Power consumption has become one of the bottlenecks limiting the future development of integrated circuits. Tunnel FETs(TFETs) and negative capacitance FETs(NCFETs) can break the subthreshold swing limitation(60 mV/de...
详细信息
Power consumption has become one of the bottlenecks limiting the future development of integrated circuits. Tunnel FETs(TFETs) and negative capacitance FETs(NCFETs) can break the subthreshold swing limitation(60 mV/dec at room temperature) of conventional metal-oxide-semiconductor field-effect transistor(MOSFET) to reduce the operating voltage and thus power consumption. However, induced by the band-to-band tunneling mechanism, TFETs have a subthreshold swing degradation issue and relatively low ON current. Although NCFETs with ferroelectric/dielectric gate stack can theoretically maintain a high ON current comparable to conventional MOSFET, the physical origin of sub-60 SS is controversial and the mechanism of switching behavior in NCFET is still not clear. In this work, by experimentally investigating the whole negative differential capacitance process and its gate voltage amplification coefficient, an intrinsic issue of SS degradation with increased gate voltage is also found in NCFET for the first time. Based on the physical investigation and simulation results, it is shown that the intrinsic SS degradation in NCFET is resulting from the instant dielectric polarization response. Both the decrease of dielectric thickness and the increase of dielectric constant may lead to the severer SS degradation, which is not favorable for scaled NCFETs.
Dear editor,Recently, ferroelectric(FE)-based negative capacitance FET(NCFET) with ferroelectric/dielectric(FE/DE) gate stack has attracted extensive attention due to its capability of sub-60 mV/dec subthreshold swing...
详细信息
Dear editor,Recently, ferroelectric(FE)-based negative capacitance FET(NCFET) with ferroelectric/dielectric(FE/DE) gate stack has attracted extensive attention due to its capability of sub-60 mV/dec subthreshold swing(SS) at room temperature while maintaining a high on-state current compared with MOSFET [1].
Dear editor,Strain technology has become a common solution in the semiconductor manufacturing industry since 90-nm technology node to overcome the severe carrier mobility degradation of nanoscale microelectronic devic...
详细信息
Dear editor,Strain technology has become a common solution in the semiconductor manufacturing industry since 90-nm technology node to overcome the severe carrier mobility degradation of nanoscale microelectronicdevices [1]. Since layout may affect the stress distribution, layout dependent effect(LDE)becomes a serious issue in advanced technology nodes.
Artificial intelligence(AI) has experienced substantial advancements recently, notably with the advent of large-scale language models(LLMs) employing mixture-of-experts(MoE) techniques, exhibiting human-like cognitive...
详细信息
Artificial intelligence(AI) has experienced substantial advancements recently, notably with the advent of large-scale language models(LLMs) employing mixture-of-experts(MoE) techniques, exhibiting human-like cognitive skills. As a promising hardware solution for edge MoE implementations, the computing-in-memory(CIM) architecture collocates memory and computing within a single device, significantly reducing the data movement and the associated energy consumption. However, due to diverse edge application scenarios and constraints, determining the optimal network structures for MoE, such as the expert's location,quantity, and dimension on CIM systems remains elusive. To this end, we introduce a software-hardware co-designed neural architecture search(NAS) framework, CIM-based MoE NAS(CMN), focusing on identifying a high-performing MoE structure under specific hardware constraints. The results of the NYUD-v2 dataset segmentation on the RRAM(SRAM) CIM system reveal that CMN can discover optimized MoE configurations under energy, latency, and performance constraints, achieving 29.67×(43.10×) energy savings,175.44×(109.89×) speedup, and 12.24× smaller model size compared to the baseline MoE-enabled Visual Transformer, respectively. This co-design opens up an avenue toward high-performance MoE deployments in edge CIM systems.
EEPROM is an important part for interface circuit of sensor. It saves the calibration data and parameter setting data by non-volatile storage. A new low- power Erasable and electrically programmable read only memory (...
详细信息
EEPROM is an important part for interface circuit of sensor. It saves the calibration data and parameter setting data by non-volatile storage. A new low- power Erasable and electrically programmable read only memory (EEPROM) circuit for sensor interface circuit is introduced in this paper. The data stored in EEPROM can be reloaded automatically by the power-on-reset technique when the power is on. The module consumes power only during power-on-reset and data changing by status optimization and low-power design. This EEPROM circuit has been used in a MEMS accelerometer readout circuit verified by 0.35.μm CMOS EEPROM process. The results are satisfying in that the module can write and store 25bit data and automatically reloads data within 0.1.μs after power-on and then turns to low power mode.
Dear editor,Image sensors have been rapidly developed for decades and widely used in many different fields [1, 2]. To achieve high resolution, the pixel size has been scaled down to 1 μm for mass production [3]. Neve...
详细信息
Dear editor,Image sensors have been rapidly developed for decades and widely used in many different fields [1, 2]. To achieve high resolution, the pixel size has been scaled down to 1 μm for mass production [3]. Nevertheless, the pixel size has to be further reduced to meet the growing demand for higher resolution. In CMOS image sensor(CIS), the pixel is composed of a photodiode, reset transistor,
This paper presents a pixel circuit with high frame rate, high linearity, and low power consumption. This pixel circuit is applied in a digital readout integrated circuit (DROIC) of 320×256 infrared focal plane a...
详细信息
This paper presents a low power readout integrated circuit (ROIC) with 12-bit two-stage column-wise ADCs for 25μm-pitch 640×480 silicon diode uncooled infrared imagers. A novel two-stage column-wise ADC composed...
详细信息
A low-power and high-dynamic range ADC-based 320×256 size digital readout integrated circuit for infrared focal plane arrays is proposed in this paper. Compared with the traditional structure, the proposed pixel-...
详细信息
With the development of power integrated circuits, LDMOS devices assume a significant role in power device applications due to their elevated input impedance and enhanced conversion rate. Nonetheless, a contradictory ...
详细信息
暂无评论