Kenichiro Tahira

Spectral analysis of deep level transient spectroscopy (SADLTS)

A spectral analysis of Deep Level Transient Spectroscopy (SADLTS) is proposed. This method analyzes the transient junction capacitance C(t)=\intbaS(λ)exp (-λt)dλ in order to determine the finite continuous emission rate spectrum S(λ). SADLTS permits one to obtain more detailed information with a single-temperature scan and spectral analysis instead of the conventional multi-temperature scan or single-exponential analysis. Even if S(λ) includes two peaks at λ1 and λ2, those peaks can be distinguished for λ2/λ1>2. As an example of the application of SADLTS, deep levels in Si:Au were experimentally investigated. According to the three-dimensional S(λ)-T2/λ-1/T representation, the apparent single peak in the conventional DLTS was found toconsist of two adjacent levels with activation energies and capture cross sections of EB1=0.51 eV, σB1=4.0×10-15 cm2 and EB2=0.47 eV, σB2=1.1×10-15 cm2, respectively.

Distribution of Deep Level Parameters in Spectral Analysis of DLTS (SADLTS)

It is shown that one can estimate the broadening parameters of both the capture cross section σ and the activation energy E of a deep level from the emission rate spectrum S(λ, T) of the transient capacitance ΔC(t, T) obtained in a spectral analysis of the DLTS(SADLTS) proposed in our previous paper. Numerical examples are given to illustrate typical cases of interest for an idealized model of independent Gaussian distributions, f(σ) and g(E). In general, S(λ, T) is a convolution of f and g with its higher and lower λ side characterized by broadening parameters for σ and E, respectively. A simple formula is derived in order to obtain the central value E of the activation energy.

Spectral analysis of deep level transient spectroscopy (SADLTS) of DX-centers in AlxGa1−xAs:Sn

Emission rate spectra S(λ, T) of DX-centers in AlxGa1-xAs:Sn for x=0.35, 0.45, 0.55, 0.65 and 0.75 were obtained from the transient capacitance waveform ΔC(t, T) by the spectral analysis proposed in our previous paper. S(λ, T)s were distributed over the range λ~10~103 between 170~200 K and showed double peak features for x=0.35, 0.45 and 0.55. S(λ, T)s broaden, shift to higher λ and decrease their peak heights at higher temperatures. Assuming Gaussian distribution for both capture cross section σ and activation energy E, the width of the activation energy ΔE is estimated to be 10~16 meV and largest at x=0.45, while the relative broadening in the capture cross section is found to be Δσ/σ0~1/3 from the shape of S(λ, T). The possible origin of the double peaks is tentatively assigned to DX-to-X conduction band and DX-to-L conduction band excitation for x=0.45 and 0.55.

Estimation of Schottky Contacts to Porous Si by Photoacoustic Spectroscopy

The photoacoustic (PA) spectra of porous Si (PS) were measured using a microphone under various surface treatments such as vibration during anodisation, vibration during rinsing, and acid surface treatment before contact deposition. The Fourier transform infrared (FTIR) spectra were also compared for the estimation of surface conditions. To obtain good contacts for electroluminescence (EL) devices, where a large injection current during operation is required, the vibration during rinsing in distilled (DI) water and hydrochloric (HCl) acid surface treatment are important. The PA signals increase with such vibration because of the enhancement of the pressure effects in pores of PS. Moreover, the decrease in PA signals caused by the surface change of pores can be also observed with the surface treatments. As the good contacts are obtained both by the vibration and surface treatments, the surface states in the pores seem to be important for the contacts to PS. This process can be monitored by PA spectroscopy in a noncontact manner because the PA signals from the PS are enhanced in the pores by the pressure effects, and PA spectroscopy can be sensitive to the change in surface of pores of PS.

Multiexponential Analysis of Deep Level Transient Spectroscopy (C2-MEDLTS)

A novel method of analyzing the deep impurity levels in semiconductors is proposed. This method analyzes the square of the transient junction capacitance waveforms as multiexponentials by using the nonlinear least squares method and is referred to as C2-MEDLTS. The effect on the emission rate τ of the deep levels in the nonionized region (λ-effect) is included. It allows correct evaluation of the activation energies and capture cross sections even for the high concentrations of closely spaced deep levels (NT/ND 1).

Photoacoustic Spectra for Porous Silicon Using Piezoelectric Transducer and Microphone.

Photoacoustic (PA) spectra are measured for porous Si (PS) using piezoelectric detectors (PPT) and compared with the microphone PA (MPA) spectra to examine the nonradiative properties of PS that has complicated nanostructures. Three peaks (peaks 1 and 2, and a small peak at 600 nm) are observed in samples 2 and 3 with porosities p=60 and 70%, respectively, while only a single peak (peak 1) is observed in sample 1 with p=30%. Peak 2 is related to the band gap of the Si substrate. In contrast, peak 1 seems to be due to the absorption change at the band gap unique to the PS. This band gap shifts to higher energy for PS samples with higher porosity. We propose that this band gap is defined by the largest size nanocrystals, where PS has nanocrystallite size distribution. The small peak at 600 nm is related to the PL peak and it is observed only in PPT spectra. These results suggest that different mechanisms are operating in PPT spectra and MPA ones for the enhancement of the PA spectra.

Double-Peak Emission Rate Spectrum of DX-Centers in AlxGa1-xAs

Emission rate spectra S(λ) of the DX-centers in AlxGa1-xAs were studied for dopants, such as Te, Sn, Si and Se, using the SADLTS (spectral analysis of deep level transient spectroscopy) method. The double peak emission rate spectra were observed for these DX-centers in AlxGa1-xAs at around x=0.5 of Al mole fraction. The band effect (crossover of L-band and \varGamma-band) is the dominant contribution to the rather broadened spectra of DX-centers appearing in the conventional DLTS.

Transient Characteristics of PME Effect –Theory–

Assuming a single life time τ for both of the minority and majority carriers, a surface recombination velocity s, and an absorption coefficient k, the spatial and time distribution of excess minority carriers injected in a semiconductor by a monochromatic light pulse is expressed in terms of error functions with space and time variables normalized to diffusion length L and τ. The result is characterized by two material parameters α=kL and γ=sτ/L. PME output proportional to the surface density of excess carriers p(0, t) is shown to decay much faster than the PC output because for large α inward diffusion of carriers accelerates the decay of p(0, t) but has no effect on the total number of carriers. Results of preliminary measurements on n-Ge samples using a Q-switched ruby laser is shown to be in reasonable agreement with these theoretical conclusions.

A high Q temperature insensitive inductive transistor circuit

Abstract The inductive transistor circuit of ‘reactance diode’ type is generalized to the form in which the base resistance is replaced by a series combination of a resistance and another reactance diode, and the frequency and temperature dependences of the equivalent impedance of the circuit are studied. The equivalent impedance shows a strong frequency dependence. The equivalent resistance passes through a minimum at a frequency nearly equal to the geometrical mean of the two cut-off frequencies of the transistors. The sign of this minimum resistance can be adjusted to be either positive or negative. The series equivalent inductance increases monotonically as a function of frequency below the higher cut-off frequency of the two transistors. The increase in both the equivalent resistance and the series inductance results from the increase in the value of the resistances which were connected in series with the second reactance diode. The temperature dependence of these quantities can be explained by assuming a reasonable temperature dependence for cut-off frequencies of the transistors. Compensation of these temperature effects is possible by giving appropriate negative temperature coefficients to the external base resistances. The magnitudes of these temperature coefficients fall well within the range realizable by the use of semiconductor resistances. On the basis of these features of this circuit a design procedure having a temperature-insensitive high Q inductance is proposed.

Vanadium-Related Deep Levels in n-Silicon Detected by Junction Capacitance Waveform Analysis

Multiexponential (ME-) and spectral analysis (SA-) Deep Level Transient Spectroscopy (DLTS) are made on vanadium-related deep levels in n-silicon. We resolved three traps with large amplitude and three more traps with comparatively small amplitude, VA (ΔE=0.08 eV, σ=3.2×10-18 cm2), VB (ΔE=0.28 eV, σ=2.6×10-14 cm2) and VC (ΔE=0.24 eV, σ=8.4×10-18 cm2). These vanadium-related levels do not have appreciable broadening of the emission rate spectrum.