Electron mobility in the inversion layers of fully depleted SOI films

The dependences of the electron mobility μ_eff in the inversion layers of fully depleted double–gate silicon-on-insulator (SOI) metal–oxide–semiconductor (MOS) transistors on the density N _e of induced charge carriers and temperature T are investigated at different states of the SOI film (inversion–accumulation) from the side of one of the gates. It is shown that at a high density of induced charge carriers of N _e > 6 × 10¹² cm^–2 the μeff(T) dependences allow the components of mobility μ_eff that are related to scattering at surface phonons and from the film/insulator surface roughness to be distinguished. The μ_eff(N _e ) dependences can be approximated by the power functions μ_eff(N _e) ∝ N _e ^?n . The exponents n in the dependences and the dominant mechanisms of scattering of electrons induced near the interface between the SOI film and buried oxide are determined for different N _e ranges and film states from the surface side.     

Performance characteristics of <Emphasis Type="Italic">p</Emphasis>-channel FinFETs with varied Si-fin extension lengths for source and drain contacts

The length of Source/Drain (S/D) extension (L_SDE) of nano-node p-channel FinFETs (pFinFETs) on SOI wafer influencing the device performance is exposed, especially in drive current and gate/S/D leakage. In observation, the longer L_SDEpFinFET provides a larger series resistance and degrades the drive current (I_DS), but the isolation capability between the S/D contacts and the gate electrode is increased. The shorter L_SDE plus the shorter channel length demonstrates a higher trans-conductance (G _m ) contributing to a higher drive current. Moreover, the subthreshold swing (S.S.) at longer channel length and longer L_SDE represents a higher value indicating the higher amount of the interface states which possibly deteriorate the channel mobility causing the lower drive current.     

Power switching transistors based on gallium nitride epitaxial heterostructures

The results of the development of power switching transistors based on epitaxial gallium nitride heterostructures to create an energy-efficient conversion technique are presented. The developed powerful GaN transistor operates in enrichment mode with unlocking threshold voltage V _th = +1.2 V and a maximum drain-source current I _ds = 0.15 A/mm at the drain-source voltage V _ds = +8 V. The drain-source breakdown voltage in the closed state is V _b = 300 V at the drain-source distance L _ds = 8.5 μm and drain-source voltage V _ds = 0 V.     

1/<Emphasis Type="Italic">f</Emphasis> noise in metal insulator semiconductor transistors with different conductivity types and different topological dimensions of the channel

An experimental investigation of 1/f noise in metal insulator semiconductor transistors with different types of channel conductivity and different topological sizes of the gate region is presented. The transistors are produced on the basis of the standard 1.2-μm complementary metal oxide semiconductor technology and can be used as reading elements in 2D multielement detectors of infrared radiation. The level of 1/f noise is determined in relation with the channel conductivity type. It is shown that the p-channel transistors exhibit a lower level of 1/f noise compared to that of the n-channel transistors (by about one order of magnitude). The dependence of 1/f noise on the topological size of the gate region is obtained. The transistors with the smallest channel width are found to produce the highest noise level.     

High-voltage MIS-gated GaN transistors

Transistors with a high electron mobility based on AlGaN/GaN epitaxial heterostructures are promising component types for creating high-power electronic devices of the next generation. This is due both to a high charge-carrier mobility in the transistor channel and a high electric durability of the material making it possible to achieve high breakdown voltages. For use in power switching devices, normally off GaN transistors operating in the enrichment mode are required. To create normally off GaN transistors, the subgate region on the basis of p-GaN doped with magnesium is more often used. However, optimization of the p-GaN epitaxial-layer thickness and doping level makes it possible to achieve a threshold voltage close to V _th = +2 V for the on-mode of GaN transistors. In this study, it is shown that the use of a subgate MIS (metal–insulator–semiconductor) structure involved in p-GaN transistors results in an increase in the threshold voltage for the on-mode to V _th = +6.8 V, which depends on the subgate-insulator thickness in a wide range. In addition, it is established that the use of the MIS structure results in a decrease in the initial transistor current and the gate current in the on mode, which enables us to decrease the energy losses when controlling powerful GaN transistors.     

Study of RF transistor with submicron T-shaped gate fabricated by nanoimprint lithography

The results devoted to the development of a method for creating an RF transistor, in which a T-shaped gate is formed by nanoimprint lithography, are presented. The characteristics of GaAs p-HEMT transistors have been studied. The developed transistor has a gate “foot” length of the order of 250 nm and a maximum transconductance of more than 350 mS/mm. The maximum frequency of current amplification f _t is 40 GHz at the drain-source voltage V _DS = 1.4 V and the maximum frequency of the power gain f _max is 50 GHz at V _DS = 3 V.     

On the influence of a Si single-crystal real surface on the low-frequency internal friction and the behavior of an effective shear modulus

The spectra of low-frequency internal friction and the behavior of the effective shear modulus G_eff(T) in a Czochralski grown silicon single crystal were investigated after mechanical and chemical-mechanical treatment of the surface and after natural aging at room temperature over a long period of time. The mechanical treatment of the Si surface was shown to initiate unusual structural phase transitions accompanied by an inverse hysteresis in the G_eff(T) dependence. Such a behavior of G_eff(T) is assumed to be associated with the formation of incommensurate phases. It was also found that aging at room temperature for more than 10000 h leads to the onset of decomposition of the supersaturated solid solution of oxygen in silicon and to the generation of substantial stresses that initiate structural phase transitions in the samples.     

Mesoscopic fluctuations in the conductance of silicon-based MOS structures with a doped surface layer

In n-Si-based metal-oxide-semiconductor structures with a boron-doped surface layer, at the temperature T = 77 K, mesoscopic fluctuations in nondiagonal resistance-tensor component R _xy are discovered under hole depletion of the Si: B layer caused by the field effect. A sharp increase in the fluctuation in R _xy is discovered during the study of a shift from the 3D weakly perturbed hole transport to the quasi-2D percolation conduction on the back boundary of the Si: B layer. The fluctuations in R _xy are due to the restructuring that occurs in the percolation cluster when the hole shielding of the fluctuation potential of ionized acceptors is weakened. Correlation length L _c of the percolation cluster, which determines the boundary of the transition from the macroscopic to the mesoscopic system, has been estimated experimentally. Good agreement between the estimates and the computational results is observed for L _c ranging from ≈ 10 nm to ≈ 1 μm.     

Method for narrowing the directional pattern of an InGaAs/GaAs/AlGaAs multiwell heterolaser

A semiconductor laser with a new waveguide is developed. It allows significant narrowing of the directional pattern (to 4° in the plane perpendicular to the p–n junction). In the used waveguide, the minimum excess of the effective refractive index n_eff of the excitation mode over the substrate refractive index n_s (n_eff–n_s ? 1) is provided by selecting the thickness of Al_0.3Ga_0.7As confinement layers, which significantly increases the waveguide mode size and leads to directional-pattern narrowing.     

Work Function Tuning and Doping Optimization of 22-nm HKMG Raised SiGe/SiC Source–Drain FinFETs

The basic requirements on process design of extremely scaled devices involve appropriate work function and tight doping control due to their significant effect on the threshold voltage as well as other critical electrical parameters such as drive current and leakage. This paper presents a simulation study of 22-nm fin field-effect transistor (FinFET) performance based on various process design considerations including metal gate work function (WF), halo doping (N _halo), source/drain doping (N _sd), and substrate doping (N _sub). The simulations suggest that the n-type FinFET (nFinFET) operates effectively with lower metal gate WF while the p-type FinFET (pFinFET) operates effectively with high metal gate WF in 22-nm strained technology. Further investigation shows that the leakage reduces with increasing N _halo, decreasing N _sd, and increasing N _sub. Taguchi and Pareto analysis-of-variance approaches are applied using an L₂₇ orthogonal array combined with signal-to-noise ratio analysis to determine the best doping concentration combination for 22-nm FinFETs in terms of threshold voltage (V _t), saturation current (I _on), and off-state current (I _off). Since there is a tradeoff between I _on and I _off, the design with the nominal-is-best V _t characteristic is proposed, achieving nominal V _t of 0.259 V for the nFinFET and ?0.528 V for the pFinFET. Pareto analysis revealed N _halo and N _sub to be the dominant factor for nFinFET and pFinFET performance, respectively.     

Increase the threshold voltage of high voltage GaN transistors by low temperature atomic hydrogen treatment

High-electron-mobility transistors (HEMTs) based on AlGaN/GaN epitaxial heterostructures are a promising element base for the fabrication of high voltage electronic devices of the next generation. This is caused by both the high mobility of charge carriers in the transistor channel and the high electric strength of the material, which makes it possible to attain high breakdown voltages. For use in high-power switches, normally off-mode GaN transistors operating under enhancement conditions are required. To fabricate normally off GaN transistors, one most frequently uses a subgate region based on magnesium-doped p-GaN. However, optimization of the p-GaN epitaxial-layer thickness and the doping level makes it possible to attain a threshold voltage of GaN transistors close to V _th = +2 V. In this study, it is shown that the use of low temperature treatment in an atomic hydrogen flow for the p-GaN-based subgate region before the deposition of gate-metallization layers makes it possible to increase the transistor threshold voltage to V _th = +3.5 V. The effects under observation can be caused by the formation of a dipole layer on the p-GaN surface induced by the effect of atomic hydrogen. The heat treatment of hydrogen-treated GaN transistors in a nitrogen environment at a temperature of T = 250°C for 12 h reveals no degradation of the transistor’s electrical parameters, which can be caused by the formation of a thermally stable dipole layer at the metal/p-GaN interface as a result of hydrogenation.     

The temperature and concentration dependences of the charge carrier mobility in PbTe-MnTe solid solutions

The temperature dependences of conductivity (σ), the Hall coefficient (R_H), and the charge carrier mobility (μ_H) for cast and pressed samples of PbTe-MnTe solid solutions (0–2.5 mol % of MnTe) were studied in the temperature range of 80–300 K. The mobility of cast samples varies only slightly in the temperature range of 80–140 K. At higher temperatures, μ_H decreases according to the power law μ_H=aT^?v. The mobility μ_H of pressed samples exponentially increases with temperature in the temperature range of 100–160 K, which is related to the energy barriers ΔE_a formed by oxide films at the grain boundaries. The dependences of μ_H, ν, and ΔE_a on the MnTe content have anomalies in the composition range of 0.75–1.25 mol %, which are related to the concentration phase transition of the percolation type.     

Increase in the rate and discretization of the kinetics of isothermal surface generation of minority charge carriers in metal-insulator-semiconductor structures with a planar-inhomogeneous insulator

Surface generation of minority charge carriers in silicon metal-oxide-semiconductor (MOS) structures is efficient only at the initial recombinationless stage. Quasi-equilibrium between surface generation centers and the minority-carrier band is established in a time t ～ 10^?5 s. In the absence of other carrier generation channels, an equilibrium inversion state at 300 K would need t = t_∞ > 10³ years to become established. In fact, the time t ∞ is much shorter, due to excess-carrier generation via centers located at the SiO₂/Si interface over the gate periphery. This edge-related generation can easily be simulated in an MOS structure with a single gate insulated from Si by oxide layers of various thicknesses. At gate depleting voltages V _g , the role of the periphery is played by a shallow potential well under a thicker oxide, and the current-generation kinetics becomes unconventional: two discrete steps are observed in the dependences I(t), and the duration and height of these steps depend on V _g . An analysis of the I(t) curves allows determination of the electric characteristics of the Si surface in the states of initial depletion (t = 0) and equilibrium inversion (t = t_∞), as well as the parameters of surface lag centers, including their energy and spatial distributions. The functionally specialized planar inhomogeneity of a gate insulator is a promising basis for dynamic sensors with integrating and threshold properties.     

Mobility of minority charge carriers in <Emphasis Type="Italic">p</Emphasis>-HgCdTe films

Temperature dependences of electron mobility in p-Hg_1?xCd_xTe films (x=0.210–0.223) grown by molecular beam epitaxy are investigated. In the temperature range 125–300 K, mobility was found by the mobility-spectrum method, and for the range 77–125 K, it was found using a magnetophotoconductivity method suggested in this study. The method is based on the measurement of the magnetic-field dependence of photoconductivity. The magnetic field is parallel to the radiation and normal to the sample surface. The electron mobility is determined using the simple expression μ _n [m²/(V s)]=1/B _H [T]. Here, B _H is the induction of the magnetic field corresponding to a half-amplitude of the photoconductivity signal under zero magnetic field. In the temperature range 100–125 K, the results obtained by the magnetophotoconductivity and mobility-spectrum methods coincide. For the samples investigated, the electron mobility at 77 K is in the range 5–8 m²/(V s).     

The effects of monovacancies on the terrace width during sublimation from the (111) surface of a diamond-like crystal

Using the three-dimensional Monte Carlo model, the effect of monovacancies on the diffusion exchange between the steps on the (111) surface of a diamond-like crystal during sublimation was investigated. The critical terrace width L_cr (the distance from the step edge to the nearest stable vacancy islands) was determined as a function of the relationship between the energy of adatom formation at the smooth terrace E_n and the energy of adatom desorption from the terrace E_d with retention of the sum of these energies E_sub=E_n+E_d, which characterizes the sublimation of a substance. A well-pronounced maximum exists in the dependences L_cr(E_n) obtained. The formula is suggested which well describes the dependences L_cr(E_n) for various values of E_sub. The extrapolation of the dependence L_cr(E_n) to the value E_sub=4.2 eV, which is characteristic of Si, permits us to compare the model results with the experimental data of other authors. The explanation of a threefold increase in the terrace width, which is observed experimentally at about 1500 K, is suggested.     

A study of recombination centers in irradiated <Emphasis Type="Italic">p</Emphasis>-Si crystals

p-Si samples irradiated with 8-Mev electrons are studied. It is suggested that the multicomponent V₃+O or V₂+O₂ complexes are not recombination centers on the basis of an analysis of the dependences of the minority-carrier lifetime τ, the resistivity ρ, the concentration p, and the Hall mobility μ_H on the temperature of isochronous annealing T_ann. Deep donors with energy levels at ΔE_i=E_v+0.40 eV and the V₃+O₃ and the V₃+O₂ complexes affect the values of μ_H and τ. The curves of isochronous annealing are used to determine the annealing-activation energies E_ann for defects such as K centers, interstitial carbon atoms C_i, the V+B and V₂+O₂ complexes, divacancies V₂, and defects with a level at ΔE_i=E_v+0.20 eV. These energies were found to be equal to E_ann=0.9, 0.25, 1.6, 2, 1.54, and 2.33 eV, respectively.     

Increase in the electron mobility in the inversion channel of a Si-MOS transistor in the case of ion polarization of the gate oxide

The effective mobility of electrons μ* in the inversion n-type channel of a field-effect transistor increases appreciably (as a result of space-charge ion polarization of the gate oxide) from the typical values of ?820 cm² V^?1 s^?1 to the values of ?2645 cm² V^?1 s^?1, which exceed the electron mobility in bulk silicon. After polarization, the sheet concentration of Na⁺ ions at the SiO₂/Si interface exceeds 6 × 10¹³ cm^?2. The ions are almost completely neutralized by electrons in the inversion channel. As temperature T is decreased in the range from 293 to 203 K, μ* increases according to the law μ* ∝ T ^?0.82. Apparently, the observed dependence μ*(T) is caused by the combined scattering of electrons by roughness of the Si/SiO₂ interface surface, phonons, and the interface states. Depolarization of the oxide reverts μ* to the initial value. Anomalously large values of μ* are assumed to be either a consequence of the origination of pronounced structural stresses in the surface Si layer due to the oxide polarization or a result of a phase reconstruction of the inversion-channel region due to hybridization of the wave functions of electrons localized at the Na⁺ ions with the wave functions of electrons in the inversion channel.     

Analog/RF Study of Self-aligned In<Subscript>0.53</Subscript>Ga<Subscript>0.47</Subscript>As MOSFET with Scaled Gate Length

This study presents the impact of gate length scaling on analog and radio frequency (RF) performance of a self- aligned multi-gate n-type In_0.53Ga_0.47As metal oxide semiconductor field effect transistor. The device is fabricated using a self-aligned method, air-bridge technology, and 8 nm thickness of the Al₂O₃ oxide layer with different gate lengths. The transconductance-to-normalized drain current ratio (g _m/I _D) method is implemented to investigate analog parameters. Moreover, g _m and drain conductance (g _D) as key parameters in analog performance of the device are evaluated with g _m/I _D and gate length variation, where g _m and g _D are both showing enhancement due to scaling of the gate length. Early voltage (V _EA) and intrinsic voltage gain (A _V) value presents a decreasing trend by shrinking the gate length. In addition, the results of RF measurement for cut-off and maximum oscillation frequency for devices with different gate lengths are compared.     

Kinetics of ambipolar diffusion and drift currents of nonequilibrium carriers in semiconductors

A radically new method is suggested for drift mobility determination in semiconductors; this method is based on the measurement of how much time it takes to attain a peak value of the diffusion-drift current of nonequilibrium charge carriers excited by short pulses of light from a high-absorption region through the one of the contacts. The estimations show that the method is applicable if the drift flow exceeds the diffusion flow. This condition is satisfied at voltages exceeding those corresponding to the highest rate of t_max(U) decay. Mobility μ can be calculated by the formula μ=d²(2Ut_max)^?1, where d is the distance between contacts, U is the applied voltage, and t_max is the time the peak value of the photocurrent is attained.     

Effect of thallium doping on the mobility of electrons in Bi<Subscript>2</Subscript>Se<Subscript>3</Subscript> and holes in Sb<Subscript>2</Subscript>Te<Subscript>3</Subscript>

The Shubnikov–de Haas effect and the Hall effect in n-Bi_2–xTl_xSe₃ (x = 0, 0.01, 0.02, 0.04) and p-Sb_2–xTl_xTe₃ (x = 0, 0.005, 0.015, 0.05) single crystals are studied. The carrier mobilities and their changes upon Tl doping are calculated by the Fourier spectra of oscillations. It is found shown that Tl doping decreases the electron concentration in n-Bi_2–xTl_xSe₃ and increases the electron mobility. In p-Sb_2–xTl_xTe₃, both the hole concentration and mobility decrease upon Tl doping. The change in the crystal defect concentration, which leads to these effects, is discussed.