Thermal noise of a silicon double-injection plasma diode

Measurements of the thermal noise of a silicon p⁺−π−n⁺ diode operating in the Lampert-insulating regime agree within 6 per cent with the prediction S_i = 4kT Re(Y). The noise measurements were performed in the cube-law regime with d.c.-currents from 100 μA to 4mA at room temperature.     

Experimental studies of switching in metal semi-insulating n-p silicon devices

The switching properties of silicon structures comprising a p⁺-n junction and a metal electrode separated from the n-section of the p⁺-n junction by a semi-insulating (leaky) layer are presented. Two basic types of structure were studied: devices with relatively light doped n-sections, and those with relatively heavily doped sections.

The switching voltage of the first group was found to be proportional to the product of the doping density, N_d and the square of the width of the n-section, and to be only very weakly temperature-dependent. The capacitance-voltage relationship of the device in the high-impedance mode was found to be of the form C⁻¹ ∞ V^1/2, and these measurements established that switching occurred just as the depletion region of the n-section under the gate electrode reached through to the p⁺-n junction. It was thus established that these devices were operating in the punch-through mode.

In the second group of devices, the doping density of the n-section was increased by diffusing an n-well into the section. The switching properties were found to be quite different from the punch-through devices. The switching voltage was found to be independent of the width of the n-section and proportional to N_d^−3/4. Capacitance measurements also showed that the depletion region in the n-section under the oxide at switching, varied with the doping concentration, and was substantially less than the width of the n-layer. It was thus concluded that switching in these devices was of the avalanche-mode type.     

Simulation and analysis of amorphous silicon image sensor having a p−i−n structure

The current-voltage characteristics of a p−i−n a-Si : H contact image sensor under dark and illuminated conditions have been simulated by solving the Poisson's equation and the continuity equations, and the results are correlated with the experiments. The dependence of the dark and photo-currents on the parameters such as the density of states in the gap, intrinsic layer width, dopant concentrations of p⁺ layer and n⁺ layer are discussed.     

Process and device characterization of high voltage gallium arsenide P-i-N layers grown by an improved liquid phase epitaxy method

GaAs P-i-N layers with an i-region net doping of less than 10¹² cm⁻³ were grown on P⁺ and N⁺ substrates by a modified liquid phase epitaxy (LPE) method. Doping profiles and structural data obtained by varius characterization techniques are presented and discussed. A P⁺-P-i-N-N⁺ diode with a 25 μm-wide i-region exhibits a breakdown voltage of 1000 V, a t_rr of 50 ns, and reverse current densities (at V_R = 800 V) of − 3 × 10⁻⁶ A/cm² at 25°C and 10⁻² A/cm² at 260° C.     

Modeling multi-band effects of hot electron transport in silicon by self-consistent solution of the Boltzmann transport and Poisson equations

This paper presents a self-consistent numerical technique for the solution of the multi-band Boltzmann transport equation (BTE) and the Poisson equation in silicon. The effects of high energy bands ( 3 eV) are modeled in the formulation. The numerical technique utilizes a new curvilinear boundary-fitted coordinate grid which is tailored for self-consistent calculations. A new Scharfetter-Gummel like discretization of the BTE is presented. The numerical algorithm is tested on a n⁺ − n − n⁺ device structure.     

Unreliable sensor grids: coverage, connectivity and diameter

We consider an unreliable wireless sensor grid network with n nodes placed in a square of unit area. We are interested in the coverage of the region and the connectivity of the network. We first show that the necessary and sufficient conditions for the random grid network to cover the unit square region as well as ensure that the active nodes are connected are of the form p(n)r²(n)  log(n)/n, where r(n) is the transmission radius of each node and p(n) is the probability that a node is “active” (not failed). This result indicates that, when n is large, even if each node is highly unreliable and the transmission power is small, we can still maintain connectivity with coverage.

We also show that the diameter of the random grid (i.e., the maximum number of hops required to travel from any active node to another) is of the order . Finally, we derive a sufficient condition for connectivity of the active nodes (without necessarily having coverage). If the node success probability p(n) is small enough, we show that connectivity does not imply coverage.     

The physics and fabrication of in situ back-gated (311)A hole gas heterojunctions

This paper reports the fabrication of an in situ back-gated hole gas on the (311)A surface of GaAs. The hole density can be varied from fully depleted to p_s = 2.1 × 10¹¹ cm⁻² with mobilities of up to μ = 1.1 × 10⁶ cm²V⁻¹ s⁻¹. It is seen that for carrier densities down to p_s = 4 × 10¹⁰ cm⁻² the mobility in the [ ] direction is greater than that in the [ ] direction. Using a combination of front- and back-gates we are able to keep the carrier density constant and deform the hole gas wavefunction such that the holes are pushed up against or moved further away from the heterointerface. Thus we are able to separately investigate the various scattering mechanisms that determine the mobility, and compare the experimental data with theoretical calculations based on the shape of the wavefunction.     

Nitrogen vacancy scattering in GaN grown by metal–organic vapor phase epitaxy

Electron mobility limited by nitrogen vacancy scattering was taken into account to evaluate the quality of n-type GaN grown by metal–organic vapor phase epitaxy. Two assumptions were made for this potential for the nitrogen vacancy (1) it acts in a short range, and (2) does not diverge at the vacancy core. According to the above assumptions, a general expression to describe the scattering potential U(r)=−U₀exp[−(r/β)ⁿ], (n=1,2,…,∞) was constructed, where β is the potential well width. The mobilities for n=1,2, and ∞ were calculated based on this equation, corresponding to the simple exponential, Gaussian and square well scattering potentials, respectively. In the limiting case of kβ1 (where k is the wave vector), all of the mobilities calculated for n=1,2, and ∞ showed a same result but different prefactor. Such difference was discussed in terms of the potential tail and was found that all of the calculated mobilities have T^−1/2 temperature and β⁻⁶ well width dependences. A mobility taking account of a spatially complicate scattering potential was studied and the same temperature dependence was also found. A best fit between the calculated results and experimental data was obtained by taking account of the nitrogen vacancy scattering.     

Analysis of I–V measurements on PtSi-Si Schottky structures in a wide temperature range

I–V Measurements on PtSi-Si Schottky structures in a wide temperature range from 90 to 350 K were carried out. The contributions of thermionic-emission current and various other current-transport mechanisms were assumed when evaluating the Schottky barrier height Φ₀. Thus the generation-recombination, tunneling and leak currents caused by inhomogeneities and defects at the metal-semiconductor interface were taken into account.

Taking the above-mentioned mechanisms and their temperature dependence into consideration in the Schottky diode model, an outstanding agreement between theory and experiment was achieved in a wide temperature range.

Excluding the secondary current-transport mechanisms from the total current, a more exact value of the thermionic-emission saturation current I_te and thus a more accurate value ofΦ_b was reached.

The barrier height Φ_b and the modified Richardson constant A^** were calculated from the plot of thermionic-emission saturation current I_te as a function of temperature too. The proposed method of finding Φ_b is independent of the exact values of the metal-semiconductor contact area A and of the modified Richardson constant A^**. This fact can be used for determination of Φ_b in new Schottky structures based on multicomponent semiconductor materials.

Using the experimentally evaluated value A^** = 1.796 × 10⁶ Am⁻²K⁻² for the barrier height determination from I–V characteristics the value of Φ_b = 0.881 ± 0.002 eV was reached independent of temperature.

The more exact value of barrier height Φ_b is a relevant input parameter for Schottky diode computer-aided modeling and simulation, which provided a closer correlation between the experimental and theoretical characteristics.     

Theory of open circuit photo-voltage in degenerate abrupt p-n junctions

In this paper we have presented a comprehensive theory of photovoltage in degenerate abrupt p-n junction. In contrast to the earlier work[9], we use a generalized diffusion-mobility relation for charge carriers which is valid for extreme degeneracy. The general expression for the photo e.m.f. is used to obtain explicit expression for the case of lowinjection levels for two types of p−n junction. (i) When both the neutral regions are extremely degenerate and (ii) when one region is extremely degenerate but other is non-degenerate. Some numerical results of photo e.m.f. for typical p⁺n and n⁺p junctions are presented.     

Carrier recombination in single crystal PbSe

The photoconductivity decay curves after illumination of single crystal n- and p-type PbSe were analysed assuming recombination through different localized impurity levels in conjunction with direct recombination. The lifetimes deduced for direct (Auger and radiative) recombination below 250 K were in agreement with the calculated values for carrier concentrations 2·10¹⁷ cm⁻³. Furthermore, the existence of up to three impurity levels was concluded from the longer lifetime-components present in the decay curves. Appropriate approximations of the general recombination theory yielded energies separated between 20 and 50 meV from the nearer band edge and minority carrier cross sections 10⁻¹⁷−4·10⁻¹⁹ cm² in the temperature range 250-100 K, and majority carrier cross sections 10⁻¹⁹−10⁻²⁰ cm² at T < 100 K for these levels.     

Lateral tunneling transistors fabricated by plane-dependent Si-doping in nonplanar epitaxy on GaAs (311)A and (411)A substrates

We have demonstrated the lateral tunneling transistors on GaAs (311)A and (411)A patterned substrates by using the plane-dependent Si-doping technique. Lateral p⁺-n⁺ tunneling junctions are formed by growing heavily Si-doped layers on patterned substrates. Current—voltage curves for both transistors show gate-controlled negative differential resistance characteristics. Furthermore, the peak current density of the lateral tunneling diodes fabricated on the (311)A patterned substrates increases as buffer layer thickness is increased, and a typical peak current density of 58 A/cm² for p = 6 × 10¹⁹ cm⁻³ and n = 7 × 10¹⁸ cm⁻³ is obtained when the buffer layer thickness is 1.2 μm. This study shows that plane-dependent Si-doping in non-planar epitaxy is a promising technique for fabricating tunneling transistors.     

Optically detected impurity D and D transitions in GaAs quantum wells

We have carried out a far-infrared magnet-optical study on shallow donor states confined in GaAs quantum wells (QWs), applying a recently developed optical detection technique. We have observed, in addition to cyclotron resonance, the 1s → 2p₊ transition of neutral donors (D⁰), and singlet and triplet transitions of negative donor ions (D⁻); the latter observation verifies the existence of D⁻ ions in well-only doped QWs under optical pumping. This is the first observation of optically detected impurity resonances in confined systems and demonstrates the power and utility of this technique for such studies.     

Rise time of silicon p-n-p photodiodes

Exact analytical expressions are derived for the short circuit photodiode currents excited by light pulses, under the assumption that the drift carrier velocity linearly depends on the electric field in the depletion layer. Reflection from the back surface of the photodiode is taken into account. Using the obtained expressions it is possible to establish a connection between the rise time t_rise and the product W_eff of the absorption coefficient (λ) and effective depletion layer width W_eff(W) at various ratios of the diode thickness and the effective depletion layer width. The influence of the RC-constant (where C is the photodiode effective capacity and R is the sum of the diode series and loading resistances) on the rise time is also analyzed.

One of the most important conclusions is that generally the rise time is larger for p-n-n⁺ photodiode configurations than for n-p-p⁺ configurations at the same substrate resistivity.     

Hole and electron mobilities in heavily doped silicon: comparison of theory and experiment

Most device models for npn or pnp transistors assume that hole (electron) mobilities in n-type and p-type silicon are equal. Partial-wave phase shift calculations for the contributions of carrier-dopant ion scattering to the carrier mobilities lead to unequal minority hole (electron) and majority hole (electron) mobilities at the same doping density. These calculations are valid over the doping range of 2 x 10¹⁹ to 8 x 10¹⁹ cm⁻³ in n-type and p-type silicon and contain the assumptions that the holes and electrons move in isotropic parabolic energy bands and are scattered by the screened Coulomb potentials of the dopant ions. When the effects of carrier-acoustic phonon and carrier-carrier scatterings are included, these calculations agree to within the spread of experimental value for the majority mobilities reported in the literature. This agreement is a substantial improvement by factors of 2–4 over the results of earlier theories such as first order Born and nondegenerate theories. The results of this work, particularly the inequality of minority and majority carrier mobilities, have implications for the modeling of both bipolar and field effect transistors.     

Optimal fault margin in a computer system

This paper suggests a stochastic model how to determine a fault margin in a computer system, who fails when the total number of hidden faults exceeds a threshold level N of tolerance: A fault occurs at a non-homogeneous Poisson process, and (i) becomes system failure with probability p₁, (ii) becomes hidden fault with probability p₂ and is accumulated, or (iii) is removed with probability p₃. The expected cost rate to system failure is derived, and an optimal number N^* to minimize it is discussed. A numerical example is finally given.     

Breakdown voltage of diffused epitaxial junctions

The quantitative aspects of the breakdown-voltage calculations in reach-through-limited p⁺−n−n⁺ junctions are revisited, using numerical simulation. It is shown that the conventional abrupt-junction approximation may underestimate the breakdown voltage of diffused epitaxial junctions by as much as 60%, depending on junction depth. Oppositely, but less erroneously, a combined abrupt/linearly-graded approximation overestimates the breakdown voltage by at most 15%. A set of numerically calculated plots are provided for the design of low-voltage power devices.     

Frequency dependence of of MOS capacitors

The frequency dependence of ΔV/Δ(C⁻²) of an MOS capacitor, which plays an important role in determining the semiconductor doping profile, is studied theoretically and experimentally. Useful expressions relating the measurable quantities to the doping profile are derived systematically. It is shown how interface states and majority carriers influence the frequency dependence of ΔV/Δ(C⁻²) and give rise to errors in profile determinations. The techniques of measuring the various types of the frequency dependence of ΔV/Δ(C⁻²) are also described.     

Quantized conductance in a Si/Si0.7Ge0.3 split-gate device and impurity-related magnetotransport phenomena

We have defined a quantum point-contact by the split-gate technique in a Si/SiGe heterostructure containing a two-dimensional electron gas with an elastic mean free path of about 1.3 μm. The conductance of this device shows typical steps very close to multiples of 4e² h⁻¹. Upon application of a perpendicular magnetic field the spin and valley degeneracies are lifted and magnetic depopulation of the one-dimensional subbands can be observed. The appearance of Aharonov-Bohm oscillations for B ≥ 2T and of resonant tunneling peaks close to “pinch-off” indicates the presence of impurities close to the constriction.     

Theoretical calculation and experimental evidence of the real and apparent bandgap narrowing due to heavy doping in p-type Si and strained Si1−xGex layers

Based on an analytical approach developed by Jain and Roulston, the different contributions to the bandgap narrowing at T = 0 K due to heavy doping in highly p-type doped Si and strained Si_1-xGe_x-layers are calculated for x < 0.3. The valence band in Si and in strained Si_1-xGe_x layers is not parabolic, it is highly distorted. To take the non-parabolicity into account a dopant concentration-dependent density of states effective mass is defined. within the framework of this formalism we find that the bandgap narrowing (BGN) in Si is not appreciably affected due to the band distortion. The situation for strained Si_1-xGe_x layers is quite different, in that the BGN increases significantly at doping levels exceeding 10¹⁹ cm⁻³. In the earlier published work, BGN of the Si_1−xGe_x layers was either slightly smaller or about the same as in Si. Now at high doping levels, BGN becomes considerably higher than in Si. We will show that the effect of the strain on the Fermi energy is much larger than on the BGN, which will cause a large change in the effective valence band offset. Comparison will be made then between our theoretical calculations and experimental results obtained on two different device structures. The modified effective valence band offset that we have calculated is in very good agreement with the experimental value derived from the published work on long-wavelength optical detectors. The apparent bandgap narrowing in strained p-type Si_1-xGe_x-layers is also calculated and compared with the experimental results on Heterojunction Bipolar Transistors fabricated in our laboratory.