Masato Toita

RTS in submicron MOSFETS and quantum dots

In the present paper, emphasis is laid on those RTS showing a capture process, which deviates from the standard Shockley-Read-Hall kinetics. A modified two-step approach is proposed. In this case the charge carrier quantum transitions represent a primary process X(t), which involves two or three quantum states. The measurable quantity is the current modulation, which has discrete states, too. The current modulation is then represented by a secondary process Y(t). The proposed model can explain some of the complex switching phenomena being measured in nanoscale devices. The quadratic dependence of the capture rate on the current and the noise spectral density dependence on the current and temperature are analysed. It is shown that the occupation time probability density for emission is given by a superposition of two exponential dependencies, whereas the capture time constant distribution is purely exponential.

Noise in Submicron Metal–Oxide–Semiconductor Field Effect Transistors: Lateral Electron Density Distribution and Active Trap Position

Experiments were carried out for the n-channel devices, processed in a 0.3 µm spacer less complementary metal–oxide–semiconductor technology. Random-telegraph-signal measurements were performed for the constant gate voltage. It is supposed that electron concentration in the channel decreases from the source to the drain contact. Lateral component of the electric field is inhomogeneous in the channel and it has a minimum value near the source and reaching the maximum value near the drain electrode. Drain current is given by two components – diffusion and drift ones. Diffusion current component is independent on the x-coordinate and it is equal to the drift current component for the low electric field. The model explaining the experimentally observed capture time constant dependence on the lateral electric field and the trap position is given. From the dependence of the capture time constant τc on the drain current could be calculated longitudinal coordinate of the trap position.

Amplitude of RTS noise in MOSFETs

Low frequency noise of NMOS and PMOS field effect transistors was measured in wide temperature range as a function of applied electric field intensity in longitudinal and perpendicular direction and the influence of sample geometry on 1/f noise and RTS noise was examined for various gate lengths. Relative amplitude of RTS noise given by number of carriers under the gate and its dependence on channel and gate bias was analyzed.

GRT model of RTS noise in MOSFETs

Random Telegraph Signal (RTS) noise in submicron MOSFETs showing a capture process, which deviates from the standard Shockley-Read-Hall kinetics, is analyzed using generation-recombination-tunneling model of current modulation in order to explain quadratic dependence of capture rate on current. Proposed model of two-step charge carrier quantum transitions involving secondary trap at the channel and gate oxide interface better represents observed complex switching phenomena in nanoscale devices, as is confirmed by presented experimental results.

RTS noise amplitude and electron concentration in MOSFETs

Random Telegraph Signal (RTS) noise was measured in submicron MOSFETs under various biasing conditions from sub threshold to inversion region and dependence of amplitude and mean time of capture and emission on electric field and electron concentration analyzed. Numerical model of charge carrier transport in the channel was used to estimate trap position between the source and drain electrodes from experimental characteristics.

RTS in Submicron MOSFETs: Lateral Field Effect and Active Trap Position

Experiments were carried out for n‐channel devices, processed in a 0.3 μm spacerless CMOS technology. The investigated devices have a gate oxide thickness of 6 nm and the effective interface area is AG = 1.5 μm2. The RTS measurements were performed for constant gate voltage, where the drain current was changed by varying the drain voltage. The capture time constant increases with increasing drain current. The model explaining the experimentally observed capture time constant dependence on the lateral electric field and the trap position is given. From the dependence of the capture time constant τc on the drain current we can calculate x‐coordinate of the trap position. Electron concentration in the channel decreases linearly from the source to the drain contact. Diffusion current component is independent on the x‐coordinate and it is equal to the drift current component for the low electric field. Lateral component of the electric field intensity is inhomogeneous in the channel and it has a minimum value ...

Zero Cross Analysis of RTS Noise

RTS noise of Si MOSFETs and GaN HFET was analysed by means of zero cross method. Noise spectral density of crossing events in 1 ms to 100s windows was flat over 10−5 to 103 frequency range without apparent 1/f noise. The fluctuation of crossing rate was same in every sample, although other noise components next to RTS noise are quite different in Si and GaN devices. Correlation analysis of pulse length in doublets and triplets of successive pulses didn’t reveal any mutual dependence.

RTS and 1/f Noise in Submicron MOSFETs

The capture and emission time constants dependence on drain current for constant gate voltage and variable drain voltage show that probability for charge carrier capture decreases with increasing lateral electric field while emission process is independent on lateral field intensity. We have performed analyses of SiO2 gate insulating layer from VA characteristics measured in wide temperature range (gate electrode area 1.5 μm2, insulating layer thickness 6 nm. Leakage current in reverse mode of n‐MOS sample (for gate electrode negative) is for applied voltage lower than 1 V exponential function of applied voltage with saturation current I0 = (1–3)×10−16 A and ideality factor near to 1. Saturation current value corresponds to Schottky barrier high about 1.2 eV. We suppose that in SiO2 gate insulating layer and on the interface Si‐SiO2 there are oxygen vacancies and interstitials. High density of overlapping energy localized states creates in SiO2 impurity conduction band about 1.2 eV above the Si conduction...

1/f Noise Behaviour in High Mobility PMOSFETs with Optimized In‐plane Channel Direction

PMOS transistors with a strained SiGe channel have higher hole mobility compared with conventional pure‐silicon counterparts. In this paper, we report on the hole mobility and low‐frequency noise characteristics of the SiGe channel PMOS FETs. Epitaxial layers consist of 15nm bottom‐Si, lOnm Si0.7Ge0.3, 9nm cap‐Si are deposited selectively on the active region. Hole mobility in the SiGe channel PMOS FETs depends on the in‐plane channel direction. The maximum hole mobility was obtained for 〈100〉 direction device and the minimum was obtained for 〈110〉 direction device. Low frequency noise behaviour for SiGe channel PMOS FET was investigated and compared with pure Si PMOS FETs. The lowest input‐referred noise power (Svg) was obtained for the 〈100〉 direction SiGe PMOS FET. The highest trans‐conductance in the 〈100〉 channel results in lowest Svg. Hooge’s alpha (αH) for the SiGe channel PMOS FET was approximately 1/4 of the αH for the Si channel PMOS FET.

RTS in Submicron MOSFETs: High Field Effects

The downscaling of electronic devices makes high field transport effects more important. In deep submicron technology high transversal and high lateral electric field exists. Application of drain voltage 1V results in electric field, which exceeds the silicon critical electric field. Electron temperature is then higher than lattice one and field dependent electron mobility must be considered. Due to small gate area we were able to activate one trap only and then in time domain two levels signal was observed. A systematic analysis of two level RTS signal was made to obtain information on capture and emission processes as a function of gate voltage, drain current and temperature for low and high lateral electric field. With increasing drain voltage capture time increases, while dependence on gate voltage is almost the same as for low drain voltage. For constant gate voltage and variable drain voltage emission process is independent on lateral field intensity, while capture time increases with lateral field ...