In this article, we provide a comprehensive review of defect formation at the atomic level in interfaces and gate oxides, focusing on two primary defect types: interface traps and oxide traps. We summarize the current theoretical models and experimental observations related to these intrinsic defects, as they critically impact device performance and reliability. By integrating theoretical insights with experimental data, this review provides a thorough understanding of the atomic-scale interactions that govern defect formation.

@en

Due to the deficient passivation of the interface between silicon carbide and silicon dioxide, the defect-induced capture and release of trapped charges triggered by external Bias Temperature Stress (BTS) leads to parameter shifts and degraded device performance. This study models the trap-induced transient current in silicon carbide metal-oxide-semiconductor capacitors, providing insight into how capacitance and conductance change during C-V measurements under conditions of high temperature, varied frequency, and varied applied voltage.@en

The demand for accurate temperature sensing in extreme temperatures is increasing. Traditional silicon-based integrated temperature sensors usually cannot survive above 200 °C. Many researchers have started to focus on semiconductors with a large bandgap. Among them, silicon carbide (SiC) is the most promising one. Nevertheless, most reported SiC sensors are in the form of discrete components and are not compatible with integrated electronics. In this work, we demonstrate an open 4H-SiC CMOS technology, and the fabrication steps are detailed. The temperature sensing elements in this technology, including resistors based on different implanted layers and MOSFETs, are characterized up to 600 °C. At room temperature, the resistive-based elements demonstrate large negative temperature coefficients of resistance (TCRs). With increasing temperature, the TCR starts to decrease and even becomes positive. The TCR change is due to the interplay between increasing dopant ionization rate and decreasing mobility as a function of temperature. The resistance change with temperature fits well into the Steinhart-Hart model and second-order polynomial equation. The p-type diode-connected MOSFET has a sensitivity of 4.35 mV/°C with a good linearity. The nMOS-based sensor has a maximum sensitivity of -9.24 mV/°C but a compromised linearity. The characterization of these sensing elements provides important results for potential users who will work on SiC integrated temperature sensing with this technology.

@en

In this work, a highly linear temperature sensor based on a silicon carbide (SiC) p-n diode is presented. Under a constant current biasing, the diode has an excellent linear response to the temperature (from room temperature to 600°C). The best linearity (coefficient of determination ${R}^{{2}}$ = 99.98%) is achieved when the current density is 0.53 mA/cm2. The maximum sensitivity of the p-n diode is 3.04 mV/°C. The temperature sensor is fully compatible with Fraunhofer Institute (FHG) IISB's open SiC CMOS (complementary metal-oxide-semiconductor) technology, thus enabling the monolithic integration with SiC readout circuits for high-temperature applications. The sensor also features a simple fabrication process. To our knowledge, the presented device is the first SiC diode temperature sensor that does not require a mesa etch or backside contacts.

@en

This paper presents p-n diode temperature sensors and MOSFET temperature sensors in low-voltage silicon carbide (SiC) CMOS technology. The reported temperature sensors directly make use of the existing doping layers in the CMOS process, thus enabling the monolithic integration of the SiC temperature transducer and the SiC readout electronics. The temperature sensor is characterized from 25 to 200°C. The diode-based temperature sensor has a maximum sensitivity of 3.27 mV/°C and a maximum R 2 of 99.81%. The MOSFET-based temperature sensor achieved a maximal sensitivity of 16.5 mV/°C, however, with less linearity (R 2 max = 99.11%). This technology shows a unique potential for implementing harsh environment smart temperature sensors.@en