Thermal Generation-recombination Research Articles

Model for elimination of lifetime-limiting carbon vacancy defects in SiC by thermal treatment

In 4H-SiC semiconductors, carbon vacancies act as traps, which limit the carrier lifetime. During high-temperature treatment of 4H-SiC, the concentration of carbon vacancies can be increased or decreased by several atomistic processes, including the diffusion of carbon vacancies and carbon self-interstitials, and the thermal generation–recombination of defects. In this work, an analytic process model has been developed and calibrated against a collection of measured data. The model describes the concentration of carbon vacancies after thermal processing for a wider range of process conditions than previous works. For inert annealings, bulk recombination, bulk generation, and diffusion of carbon vacancies and carbon interstitials play a critical role. For oxidation processes, carbon interstitials are injected at the oxidizing surface. The injection rate of carbon interstitials at the oxidizing surface and their diffusivity from the surface into the bulk govern the reduction of carbon vacancies via bulk recombination. Basic properties of carbon vacancies and carbon self-interstitials in 4H-SiC, such as the thermal equilibrium concentrations, diffusivities, and bulk recombination rates, are reflected by model parameters and have been determined by model calibration for the temperature range of 1150–1950 °C. High-quality epitaxial films and low-quality substrates are described consistently, when assuming that carbon interstitials can be trapped by defects present only in the substrate.

Perspective on III–V barrier detectors

In a photodiode made from a narrow bandgap III–V material such as InSb, the dark current is usually dominated by thermal generation-recombination (GR) in the depletion region. In an XBn or XBp barrier detector, the GR current is suppressed by confining the depletion region to a wide bandgap barrier material with a band alignment that blocks majority carriers. Diffusion limited barrier detectors are essentially unipolar and represent a device architecture with unity gain that is fundamentally different from that of the traditional photodiode. High performance barrier detector arrays spanning the mid- and long-wave infrared atmospheric transparency windows are currently being produced with both bulk alloy and type II superlattice (T2SL) absorbers several micrometers thick. In T2SLs, 5–10 μm diffusion lengths have been demonstrated for both InAs/GaSb XBp and InAs/InAsSb XBn devices. The former exhibit minority electrons with a short lifetime and a high mobility, while the latter exhibit minority holes with a long lifetime and a low mobility. The contrasting behavior is understood in terms of competing GR and Auger recombination mechanisms, and a transition between metallic and nonmetallic conduction. These properties present unique challenges for the future design of monolithic dual band photodetectors.

Low-Temperature Noise Performance of SuperSpec and Other Developments on the Path to Deployment

SuperSpec is a compact on-chip spectrometer operating at mm and sub-mm wavelengths which will enable the construction of sensitive multibeam spectrometers. SuperSpec employs a filter bank architecture, consisting of lithographically patterned niobium superconducting microstrip mm-wave resonators. The power admitted by each resonator is detected by a titanium nitride lumped-element kinetic inductance detector (KID) with resonant frequency from 100 to 200 MHz. We present a characterization of the detector noise performance down to 10 mK measured in a dark setting. We report a device NEP of $$2.7 \times 10^{-18}\, \hbox {W Hz}^{-1/2}$$ at 210 mK, which is below the expected photon noise level at high-altitude ground-based observatories. The NEP decreases to a constant value of approximately $$7.0 \times 10^{-19}\, \hbox {W Hz}^{-1/2}$$ below 130 mK. The white noise is well modeled by thermal generation–recombination noise (GR noise) down to 130 mK and a noise floor at low temperatures. Moreover, the addition of low-pass coaxial filters further reduces the noise floor to achieve an NEP of $$5.7 \times 10^{-19} \,\hbox {W Hz}^{-1/2}$$ below 100 mK. We discuss a photolithographic technique to adjust KID resonances that results in an $$f_{0}$$ designed versus measured scatter of $$1.7 \times 10^{-5}$$ , which will allow a significant reduction in resonators lost to clashes in full-scale designs. Finally, we present a demonstration of a new ROACH-2-based readout system operating below 500 MHz and show preliminary data indicating the suitability of this system for future highly multiplexed KID arrays.

Dependence of dark current on carrier lifetime for InGaAs/InP avalanche photodiodes

Effects of carrier lifetime of all layers on the dark current have been studied for the separate-absorption-grading-charge-multiplication InGaAs/InP avalanche photodiodes (APDs). The results indicate that the remarkably increasing of the dark current at the punch-through voltage strongly depends on the carrier lifetime of InGaAs absorption layer. According to the simulation results, we can estimate the carrier lifetime of InGaAs absorption layer and InP multiplication layer to be about 100 ns and 20 ps for our fabricated device. And we can see that the dark current of APDs near the breakdown voltage is mainly dominated by the thermal generation–recombination current from the InGaAs absorption layer and trap-assisted-tunneling current from the InP multiplication layer.

Effects of Electrostatic Discharge Stress on Current-Voltage and Reverse Recovery Time of Fast Power Diode

Fast recovery diodes (FRDs) were developed using the <TEX>$p^{{+}{+}}/n^-/n^{{+}{+}}$</TEX> epitaxial layers grown by low temperature epitaxy technology. We investigated the effect of electrostatic discharge (ESD) stresses on their electrical and switching properties using current-voltage (I-V) and reverse recovery time analyses. The FRDs presented a high breakdown voltage, >450 V, and a low reverse leakage current, < <TEX>$10^{-9}$</TEX> A. From the temperature dependence of thermal activation energy, the reverse leakage current was dominated by thermal generation-recombination and diffusion, respectively, at low and high temperature regions. By virtue of the abrupt junction and the Pt drive-in for the controlling of carrier lifetime, the soft reverse recovery behavior could be obtained along with a well-controlled reverse recovery time of 21.12 ns. The FRDs exhibited excellent ESD robustness with negligible degradations in the I-V and the reverse recovery characteristics up to <TEX>${\pm}5.5$</TEX> kV of HBM and <TEX>${\pm}3.5$</TEX> kV of IEC61000-4-2 shocks. Likewise, transmission line pulse (TLP) analysis reveals that the FRDs can handle the maximum peak pulse current, <TEX>$I_{pp,max}$</TEX>, up to 30 A in the forward mode and down to - 24 A in the reverse mode. The robust ESD property can improve the long term reliability of various power applications such as automobile and switching mode power supply.

Numerical Analysis of Current–Voltage Characteristics of LWIR nBn and p-on-n HgCdTe Photodetectors

We performed numerical analysis of the current–voltage characteristics of long-wavelength infrared unipolar HgCdTe nBn photodetectors and compared those results with those of conventional p-on-n HgCdTe photodiodes. A computer program was applied to explain in detail the impact of the charge carrier generation and recombination processes on current densities. In our model the carrier diffusion, thermal generation–recombination, band-to-band tunneling, trap-assisted tunneling (via states located at mercury vacancies as well as dislocation cores), and impact ionization are included as potential limiting mechanisms. To validate the model, we compared the theoretical predictions with experimental data of high-quality p-on-n photodiodes published in the literature.

Two dimensional simulation and modeling of the electrical behavior in nanocrystalline silicon thin-film transistors

Nanocrystalline silicon thin-film transistors present technological interest in that they combine many of the advantages of amorphous with those of polycrystalline Si structures. Progress in practical implementation of this technology is hampered by limited understanding of the conduction mechanisms in these structures and of the underlying relationship between device behavior and process manufacturing parameters. These mechanisms are explored through detailed simulation that includes model calibration and correlation with experimental results, as well as parametric sensitivity evaluation of this class of devices over the entire range of applied voltage. Through fitting of the tests results, a unique set of density of states was identified that characterizes the particular technology used. The leakage current was attributed to the band to band tunneling and thermal generation-recombination mechanisms. For devices with channel length of less than 20μm, the kink effect was observed in the output characteristics for high drain voltages and the impact ionization coefficient was determined.

Revised Shockley–Read–Hall lifetimes for quantum transport modeling

The inclusion of quantization effects on the carrier densities is now the state of the art in modern semiconductor device simulators and yields, for example, quantum-corrected threshold voltages and quantum-mechanical models of the channel mobility. However, the effect of charge quantization on nonradiative thermal generation–recombination has not received much attention. In this article, Shockley–Read–Hall recombination is examined for situations in which electrons and/or holes are confined in semiconductor devices. For the transitions between band states and a single deep level, a previously developed multiphonon description is adopted. It is found that the lifetimes have to be altered due to the same quantized local density of states that also accounts for the carrier distribution. Numerical evaluation of this model for one-dimensional potentials and small phonon energies results in spatially varying lifetime profiles that exhibit two opposite regimes. The additional nonclassical offset of the subband eigenenergies causes an increased lifetime in the limit of strong quantum confinement. For nondegenerate statistics, an analytical high-temperature approximation is presented for this limit, where the activation energy of the lifetime is increased by the lowest-subband offset. In the absence of confinement, however, high electric fields reduce the lifetime due to carrier tunneling into the bandgap.

A novel multi-heterojunction HgCdTe long-wavelength infrared photovoltaic detector for operation under reduced cooling conditions

A new multi-heterojunction photovoltaic infrared photodetector is presented. The device has been developed specifically for operation at elevated temperatures for detection of infrared radiation in the mid- and long-wavelength range of the infrared spectrum. The new structure solves the perennial problems of poor quantum efficiency and low dynamic resistance found in conventional long-wavelength infrared photovoltaic detectors when operated near room temperature. Analysis indicates that practical devices with properly optimized multiple heterojunction layers are capable of achieving the performance limits imposed by the statistical nature of thermal generation-recombination processes. In order to demonstrate the technology, multiple heterojunction devices have been fabricated on epilayers grown by isothermal vapour phase epitaxy of HgCdTe and in situ As p-type doping. The detector structures were formed using a combination of conventional dry etching, angled ion milling and angled thermal evaporation for contact metal deposition. These multi-junction HgCdTe heterostructure devices exhibit performances comparable to, or better than, existing long-wavelength photoconductors and photoelectromagnetic detectors operated under the same conditions. A typical of optically immersed multiple heterostructure photovoltaic detectors of approximately can be achieved at a signal wavelength of and for operation at 230 K.

Multi-heterojunction large area HgCdTe long wavelength infrared photovoltaic detector for operation at near room temperatures

This paper describes a new multi-heterojunction n+pp photovoltaic infrared photodetector. The device has been developed specifically for operation at temperatures of 200–300K in the long wavelength (8–14 µm) range of the infrared spectrum. The new structure solves the perennial problems of poor quantum efficiency and low dynamic resistance found in conventional long wavelength infrared photovoltaic detectors when operated near room temperature. Computer simulations show that devices with properly optimized multiple heterojunctions are capable of achieving the performance limits imposed by the statistical nature of thermal generation-recombination processes. In order to demonstrate the technology, multiple heterojunction devices have been fabricated on epilayers grown by isothermal vapor phase epitaxy of HgCdTe and in situ As p-type doping. The detector structures were formed using a combination of conventional dry etching, angled ion milling, and angled thermal evaporation for contact metal deposition. These multi-junction n+pp HgCdTe heterostructure devices exhibit performances which make them useful for many applications. D* of optically immersed multiple heterostructure photovoltaic detectors exceeding 108cmHz1/2/W were measured at λ=10.6 µm and T=300K.

Off-current in polycrystalline silicon thin film transistors: An analysis of the thermally generated component

The thermal generation component of polycrystalline silicon TFTs off-current is analysed experimentally and theoretically. In order to minimize the field-enhanced component of the leakage current, hot-hole injection, obtained by stressing the device at negative gate voltage and high source-drain voltage, has been used to reduce the electric field at the drain junction. After stress, the electrical characteristics in the off-regime are channel length independent and do not depend on gate voltage. This behaviour has been associated with the thermal generation-recombination processes occurring at the drain junction. Two-dimensional numerical simulations have been carried out with the program HFIELD, which has been modified to take into account the presence of gap states in polysilicon, and to incorporate the thermal generation-recombination processes by using the Shockley-Read-Hall statistics. Numerical simulations confirm that the generation occurs in the depletion region of the drain junction. The experimental I d- V ds characteristics measured at negative gate voltage have been compared with the calculated characteristics. The best fit with the experimental data was obtained only by using a rather short carrier lifetime (10 −12 s). The simulations indicate that a decrease of the density of states produces a lower off-current owing to a longer carrier lifetime and to a reduction of the drain junction depletion layer.

Investigation of the generation–recombination currents in HgCdTe midwavelength infrared photodiodes

The I–V characteristics of HgCdTe midwavelength infrared (MWIR) double-layer heterojunction photodiodes at temperatures below 120 K are dominated by thermal generation–recombination in the depletion region. The diode R0A’s, determined from the slope of the I–V curves, are in excess of 1 × 108 Ω cm2 at 77 K when measured in the dark but are much lower when measured with ambient backgrounds. In order to better understand this phenomenon, the I–V characteristics of numerous MWIR HgCdTe photodiodes were measured as functions of background flux using a blackbody source and several narrowband spectral filters. Based on the measured R0A at 100 K in the dark, where thermal G–R currents dominate, the G–R lifetimes are approximately equal to the Auger lifetime for the minority carriers in the lightly doped n-type base layer. At backgrounds in excess of 1 × 1012 photons/cm2 s, the R0A decreases inversely with background flux. The magnitude of the effect at wavelengths much shorter than the cutoff agrees well with estimates based on the depletion modulation theory. At wavelengths very close to the cutoff, the effect is largest and agrees well with a theory based on optical generation of carriers in the depletion region. Thus, the background dependence of R0A is due to the combined effects of depletion modulation and optical generation in the depletion region.

Numerical simulation of hot-electron phenomena

An accurate two-dimensional numerical model for MOS transistors incorporating avalanche processes is presented. The Laplace and Poisson equations for the electrostatic potential in the gate oxide and bulk and the current-continuity equations for the electron and hole densities are solved using finite-difference techniques. The current-continuity equations incorporate terms modeling avalanche generation, bulk and surface Shockley--Read--Hall thermal generation-recombination, and Auger recombination processes. The simulation is performed to a depth in the substrate sufficient to include the depletion region, and the remaining substrate is modeled as a parasitic resistance. The increase in the substrate potential caused by the substrate current flowing through the substrate resistance is also included. The hot-electron distribution function is modeled using Baraff's maximum anisotropy distribution function. The model is used to study hot-electron phenomena including negative-resistance avalanche breakdown in short-channel MOSFET's and electron injection into the gate oxide. The model accurately predicts the positive-resistance branch of the drain current-voltage characteristic and could, in principle, predict the negative-resistance branch and the sustain voltage. The gate injection current is computed by summing the flux of electrons scattered into the gate oxide by each mesh volume element. The number of electrons in each element whose component of momentum normal to the oxide is sufficient to surmount the oxide potential barrier is approximated using Baraff's distribution function, and scattering along the electron trajectories is modeled using an appropriate mean free path. The flux scattered into the oxide can be expressed as an iterated six-fold integral which is evaluated using the potential and electron current density distributions produced by the model. Nakagome et al. [1] recently observed two new types of gate injection phenomena: avalanche injection and secondary ionization induced injection. The former is caused by carriers heated in the drain avalanche plasma, and the latter is caused by electrons generated by secondary impact ionization in the depletion region. The model yields gate current curves qualitatively similar to the experimental results.

Numerical Simulation of Hot-Electron Phenomena

An accurate two-dimensional numerical model for MOS transistors incorporating avalanche processes is presented. The Laplace and Poisson equations for the electrostatic potential in the gate oxide and bulk and the current-continuity equations for the electron and hole densities are solved using finite-difference techniques. The current-continuity equations incorporate terms modeling avalanche generation, bulk and surface Shockley–Read–Hall thermal generation-recombination, and Auger recombination processes. The simulation is performed to a depth in the substrate sufficient to include the depletion region, and the remaining substrate is modeled as a parasitic resistance. The increase in the substrate potential caused by the substrate current flowing through the substrate resistance is also included. The hot-electron distribution function is modeled using Baraff’s maximum anisotropy distribution function. The model is used to study hot-electron phenomena including negative-resistance avalanche breakdown in short-channel MOSFET’s and electron injection into the gate oxide. The model accurately predicts the positive-resistance branch of the drain current–voltage characteristic and could, in principle, predict the negative-resistance branch and the sustain voltage. The gate injection current is computed by summing the flux of electrons scattered into the gate oxide by each mesh volume element. The number of electrons in each element whose component of momentum normal to the oxide is sufficient to surmount the oxide potential barrier is approximated using Baraff’s distribution function, and scattering along the electron trajectories is modeled using an appropriate mean free path. The flux scattered into the oxide can be expressed as an iterated six-fold integral which is evaluated using the potential and electron current density distributions produced by the model. Nakagome et al. [1] recently observed two new types of gate injection phenomena: avalanche injection and secondary ionization induced injection. The former is caused by carriers heated in the drain avalanche plasma, and the latter is caused by electrons generated by secondary impact ionization in the depletion region. The model yields gate current curves qualitatively similar to the experimental results.