Solid-state Electronics最新文献

In this paper, accurate temperature-dependent static model for Silicon-Carbide (SiC) power MOSFET is presented. The proposed model is formed by two equations relating to linear and saturation operating regions. In this model, new formalism of the saturation drain current is introduced to consider the peculiar features observed in the I-V static characteristics of the SiC power MOSFET: a) moderate inversion region, or region of low gate voltages and b) quasi-saturation region, region of high gate voltages at which the drain current becomes less sensitive to the increase of gate voltage. In addition, the model captures with high-precision the transition region between linear and saturation region, pinch-off region, noticed in the output characteristics of the SiC power MOSFETs. It will be shown that the model equations ensure continuity and smooth transition between all operating regions. Temperature scaling of the model is carried out by its temperature scaling parameters. The proposed compact model is simple and efficient using reduced number of technology independent parameters. Simple parameter extraction procedure is described that uses an optimizer algorithm based on good experimental initial guess. Excellent agreement is obtained by comparing model to TCAD simulation and device measurement.

The AlGaN/GaN high electron mobility transistors (HEMTs) with T-gate that suitable for high frequency applications were fabricated. A novel method to extract the bias-dependent source and drain parasitic series resistances (R_s and R_d) of AlGaN/GaN HEMTs is proposed. By analyzing the distributed capacitance and current generator network in the velocity saturated regions of the AlGaN/GaN HEMTs, a new restriction relationship between small-signal equivalent circuit elements is found. The R_s and R_d can be determined under active bias through wideband S-parameter measurements, which can better reflect the physical mechanism of AlGaN/GaN HEMTs under normal operation. The S-parameters and extrinsic transconductance calculated based the small-signal equivalent circuit element values extracted by the method proposed in this paper are very consistent with the experimental values, which reflects the accuracy of this element extraction method. In this paper, the physical mechanism that causes R_s and R_d to vary with bias voltage is also studied. This study has a deeper insight into the bias-dependence of R_s and R_d, which modifies the understanding for physical mechanisms of AlGaN/GaN HEMTs. The research results provide new ideas for establishing small-signal equivalent circuit models containing more physical effects and is of great significance to GaN-based integrated circuit design.

The metal and two-dimension (2D) semiconductor contact interfaces have a more considerable contact resistance hindering carrier injection, which makes the performance of 2D semiconductor devices less than the theory. The contact properties of Ni, Au, and Mo with MoS₂ are simulated by the first-principles method. The interface dipole caused by the interface charge redistribution changes the work function difference at the metal-MoS₂ interface, so the interface charge redistribution is one of the important factors for correctly evaluating the contact properties. Due to the metal-induced gap states (MIGS) at metal-monolayer (ML) MoS₂ interfaces, the Fermi level is strongly pinned to fixed energy, and the Schottky barrier height (SBH) cannot be regulated efficiently by the metal work function. Although the work function of Au is bigger than Ni, the Fermi level of Au is pinned at a higher position. In the meantime, the bandgap of MoS₂ narrows and metallization occurs due to the larger MIGS. In the Mo-MoS₂ interface, the Fermi level is pinned near the conduction band minimum of MoS₂. The contact resistances (R_c) of the three structures are tested by the Circular Transfer Length Method (CTLM), which is consistent with the prediction of the simulation. The Mo-MoS₂ has the smallest R_c. The results indicate that contact resistance of 2D semiconductors cannot be simply predicted by soled work functions or Fermi level pinning, but is determined by several factors.

In this work, we present a machine-learning augmented compact modeling framework for modeling process induced variations in advanced semiconductor devices. The framework employs BSIM-CMG unified compact model at the core and can be used for any advanced devices like GAA nanosheets and nanowires, FinFETs etc. We have validated the model with extensive numerical simulations and experimental data such as $14\mathrm{nm}$ technology FinFET and $24\mathrm{nm}$ technology Nanowire. Our results show excellent accuracy in modeling variability in key electrical parameters of the device including off-current (I_off), on-current (I_on), threshold voltage (V_th), subthreshold swing (SS) etc. We observe that the overall accuracy of the ML-based framework strongly depends on the nature and physical behavior of the core model used for modeling the nominal device.

We present a versatile compact model for resistive random-access memory (RRAM) that can model different types of RRAM devices such as oxide-RRAM (OxRAM) and conducting-bridge-RRAM (CBRAM). The model unifies the switching mechanisms of these RRAMs into a single framework. We showcase the model’s accuracy in reproducing published experimental device DC and transient characteristics of various RRAM structures. We also demonstrate the model’s efficacy in capturing RRAM variability and conducting 1T1R circuit simulations.

A normally-off GaN MOSFET is successfully fabricated by using the selective regrowth technique (SRT) with regrown AlGaN layer on source/drain (S/D) region. The GaN MOSFET with regrown AlGaN layer and L_g of 10 μm shows enhanced electrical performance such as maximum drain current (I_D,max) of 57 mA/mm, maximum transconductance (g_m,max) of 11 mS/mm, and field-effect mobility (μ_FE) of 59 cm²/V·s, respectively, compared to the GaN MOSFET with n⁺-GaN selective regrowth in S/D region. This is because of the high 2DEG density formed by AlGaN/GaN heterojunction in S/D region. Moreover, to accommodate the poor structural quality of the narrow region regrowth of AlGaN layer on the S/D region, wide regrown AlGaN layer is applied to the GaN MOSFET. Especially, the off-state breakdown voltage improves from 25 V to 192 V with the improved structural quality of wide regrown AlGaN layer and optimized structure and the application of the 70-nm thick SiO₂ passivation. These result shows that GaN MOSFET with wide regrown AlGaN layer on S/D region is beneficial to achieving high-quality and uniform normally-off GaN MOSFETs with excellent electrical performance.

This paper presents a hybrid model for TiO₂-based memristors, integrating the dopant drift mechanism with the Schottky barrier theory. We introduce the movement of oxygen vacancies as a dynamic variable to modulate changes in memristors. Furthermore, the variation of the dominate mechanism of the TiO₂ memristors under different operating conditions is studied, which is related to the position of the internal oxygen vacancy. The proposed model accurately captures the rectification linearity, and effectively elucidates the dominant current mechanisms manifested in six distinct regions of the I-V curves. Our model exhibits better predication with reduced errors when applied to Pt/TiO₂/Pt memristors. The proposed model can well describe the dual-mechanism memristor phenomenon, and provides a reference for the subsequent study of multi-mechanism behavior in memristors.

The large numbers of high-energy carriers that occur in semiconductor devices under semi-on state conditions can cause significant device degradation. The effects of different stresses on the electrical and trapping characteristics of GaN-based high-electron-mobility transistors (HEMTs) are investigated. Test results for GaN HEMTs under semi-on state conditions show that the electrical characteristics of these devices degrade to a certain extent after they are subjected to electrical pulse stress cycles with different drain voltage, frequencies, and duty cycles; a degree of degradation also occurs in the electrical characteristics of the devices when they are subjected to direct current electrical stresses. After electrical stress is applied, the absolute amplitude of the traps in the device increases, thus indicating an increase in the trap density. The results show that voltage is the main driver for device damage, with the current playing an accelerating role through its effects on device temperature or by supplying hot electrons; therefore, the drain voltage has the most significant effect on device degradation, which is mainly due to channel high-energy hot electron injection.

Significant discrepancies were found between experimental results and the results calculated by the conventional physics-based model for the cutoff frequency and some equivalent circuit parameters of double heterojunction bipolar transistors (DHBT). In order to accurately evaluate the primary quantitative performance of DHBT, a comprehensive physics-based model was developed and validated by comparing experimental data from three research institutions. The proposed physics-based model combines the equivalent circuit of the T-topology and hybrid-π topology, and includes modification formulas for estimating the intrinsic dynamic resistance of the base–collector and base-emitter junctions, as well as the cutoff frequency, the hybrid-π input capacitance, and the gain.

In this research, we reported that manganese and copper atoms were embedded in silicon-carbon films to fabricate impedance organic vapor sensors. Gas sensitive layers were formed using electrochemical deposition of 9:1 CH₃OH/HMDS solutions, followed by thermal annealing at 500 °C for 2 h. Silicon-carbon films contain 4H-SiC, 15R-SiC and 6H-SiC polytypes, as well as amorphous diamond phases. Mott-Schottky plots were used to evaluate the silicon-carbon films conductivity type, flat band potential and carrying density. Sensor operations were examined at ambient temperature and up to 80 % relative humidity to assess their functionality. The silicon-carbon films impedance sensors detected 6–37 ppb toluene vapor. The manganese and copper embedded in silicon-carbon films detected 5–52 ppb isopropanol vapor and remained unchanged in humidity range (40–65 %). However, at humidity level up to 80 %, the sensing response range decreases by ≈1.5–2 times, with isopropanol significantly contributing to the response.