Effective Defect Density Research Articles

Revisiting ‘revisiting supervised methods for effort‐aware cross‐project defect prediction’

AbstractEffort‐aware cross‐project defect prediction (EACPDP), which uses cross‐project software modules to build a model to rank within‐project software modules based on the defect density, has been suggested to allocate limited testing resource efficiently. Recently, Ni et al. proposed an EACPDP method called EASC, which used all cross‐project modules to train a model without considering the data distribution difference between cross‐project and within‐project data. In addition, Ni et al. employed the different defect density calculation strategies when comparing EASC and baseline methods. To explore the effective defect density calculation strategies and methods on EACPDP, the authors compare four data filtering methods and five transfer learning methods with EASC using four commonly used defect density calculation strategies. The authors use three classification evaluation metrics and seven effort‐aware metrics to assess the performance of methods on 11 PROMISE datasets comprehensively. The results show that (1) The classification before sorting (CBS+) defect density calculation strategy achieves the best overall performance. (2) Using balanced distribution adaption (BDA) and joint distribution adaptation (JDA) with the K‐nearest neighbour classifier to build the EACPDP model can find 15% and 14.3% more defective modules and 11.6% and 8.9% more defects while achieving the acceptable initial false alarms (IFA). (3) Better comprehensive classification performance of the methods can bring better EACPDP performance to some extent. (4) A flexible adjustment of the defect threshold λ of the CBS+ strategy contribute to different goals. In summary, the authors recommend researchers and practitioners use to BDA and JDA with the CBS+ strategy to build the EACPDP model.

The Dissociation of Gallium–Hydrogen Pairs in Crystalline Silicon during Illuminated Annealing

Hydrogen‐rich crystalline Si samples are produced by coating Ga‐doped Czochralski‐grown silicon wafers with SiNx:H and subsequent rapid thermal anneal. This procedure introduces H2 dimers into the bulk, which enables the formation of gallium–hydrogen pairs (GaH). The change in pair concentration can be determined by change in resistivity, as pair formation consumes holes. During illumination at elevated temperature (180 °C), some of the previously formed GaH pairs dissociate into H and/or H2 within hours. A subsequent anneal step forms the pairs again. If illumination is prolonged to days at 180 °C, a second dissociation phase occurs after which GaH pairs do not form again in the dark. Therefore, probably two processes are ongoing: a reversible (fast) one and an irreversible (slow) dissociation of GaH pairs in crystalline silicon. The fraction of dissociated pairs and rate of dissociation depend on temperature and excess charge carrier concentration. This suggests an electron‐driven, thermally activated back reaction into either H dimers or atomic hydrogen. An Arrhenius analysis reveals a possibly injection‐dependent activation energy of the reversible dissociation process in the range of 0.64–0.71 eV. Lifetime measurements reveal a drastic increase in effective defect density during the second decrease in pair concentration.

Effect of Channel Thickness on Performance of Ultra-Thin Body IGZO Field-Effect Transistors

Amorphous indium–gallium–zinc oxide (a-IGZO) is a promising channel material for an upper transistor in monolithic three-dimensional devices. Although the field-effect transistors (FETs) with a rather thick channel thickness >10 nm have been intensively examined, less information is available for the IGZO FETs with an ultra-thin body (< 10 nm). In this study, the FETs with the IGZO channel layer ranging from 2 to 20 nm were investigated in detail. As the channel thickness decreased from 20 to 7 nm, the mobility and subthreshold swing (SS) values were improved. In contrast, the deterioration in mobility and SS occurred when the IGZO thickness was less than 7 nm. The physical rationale for the strong IGZO thickness dependence on performance of the resultant FETs was discussed based on subgap density-of-state distribution and mobility models such as percolation and surface-roughness scattering mechanisms using a technological computer-aided design simulation with a quantum mechanical model. IGZO FET with an IGZO thickness of 7 nm exhibited the best performance, which was attributed to the synergic balance by percolation efficiency and reduction in effective subgap defect density of IGZO.

Impact of Dimensions of Memory Periphery FinFETs on Bias Temperature Instability

Fin height and width dependence of negative and positive Bias Temperature Instability (N/PBTI) on logic for memory high- $\kappa $ metal gate (HKMG) FinFET transistors is reported for the first time. It was observed that NBTI degradation is less severe when increasing the physical height of the silicon fin. The increased fin height results in a lower effective defect density, believed to be related to a reduced role of the defective fin corners and/or top surface. In addition, activation energies for the capture process in tall fins during NBTI stress show lower values while charge trapping in standard height fins is highly temperature dependent. PBTI results reveal a similar, albeit less severe, impact of fin height, suggesting an impact of fin height on the high- $\kappa $ layer, with again an increased defectivity at the fin corners and/or top surface, whose effective role is reduced in the case of taller fin. On the other hand, PBTI shows limited temperature dependence, independent of fin height.

Unravelling additive-based nanocrystal pinning for high efficiency organic-inorganic halide perovskite light-emitting diodes

Organic-inorganic halide perovskite light emitting diode (PeLED) as a narrow band emitter is an emerging research field. To overcome limited electroluminescence efficiency of PeLEDs, trap-assisted non-radiative recombination in polycrystalline perovskite films should be reduced and the electron-hole balance in the PeLEDs must be improved. In this work, we investigated a practical way to effectively overcome above-mentioned issues by unravelling additive-based nanocrystal pinning (A-NCP) process using the carefully controlled electron transporting organic material solutions diluted in a volatile non-polar solvent. We found that without affecting the intrinsic crystal structure, A-NCP improved the radiative recombination rate by reducing effective defect density at grain boundaries due to the defect healing effect. Moreover, it induced the improved electron-hole balance in the dominantly p-type CH3NH3PbBr3 based PeLEDs, leading to the highest efficiency of 8.79% ever reported to date among organic-inorganic halide perovskite-based green PeLEDs. Therefore, our work gives the effective approaches for efficient PeLEDs from the investigations of the role of A-NCP incorporating a tiny amount of an electron transporting molecule as an additive to increase radiative recombination rate of polycrystalline perovskite films.

E-beam irradiation effects on IR absorption bands in single-walled carbon nanotubes

Abstract We have measured the absorption and Raman spectral change induced by the irradiation of e-beam. By the irradiation of e-beam on SWNTs thin films, the intensity of defect related Raman band increase, and the peak energy of IR absorption bands shift to the higher energy side. These results indicate that the origin of infrared band is due to the plasmon resonance of finite-length SWNT. We have estimated the effective tube length and defect density from IR absorption peak energy.

Advances in 3x150 mm Hot-Wall and 6x150 mm Warm-Wall SiC Epitaxy for 10kV-Class Power Devices

Initial results from a custom, 3x150 mm horizontal hot-wall, SiC Vapor-Phase epitaxial growth reactor (VPE), with full planetary motion are reported. Epitaxial layer properties are compared with those grown on previously reported 1x150 mm hot-wall and 6x150 mm warm-wall reactors for epitaxial wafers of up to 100 microns thickness. The median intra-wafer thickness and doping uniformity (σ/mean) for the 3x150 mm reactor are 0.9% and 6.10%, respectively. The intra-run thickness and doping variation are 0.18% and 1.01%. The projected 5x5 mm useable area for layers of nominal 5 to 30 μm thickness for all the reactor configurations (as measured by a Candela CS2 optical surface analyzer) are approximately 92%, corresponding to an effective defect density of about 0.35 cm-2. Useable areas as high as 86.5% and 1-2 nm rms roughness have been observed for layers of over 100 μm thickness. Initial morphological results using a higher resolution, LaserTec SICA6X confocal differential interference contrast (CDIC) microscope, are also presented.

An effective medium approach for modeling polycrystalline silicon thin film solar cells

A simple and fast 1D effective medium simulation model for polycrystalline silicon thin film solar cells was defined, by using the device simulator AFORS-HET. The model was calibrated with current–voltage measurements of solar cells under illumination and in t he dark. The experimentally determined variation of the solar cell characteristics with varying absorber doping and absorber thickness can be reproduced by the model. Through inverse modeling, the effective defect density in the absorber layer was determined and it was shown that the open circuit voltage VOC of the cell is limited by defects localized in the bulk of the absorber layer. Using the calibrated simulation model for sensitivity analyses, different optimization strategies for this type of solar cells have been assessed. Firstly, the optimum absorber thickness as a function of absorber doping and defect densities has been determined and secondly, the VOC-increase when using a layer of hydrogenated amorphous silicon (a-Si:H) to form a heterojunction back surface field has been simulated as a function of the absorber defect density.

Gettering effect in grain boundaries of multi‐crystalline silicon

AbstractIn this work, we analyze the effect of three gettering procedures on the variation of the grain boundaries (GBs) defect density in multicrystalline silicon (mc‐Si). The effective defect density (NB) was calculated using a theoretical model where we consider the potential barrier induced by the GB as being due to structural defects and impurities. Results are compared to those obtained from C‐V measurements. The potential barrier was evaluated from the dark current‐voltage (I‐V) characteristic performed across the GB. In addition to the Rapid Thermal Annealing (RTA), we use aluminum (Al) in the first gettering procedure, in the second we use porous silicon (PS), whereas in the third one, we realize a chemical damage (grooving). Mc‐Si wafers were annealed in an infrared furnace in the same conditions, at temperatures ranging from 600 °C to 1000 °C (© 2012 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)

Regression Analysis for Transport Electron Scattering Caused by Structural Defects in InSb Quantum Wells: Application of Matthiessen's Formula

The graphical representation and numerical interpretation of the results obtained by a regression analysis for Matthiessen's formula have been investigated for the electron scattering due to micro-twins (MTs) and threading dislocations (TDs) in InSb quantum wells (QWs) at room temperature. By plotting the reciprocal of the total mobility vs the summation of the linear terms due to MTs and TDs, a two-dimensional graphical representation that clearly exhibits the “goodness of fit” of the regression analysis was depicted. The usefulness of numerical indexes, “effective defect density” and “equivalent defect scattering”, was discussed: “effective defect density” made it possible to evaluate the electron scattering in InSb QWs by using one type of structural defects, either MTs or TDs. “Equivalent defect scattering” enables one to intuitively grasp the difference in magnitude of electron scattering between MTs and TDs in InSb QWs.

Growth of Ultra-thin Titanium Dioxide Films by Complete Anodic Oxidation of Titanium Layers on Conductive Substrates

ABSTRACTThe growth of thin and ultra-thin titanium dioxide layers was investigated. Oxide films were grown by galvanostatic and potentiodynamic anodisation of evaporated titanium layers on conductive substrates. It is shown that thin-film oxidation differs significantly from anodic oxidation of solid foils or plates, due to the sudden stop of anodisation process before complete oxidation of the thick films. Depending on the pH value and the potential sweep rate, the effective defect density and the dielectric constant of the anodized layers vary from 3·1019 cm-3 to 1020 cm-3and from 16 to 27, respectively, whereas the electrolyte temperature plays only a minor role.

Effects of Nitride-Based Plasma Pretreatment Prior to SiNx Passivation in AlGaN/GaN High-Electron-Mobility Transistors on Silicon Substrates

The effects of nitride-based plasma pretreatment on the output characteristics of AlGaN/GaN high-electron-mobility transistors (HEMTs) on silicon substrates are investigated. N2 and NH3 plasma pre-treatment methods are studied to overcome the RF dispersion phenomenon caused by nitrogen-vacancy (VN)-related defect reduction. It is found that the nitride-based plasma pretreatment is effective to overcome the RF dispersion in AlGaN/GaN HEMTs on Si. The NH3 plasma pretreatment markedly reduced RF dispersion from 63 to 1%. This is considered to be attributable to the reduction of the effective VN-related defect density and elimination of carbon/oxide residuals on the surface of AlGaN/GaN HEMTs. A NH3 plasma pretreatment prior to SiNx 100 nm passivation in the AlGaN/GaN HEMTs on Si markedly improves the total output power from 15 to 18.1 dBm under the operating conditions of VDS = 15 V/VGS = -1 V.

A Novel Wafer-Yield PDF Model and Verification With 90–180-nm SOC Chips

In this paper, we describe a new wafer-yield distribution model, which agrees well with experiment using fabricated products with various process technologies. To investigate physical reasoning of the proposed model, we firstly measure effective defect density of chips regarding to spatial dependency in a wafer. It is clarified that the defect density near wafer edge is a couple of times larger than that at the rest of wafer area. Note that the increase of defect at wafer edge causes a significant yield loss in production process lines.

Ion implantation induced disorder in single‐crystal and sputter‐deposited polycrystalline CdTe characterized by ellipsometry and backscattering spectrometry

AbstractBulk single‐crystal CdTe, sputter‐deposited polycrystalline CdTe films of about 2.5 micron thickness, and single‐crystal Si (c‐Si) have been ion implanted using 350 keV Xe at fluences ranging from 1×1013 to 16×1013 cm–2 so as to create disorder in a controlled way from fully single‐crystalline to fully amorphous material. The general purpose of the investigations is to seek a parameterization of the critical point structures and establish a database for fitting the optical properties of CdTe films having different unknown grain sizes whereby the grain size will be described in terms of an effective defect density. The polycrystalline CdTe samples were magnetron sputtered onto c‐Si followed by CdCl2 and Br2‐methanol treatment to improve properties in terms of grain size and surface smoothness, respectively. The fluences for use in the ion implantation of CdTe were estimated using the SRIM (Stopping and Range of Ions in Matter) software, and cross‐checked by simultaneous implantation of bulk c‐Si samples. The optical properties were characterized by second derivative analysis and by a generalized critical point model. Although the damage created by 350 keV Xe in the simultaneously implanted c‐Si samples, as measured by both spectroscopic ellipsometry and Rutherford backscattering/ channeling spectrometry (RBS/C), agrees well with the expectations based on the SRIM simulation, the damage created in CdTe remains at a very low level even for doses several times higher than the amorphization level estimated by simulation. The character of the dechanneling of the RBS/C spectra indicates extended defects (presumably dislocation loops). This effect was similar in both single‐crystal and thin film polycrystalline CdTe, although less pronounced in thin film samples. (© 2008 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)

Analysis of structure and defects in thin silicon films deposited from hydrogen diluted silane

Thin silicon layers have been deposited from silane diluted with hydrogen. The dilution ratio R ( R = [H 2]/[SiH 4]) has been varied between R = 0 and R = 40. The structural properties of Si:H films have been studied using transmission electron microscopy imaging and Raman spectroscopy. The phase evolution from the amorphous phase into the mixed and eventually microcrystalline phase strongly depends on the hydrogen dilution. The initiation of the microcrystalline growth occurs between R = 20 and R = 25. The phase transition becomes more abrupt with increasing hydrogen dilution. Optoelectronic properties of the layers have been determined. Increasing hydrogen dilution results in films with increasing effective defect density and Urbach energy, which is related to inhomogeneous growth. The charge deep-level transient spectroscopy technique (Q-DLTS) was applied for the first time on hydrogen diluted thin silicon films in order to investigate the energy distribution of the defect states in these layers as a function of the dilution ratio R. The Q-DLTS spectra indicate a difference in defect-state distribution when the films evolve from the amorphous phase into the microcrystalline phase.

CRITICAL CURRENT DENSITIES IN Bi1.7Pb0.3-xGdxSr2Ca3Cu4O12+y (x=0.01, 0.1) SUPERCONDUCTORS PREPARED BY MELT-QUENCHING METHOD AND ANNEALED IN DIFFERENT TIME INTERVALS

We report critical current density (Jc) behaviors of the Bi 1.7 Pb 0.3-x Gd x Sr 2 Ca 3 Cu 4 O 12+y (x=0.01, 0.1) superconductors prepared by melt-quenching method and annealed in different time intervals at 840°C. The effects of annealing times are found to be very important. For example, for the sample B2, the highest value of Jc is 3.7×104 A / cm 2 at 9 K. This value is almost 1.5 and 2 times greater than that of the samples B3 and B1, respectively. With increasing annealing time, the 2223 phase increases, thus resulting in a decrease in the insulating phase. For lower annealing times, there are a large number of insulating phases. On the whole, our results indicate that there is an optimum annealing time t to form the 2223 phase, in a time interval of 120<t<192 h . It is clearly seen that Jc decreases smoothly with increasing temperature for all the samples studied. At lower temperatures, especially below 30 K, an extremely rapid decrease of Jc with increasing temperature has been observed. The experimental values also show that between the temperature interval from 9–45 K, Jc decreases by one order of magnitude. We have also observed that when the Gd – Pb substitution increases, the effective defect density increases, leading to small critical current densities.

Peak effect and Bose-glass transition of the heavy-ion irradiated Bi2212: Effect of the defect configuration

Bi 2Sr 2CaCu 2O 8+ δ single crystals were irradiated with 5.8 GeV Pb-ions to introduce columnar defects (CD). The orientation of the defect structure was altered by changing the direction of incident beam. The peak effect was observed in the sample with CD|| c-axis|| B a, and it smeared gradually as CD was tilted away from the field direction. The irreversibility field H irr was scaled with the effective defect density, except for the region where the “recoupling” transition occurs.

Electronic properties of highly P and B doped thin Si layers grown by ECR–CVD

Thin Si films strongly doped with P and B are grown at 325°C by electron cyclotron resonance chemical vapor deposition (ECR–CVD) on quartz substrates and Si wafers. Films deposited on quartz are microcrystalline. By an analysis of Hall effect data it is shown that the electrical transport in these films is controlled by potential barriers at grain boundaries. Using a simple barrier-limited transport model it is found that the effective interface defect density Q it increases with the doping level N d,a following a power law. A study of electron spin resonance shows that the in-grain mobility of electrons is much higher than the Hall mobility. At high doping levels the films grew epitaxially on Si substrates. Such films are used as emitter layers in solar cell structures and efficiencies of 14% are demonstrated.

Effective defect density for MOS breakdown: Dependence on oxide thickness

A procedure is introduced for measuring an effective density of defects that takes into account time-dependent dielectric breakdown in MOS devices. Measurements are obtained that show a surprising exponential decrease in this density with decreasing oxide thickness.