature effect on electron properties are discussed for studied devices. Temperature influence on the values of drift velocity, mobility, electron energy and electric field in different parts of conducting channel is dealt with at studied conditions. It is shown that for MOSFET with the channel length equal to 0.1 μm obtained temperature dependencies demonstrate an appreciable divergence from ones for MOSFET with channel length equal to 0.5 μm.

In this paper modeling of powerful GaAs Metal Semiconductor Field Effect Transistor (MESFET) was carried out, I-V characteristics with breakdown voltage of MESFET were obtained. With the help of this model optimal characteristics of structure for power MESFET were determined. Recommendations for value of Schottky barrier were proposed. Also breakdown voltage was calculated. More over temperature dependencies of I-V characteristics of MESFET were studied.

In this work we present the results of simulation of vertical MOS transistor with electrically variable shallow junctions in ISE TCAD. Transistor with fully silicided gate electrodes, two heavy doped delta-layers in the channel region and ZrO₂ as a gate dielectric has been simulated. The simulation used different carrier transport and mobility models. High values of on-state current have been obtained during the simulation process (~1.2 mA/μm). Different voltage regimes for middle and side gates have been assumed. Values of direct leakage current from drain to source got from simulation are relatively low and amount to approximately 0.02 μA/μm². These values show that use of electrically variable junctions and doped delta-layers really suppresses short-channel effects and reduces direct leakage current from drain to source. Technology of electrically variable junctions allows to employ vertical transistor into high-performance logic applications.

An increasing complexity of mixed analog-digital signal circuits requires optimization at higher hierarchical level. However, evolutionary optimization of mixed analog-digital signal circuits at the system level results in huge computational costs. A key to manage these computational complexities of evolutionary circuit design is an application of flexible fitness functions evaluation schedules. In this paper we compare the static, dynamic, and co-evolution fitness function evaluation schedules for multi-objective optimization of mixed analog-digital signal circuits at the system level on the base of the univariate marginal distribution algorithm. Experiments for our symmetry recognition circuit benchmark chosen indicate that the dynamic fitness function schedule is a good compromise between computational costs and optimization efficiency.

A new model for copper chemical mechanical polishing (CMP) process in the (K₃Fe(CN)₆+NH₄OH) slurry is developed in the work. A distinctive peculiarity of the model is quantitative consideration of the kinetics of the passivating layer growth and accounting of its action on the polishing rate. In accordance with the model the main stages of copper CMP are the Cu⁺ ion diffusion and tunneling of copper conductivity electrons through the passivating layer towards its interface with the slurry, as well as chemical reactions in the slurry at the passivating layer surface resulting in growth of its thickness and formation of dissoluble compounds removed from the system. The closed set of equations of the CMP kinetics is derived. Its solutions are obtained for the steady-state regime as to the chemical reactions in two limiting cases when Cu⁺ ion diffusion through the passivating layer predominates over their electromigration or vice versa. The estimates of the CMP rate and limiting values of passivating layer thickness for these two modes are carried out and give reasonable results which correlate with experimental data.

In the work the influence of vacancy clusters on strength properties of solid-solid interfaces is studied on the base of the thermodynamic approach. The conditions of cluster formation at the interface are obtained and analyzed. A model describing the influence of vacancy clusters in the contacting materials on the surface tension coefficient γ₁₂ of the interface is also developed. The dependences of the contribution of nonequilibrium vacancy clusters in γ₁₂ on a vacancy supersaturation, cluster concentrations, the number of vacancies in clusters, and temperature are analyzed.

The results of calculation of electron drift velocity in GaAs-in-Al₂O₃ and GaAs-in-AlAs quantum nanowires as well as the electric current in the armchair single-wall carbon nanotubes versus time are presented at various electric fields applied along the structures and different temperature.