We present a SPICE model for ferroelectric transistors (FEFETs) based on time-dependent Landau-Khalatnikov equation solved self-consistently with the transistor equations. The model also considers depolarization fields due to non-ideal contacts. We experimentally characterize FE films to calibrate our model, based on which we analyze the device and circuit implications of FEFETs. We discuss the dependence of the ON current and gate capacitance of FEFETs on the FE thickness and FE material parameters. A ring oscillator analysis shows delay reduction up to 97% at iso-energy for FEFETs compared with MOSFETs at VDD <; 0.4 V. FEFET-based SRAMs show 47%-68% larger read stability and 50%-57% lower access time, albeit with an increase in the write time.

In this paper, we describe an improved SPICE model for the negative capacitance field-effect transistor (NCFET). According to the law of conservation of charge, the model is built based on the relationship between the gate charge of the MOSFET and the charge reserved by the ferroelectric layer and includes the parasitic resistance and capacitance into consideration. Based on the model, the drain-induced barrier lowering (DIBL) and the negative resistance (NR) of the NCFET are analyzed. Finally, taking two well-known analog blocks (the current mirror and the latch comparator) for example, impacts of the DIBL effect and the NR property on analog circuit performances are discussed. Thanks to utilization of the NR feature, not only can impacts of the DIBL effect and the channel-length modulation (CLM) effect be alleviated to improve the mirroring accuracy, but also the comparison speed of the latch comparator be accelerated.

The use of lower precision has emerged as a popular technique to optimize the compute and storage requirements of complex deep neural networks (DNNs). In the quest for lower precision, recent studies have shown that ternary DNNs (which represent weights and activations by signed ternary values) represent a promising sweet spot, achieving accuracy close to full-precision networks on complex tasks. We propose TiM-DNN, a programmable in-memory accelerator that is specifically designed to execute ternary DNNs. TiM-DNN supports various ternary representations including unweighted {-1,0,1}, symmetric weighted {-a,0,a}, and asymmetric weighted {-a,0,b} ternary systems. The building blocks of TiM-DNN are TiM tiles--specialized memory arrays that perform massively parallel signed ternary vector-matrix multiplications with a single access. TiM tiles are in turn composed of ternary processing cells (TPCs), bit-cells that function as both ternary storage units and signed ternary multiplication units. We evaluate an implementation of TiM-DNN in 32-nm technology using an architectural simulator calibrated with SPICE simulations and RTL synthesis. We evaluate TiM-DNN across a suite of state-of-the-art DNN benchmarks including both deep convolutional and recurrent neural networks. A 32-tile instance of TiM-DNN achieves a peak performance of 114 TOPs/s, consumes 0.9-W power, and occupies 1.96 mm² chip area, representing a 300x and 388x improvement in TOPS/W and TOPS/mm², respectively, compared to an NVIDIA Tesla V100 GPU. In comparison to specialized DNN accelerators, TiM-DNN achieves 55x-240x and 160x-291x improvement in TOPS/W and TOPS/mm², respectively. Finally, when compared to a well-optimized near-memory accelerator for ternary DNNs, TiM-DNN demonstrates 3.9x-4.7x improvement in system-level energy and 3.2x-4.2x speedup, underscoring the potential of in-memory computing for ternary DNNs.

Negative-capacitance FETs (NCFETs) are a promising candidate for low-power circuits with intrinsic features, e.g., the steep switching slope. Prior works have shown potential for enabling low-power digital logic and memory design with NCFETs. Yet, it is still not quite clear how to harness these new features of NCFETs for analog functionalities. This article provides more insights into the circuit design space with new device characteristics and investigates its deployment in analog circuits, specifically, time-domain analog-to-digital converters (ADCs) and phase-locked loops (PLLs). We propose and optimize a novel digital-based clocked comparator and a capacitor-based voltage-to-time converter (VTC), which are essential building blocks in ADCs and PLLs. Evaluation results show beyond-FinFET comparison speed and enhanced linearity for the proposed NCFET-based clocked comparator and VTC, respectively. Such improvement is achieved by exploiting the steeper slope and increased output impedance of NCFETs. More details on design details and a discussion are provided in this article.

In this work, we demonstrate a CMOS static random-access-memory (SRAM) using WSe ₂ as a channel material for the first time, providing comprehensive DC analyses for transition metal dichalcogenide (TMD) material-based memory applications. A tri-gate design is adopted for the n-type MOSFET, while an air-stable, oxygen plasma induced doping scheme is introduced to implement the p-type MOSFET. DC measurements of SRAM cells demonstrate a unique dynamic tunability enabled by modulating the n-FET doping level through electrostatically gating the extended source/drain regions. Furthermore, with various read/write assist techniques, SRAM operation at low VDD of 0.8V is achieved. Our low power demonstration and its 2D ultra-thin material nature suggest promising applications of WSe ₂ for flexible electronics and Internet of Things (IoT).