Gui-Xia Yang

Amorphous tantalum pentoxide (Ta₂O₅) based resistive memory devices are sensitive to displacement damage from energetic particles. Energetic particles can displace atoms to produce oxygen vacancies in the oxide. Conductivity of the resistive devices increases with the increase in the density of oxygen vacancies, as these vacancies are responsible for the electron transport through the oxide. The displacement behaviours of tantalum and oxygen atoms in Ta₂O₅ films are studied in detail using molecular simulations. Then, the threshold displacement energies of the two spices of atoms are extracted from the simulation data to calculate the density of atom vacancies from displacement damage. Moreover, an analytical model based on Poole-Frenkel mechanism is proposed to calculate the current formed by oxygen vacancies. The model provides a good fit to the previous experiments for the degradation of off-state resistance (R_off) for Ta₂O₅-based resistive devices subjected to 800 keV tantalum ions. This model also suggests a linear relationship between the change in the reciprocal of the sixth root of R_off (Δ(1/ ⁶ √ R_off)) and the ion fluence (Φ), i.e., Δ(1/ ⁶ √ R_off)∝Φ.

The ionization defects in bipolar junction transistors (BJTs) are sensitive to ionization dose rate and molecular hydrogen in oxides. The bimolecular reaction model, a physical model with self-coincidence forms, is established in earlier works to describe formations of ionization defects in BJTs. This work proposes a deep-learning model with artificial neural network (ANN) framework to calculate the densities of ionization defects. The ANN model can be trained using data prepared by the physical model in conjunction with numerical calculations. Case-studies with a dataset of ~6.0×10 ⁵ points suggest an optimized ANN model which is composed of four hidden layers and sixteen ELU-activated (Exponential Linear Unit) neurons per hidden layer. This model provides excellent fits to the data from physical model for the dose-, dose-rate- and hydrogen-dependence of ionization defects in gate-controlled p-n-p BJTs. Our encouraging results show that the deep-learning modeling methodology may find its effective application in dynamics of ionization defects.

The Insulator Gated Bipolar Transistor (IGBT) is characterized by MOS-gating, high current density, and high blocking capabilities. The trench field-stop reverse-conducting IGBT (TFR-IGBT) is distinguished from conventional IGBT by trench MOS gate, field-stop layer and collector-short structure. As a composite structure made of metal-oxide-silicon and bipolar junction transistors, TFR-IGBT is susceptible to displacement damage. This work reports the displacement damage effects on TFR-IGBT with the fast neutron fluence up to a fluence of 10¹³ cm^-2. The transfer, forward conductive and forward blocking characteristics are degraded subsequent to neutron exposures. The suppression of the hump current in the transfer curve is observed by neutron-induced damage. This paper proposes, from a device physics perspective, the mechanism behind the characteristics degradation from displacement damage in TFR-IGBT. The dependencies of the key parameters on neutron fluence are analytically modeled. Our models provide a good fit to the experimental data of the IGP20N65H5 TFR-IGBT subjected to fission neutrons.

Ionizing radiation experiments on gated controlled lateral p-n-p bipolar transistor (GLPNP) show that the base current obtained by gate sweeping exhibits two ”unusual” phenomena, including decreasing in amplitude and narrowing in peak profile after radiation. Both interface traps and border taps are called switching states due to their similar recombination effects on substrate carriers. The two traps can attribute to the base current. The separation of interface traps and border taps is achieved by frequency-dependent charge pumping technique. Based on TCAD simulation in conjunction with device theory, this work suggests that the two ”unusual” phenomena can be explained by the combined effects of border traps and interface traps.

The anode-short MOS-Controlled Thyristor (AS-MCT) is the latest generation of MCT. Due to its MOS-gating, high-current-rise rate, normally-off and high blocking capabilities, the AS-MCT is an ideal switch for pulse discharge application. Previously, we show that the static behaviours of AS-MCT are susceptible to displacement damage as it contains bipolar structures. As a continuation of our previous work, this article reports the experimental results for the degradation of pulse discharge circuit characteristics induced by the displacement damage of its XND1 AS-MCT switch following fast neutron exposure up to 7.8×1013 cm-2. Both the charging time and the peak surge current of the pulse discharge circuit decrease significantly once the neutron flux surpasses a critical value, ~1012 cm-2. From device and circuit physics perspectives, an analytical model base on equivalent leakage current resistance is proposed to describe the degradation of the two parameters, and then, this article suggests the mechanisms behind such the degradations from the neutron exposure of AS-MCT switch.

The increase of base current is the main damage effect of p-n-p bipolar junction transistors (BJTs) subjected to ionization radiation. The Gummel-Poon model is widely used to describe the electrical performance of BJTs. For the purpose of simulating the ionization damages on bipolar circuit performance, this article develops a Gummel-Poon SPICE model that incorporates the latest formula of excess base current. Then, this article reports the ionization damage effects on a p-n-p controlled common-emitter amplifier circuit which is a key unit in bipolar integrated circuit. After γ-ray exposures, both the output voltage of the common-emitter circuit and its relaxation time decrease. Two novel phenomenons are observed for the ionization damages on the common-emitter circuit, i.e., the degradation of output voltage exhibiting time-dependent effects and sub-linear dependence on dose. Moreover, this article proposes, from perspectives of SPICE simulation and circuit theory, the mechanisms behind the characteristics of degradation from ionization damage in the common-emitter circuit. The SPICE simulations provide a good fit to the experimental data for the response of output voltage of a 2N2907 p-n-p controlled common-emitter circuit to ionization damage with a dose up to 2000 Gy(SiO2). Also, this article presents two applications of the new SPICE model in quantitatively describing the ionization damages on basic circuits, including crosstalk-related degradation and process-related statistical distribution of degradation.

This work focuses on the recombination-enhanced reactions of boron related defects in compensated silicon and their impact on the electrical performance of silicon devices. Using deep level transient spectroscopy, we measured the defect spectra in the collector of 14 MeV fusion neutron irradiated p - n - p silicon transistors during forward current injection, as well as the corresponding degradation kinetics of both current gain and leakage current on the same device. Two hole traps at E c + 0.43 eV and E c + 0.53 eV are proved to be acceptors in the lower half of the gap with significant recombination enhanced effects. These metastabe acceptors degrade the current gain by changing the space charge distribution in the p - n - p structure and are linked with carrier-induced boron-oxygen complexes responsible for the minority-carrier lifetime degradation of silicon photovoltaics.

Space particle irradiation produces ionization damage and displacement damage in semiconductor devices. The enhanced low dose rate sensitivity (ELDRS) effect caused by ionization damage has attracted wide attention. However, the enhanced low-particle-flux sensitivity effect and its induction mechanism by displacement damage are controversial. In this paper, the enhanced low-neutron-flux sensitivity (ELNFS) effect in Boron-doped silicon and the relationship between the ELNFS effect and doping concentration are further explored. Boron-doped silicon is sensitive to neutron flux and ELNFS effect could be greatly reduced by increasing the doping concentration in the flux range of 5 × 10⁹–5 × 10¹⁰ n cm⁻² s⁻¹. The simulation based on the theory of diffusion-limited reactions indicated that the ELNFS in boron-doped silicon might be caused by the difference in the concentration of remaining vacancy-related defects (V_r) under different neutron fluxes. The ELNFS effect in silicon becomes obvious when the (V_r) is close to the boron doping concentration and decreased with the increase in boron doping concentration due to the remaining vacancy-related defects being covered. These conclusions are confirmed by the p⁺-n-p Si-based bipolar transistors since the ELNFS effect in the low doping silicon increased the reverse leakage of the bipolar transistors and the common-emitter current gain (β) dominated by highly doped silicon remained unchanged with the decrease in the neutron flux. Our work demonstrates that the ELNFS effect in boron-doped silicon can be well explained by noise diagnostic analysis together with electrical methods and simulation, which thus provide the basis for detecting the enhanced low-particle-flux damage effect in other semiconductor materials.

GaAs has been demonstrated to be a promising material for manufacturing semiconductor light-emitting devices and integrated circuits. It has been widely used in the field of aerospace, due to its high electron mobility and wide band gap. In this study, the structural and photoelectric characteristics of Si-doped GaAs under different gamma irradiation doses (0, 0.1, 1 and 10 KGy) are investigated. Surface morphology studies show roughen of the surface with irradiation. Appearance of transverse-optical (TO) phonon mode and blueshift of TO peak reflect the presence of internal strain with irradiation. The average strain has been measured to be 0.009 by Raman spectroscopy, indicating that the irradiated zone still has a good crystallinity even at a dose of 10 KGy. Photoluminescence intensity is increased by about 60% under 10 KGy gamma irradiation due to the strain suppression of nonradiative recombination centers. Furthermore, the current of Si-doped GaAs is reduced at 3V bias with the increasing gamma dose. This study demonstrates that the Si-doped GaAs has good radiation resistance under gamma irradiation, and appropriate level of gamma irradiation can be used to enhance the luminescence property of Si-doped GaAs.

Silicon carbide (SiC) has been widely used for electronic radiation detectors and atomic battery sensors. However, the physical properties of SiC exposure to high-dose irradiation as well as its related electrical responses are not yet well understood. Meanwhile, the current research in this field are generally focused on electrical properties and defects formation, which are not suitable to explain the intrinsic response of irradiation effect since defect itself is not easy to characterize, and it is complex to determine whether it comes from the raw material or exists only upon irradiation. Therefore, a more straightforward quantification of irradiation effect is needed to establish the direct correlation between irradiation-induced current and the radiation fluence. This work reports the on-line electrical properties of 4H-SiC Schottky barrier diodes (SBDs) under high-dose electron irradiation and employs in situ noise diagnostic analysis to demonstrate the correlation of irradiation-induced defects and microscopic electronic properties. It is found that the electron beam has a strong radiation destructive effect on 4H-SiC SBDs. The on-line electron-induced current and noise information reveal a self-healing like procedure, in which the internal defects of the devices are likely to be annealed at room temperature and devices’ performance is restored to some extent.