Microelectronic Engineering最新文献

As integration continues in the modern semiconductor industry, copper (Cu) is used for metal lines and low dielectric constant (low-k) films are used for intermetal dielectrics (IMD) to reduce signal delays occurring in device interconnects. A diffusion barrier is essential between the Cu metal lines and the IMD to prevent Cu diffusion, and silicon carbon-nitride (SiCN) films with relatively low dielectric constants are being widely studied. In this study, SiCN films deposited from 1-(trimethylsilyl)pyrrolidine (TSPD) precursor by plasma-enhanced chemical vapor deposition (PECVD) were investigated for use as a Cu diffusion barrier in multilevel metallization process. This plasma-polymerized TSPD (ppTSPD) monolayer film as SiCN was deposited in plasma powers ranging from 15 to 30 W. The electrical properties of ppTSPD were measured and the chemical properties were analyzed by Fourier-transform infrared spectroscopy (FTIR). The dielectric constant increased with increased plasma power. The lowest dielectric constant of 3.70 and leakage current density at 1 MV/cm of 2.27$\times$10⁻⁸ A/cm² were found for ppTSPD film deposited at 15 W. To verify the Cu diffusion barrier characteristics of the ppTSPD films, a ppTSPD/ppOMCTS bilayer was introduced by using plasma-polymerized octamethylcyclotetrasiloxane (ppOMCTS) as porous low-k SiCOH films. The time-dependent dielectric breakdown (TDDB) characteristic was enhanced around five times than ppOMCTS monolayer used as a reference. The ppTSPD was suggested for fabricating SiCN films for use as a Cu diffusion barrier in multilevel metallization process.

The radio frequency (RF) chips and passive devices integrated on the through-glass-via (TGV) substrate meets the demands of miniaturization, high performance and low losses in the application. The RF chip and integrated passive devices (IPDs) are interconnected electrically by a redistribution layer (RDL) on the TGV substrate with the isolation low-k materials. The low-k materials, however, are susceptible to fracture during the thermal process in packaging due to their weak mechanical properties. In this study, the fractures of the low-k material were studied by experiments and the finite element analysis (FEA) for a RF package with integrated passive device based on TGV. The mechanical properties of the low-k material used in the FEA were tested by fabricating freestanding low-k films using microfabrication techniques. Then the fracture behaviors of the low-k material in the package and its impact factors under thermal loadings were examined. The impact factors, such as the initial defect location, direction and length, were investigated by evaluating the stress intensity factors (SIFs) at the defect tips. The results revealed that the most hazardous location in the low-k material is the region below the micro-joint of RF chip. The vertical defects along thickness in low-k film are more likely to propagate than horizontal ones. The SIF value increases linearly with the defect length both in heating and cooling conditions.

Flexible low dielectric constant (low-k) SiCOH thin films were fabricated onto flexible indium tin oxide coated polyethylene naphthalate (ITO/PEN) substrates using plasma-enhanced chemical vapor deposition (PECVD) of a tetrakis(trimethylsilyloxy)silane (TTMSS) precursor. RF plasma powers of 20 and 60 W were utilized for the deposition. The k-values of the pristine SiCOH films deposited at 20 and 60 W were 2.46 and 2.00, respectively. Both films showed hydrophobic surfaces. An inductively coupled plasma-reactive ion etching (ICP-RIE) process was then performed on the flexible SiCOH thin films using CF₄, CF₄ + O₂, and CF₄ + Ar. The surface wettability of the films increased substantially following etching, with many of the etched films being considered hydrophilic. The Fourier transform infrared (FTIR) spectra of the pristine films identified four prominent absorption bands as CH_x stretching, Si-CH₃ bending, Si-O-Si stretching, and Si-(CH₃)_x stretching vibration modes. After the etching process, the peak area ratios of Si-O-Si stretching mode increased and those of Si-(CH₃)_x stretching mode decreased. The X-ray photoelectron spectroscopy (XPS) spectra analysis determined significant concentration of fluorine on the surface of the film following etching. From the high-resolution XPS scan, it was found that the peak intensity of the C1s and Si2p peaks decreased after etching process and the peak center of the F1s peak shifted depending on etching chemistry. The k-values of the films at 20 W were fairly consistent while those of the films at 60 W increased significantly following the etching process. The increase in k-value after etching for the films at 60 W correlated with surface hydrophilicity, increase in the refractive index, and change in the peak area ratios of Si-O-Si and Si-(CH₃)_x stretching modes in the FTIR spectra.

A comprehensive study has been done on the influence of post-deposition annealing temperature on high-k cerium oxide (CeO₂) layer grown on n-type silicon (Si) substrate and its resultant interface states have been studied for Al/CeO₂/Si metal-oxide-semiconductor (MOS) devices. The high-k CeO₂ thin films were deposited by spin-coating and sintered at different annealing temperatures (T_a) in the range of 400–900 °C. The parameters such as fixed charge density (Q_eff), dielectric constant (k) of the layers, flat-band voltage (V_FB), interface defect density (D_it), etc., of the MOS device were evaluated from CV and I-V measurements. A minimum value of flat band shift (∼0.05 V) with lower Q_eff (−4.81 × 10¹¹ C/cm²) have been achieved for the T_a of 600 °C. The k and D_it were evaluated to be 22 and 1.29 × 10¹² cm⁻², respectively at the T_a of 600 °C. In addition, the CV measurements showed a very small hysteresis and very low frequency dispersion for the T_a of 600 °C sample. Energy distribution of defect states was evaluated and it was maximum towards the bottom of the conduction band. This shows that the 600 °C is the optimum annealing temperature, which results in high quality interface, and the electron affinity of the corresponding CeO₂ layers was found to be 3.29 eV as evaluated from ultraviolet photoelectron spectroscopy (UPS). Further, a maximum value of minority carrier lifetime (147 μs) has been achieved for the samples annealed at T_a of 400 °C, indicating that the post-annealing temperature plays a significant role on the properties of CeO₂ films deposited by sol-gel process. Thus, the present study demonstrates the possibility of sol-gel grown high k-CeO₂ layers suitable for MOS like devices.

In this paper, a comparative study of C-, N- and Xe-based pre-amorphization implantation (PAI) processes is proposed. The impact of the use of such processes on the agglomeration resistance and physical properties of the final Ni(Pt)Si layer, as well as the formation mechanisms via solid-state reactions and electrical performances via the transfer length measurement (TLM) method, is evaluated. It is shown that although all species are able to increase the agglomeration temperature of Ni(Pt)Si layers (up to more than 100 °C), the underlying mechanisms are different. For C- and N-based PAI processes a strong chemical effect is observed, while for Xe-based processes the amorphization depth plays an important role. Consequently, the beneficial effect of stabilizing Ni(Pt)Si layers at high temperatures using C- and N-based PAI processes has to be balanced with an increased layer resistivity (up to 30%) combined with a strong deterioration of the associated specific contact resistivity (which is multiplied by almost a factor 10). In this sense, Xe-based PAI processes seem to be a better option as they could allow to combine both requirements.

The separation band of perception, storage, and computation modules in vision systems based on traditional von Neumann architectures leads to latency and power consumption problems in data transmission, which severely limits the computational power. In recent years, in-sensor computing has gained significance in enhancing the computational performance of machine vision systems. It integrates sensing, storage and computation and is an important way to break out of the Von Neumann architecture. This study introduces an optoelectronic memristor-based image recognition algorithm to improve recognition efficiency by performing image feature extraction in a hardware array. The experimental results show that the network achieves the best accuracy of 93.26% after 30 epochs, and the loss of accuracy after weight quantization is about 1%.

This investigation identifies the critical solder joint in a typical Insulated Gate Bipolar Transistor (IGBT) module and provided new knowledge on how operating thermal loads degrade IGBT-attach, Diode-attach, and Substrate solder joints in the device. SolidWorks software is used to create three realistic 3-D Finite Element (FE) models of the typical IGBT module used in this investigation. In-service operating power and IEC 60068–2-14 thermal cycles are implemented in ANSYS mechanical package to simulate the response of the three solder joints in the FE models to the load cycles. The solder in the joints is lead-free alloy of 96.5% tin, 3% silver, and 0.5% copper (SAC305) composition. The SAC305 material properties are modelled as time and temperature dependent with Anand's visco-plastic model employed as the constitutive model. Results show that the key degradation mechanism of solder joints in IGBT module are stress, plastic strain, and strain energy magnitudes. Accumulated plastic strain in the joints is found the predominant damage factor. Critical solder joint in the module depends on the load cycle the device experiences. IGBT-attach solder joint is critical in active power load cycle. Substrate solder joint degraded most in passive thermal cum combined passive thermal and active power load cycles.

In this paper, we introduce two innovative two-stage low noise amplifiers (LNAs), each with distinct noise-matching networks. The first LNA features a low pass filter (LPF) for noise-matching in both stages, while the second uses a high pass filter (HPF) in a similar capacity. Our research focuses on evaluating the performance differences that arise from using varied matching networks within specific frequency ranges. Highlighting the critical role of appropriate network selection for optimizing gain and noise performance, our approach includes the development of two Monolithic Microwave Integrated Circuits (MMICs) using cutting-edge 0.25$\mu$m Gallium Nitride (GaN) technology. The C-band LNA, targeting a frequency range of 4–6 GHz, achieves an impressive average noise fig. (NF) of 1.5 dB and a gain of 17 dB. For the X-band range of 8–10 GHz, the LNA records a commendable average NF of 1.7 dB and a gain of 16 dB, demonstrating the effectiveness of our novel design strategies.

The development of flexible electronics, better heat dissipation capabilities, increased LED light extraction efficiency, and the implementation of inverted barrier N-polar high electron mobility transistor (HEMT) for power electronics are all made possible by adopting laser lift-off (LLO), a technology that enables the movement of discrete III-N elements onto any substrates which are otherwise not attainable. In this paper, we focus on evaluating the LLO mechanism, its application for III-N epilayers and devices, and assessing their structural and electronic characteristics to give an overview of the advancement in LLO technology for III-N microelectronics.

Developing flexible, stretchable, and self-healing wearable electronic devices with skin-like capabilities is highly desirable for healthcare and human-machine interaction. Hydrogels as a promising sensing material with crosslinked polymer networks have received widespread attention for decades. However, sensors based on hydrogels suffer from low sensitivity and stability due to their poor electrical conductivity or the movement of nanofillers in hydrogel networks. Herein, a stable, sensitive, and self-healing strain sensor is fabricated by the Ti₃C₂T_x MXene nanosheets/polyvinyl alcohol (PVA) hydrogel (T-hydrogel). The introduction of MXene increases the number of H-bonds in the PVA hydrogel network and enhances the conductivity, resulting in high sensitivity, stability, and self-healing character. The self-healing T-hydrogel-based strain sensor has a performance close to that of the original sensor. In addition, the device is capable of detecting bodily motions, indicating the potential application in the field of human health monitoring and human-computer interaction.