RESUMO
Neuromorphic computing, which mimics the structure and principles of the human brain, has the potential to facilitate the hardware implementation of next-generation artificial intelligence systems and process large amounts of data with very low power consumption. Among them, the XNOR synapse-based Binary Neural Network (BNN) has been attracting attention due to its compact neural network parameter size and low hardware cost. The previous XNOR synapse has drawbacks, such as a trade-off between cell density and accuracy. In this work, we show nonvolatile XNOR synapses with high density and accuracy using a monolithically stacked complementary ferroelectric field-effect transistor (C-FeFET) composed of a p-type Si MFMIS-FeFET at the bottom and a 3D stackable n-type Al:IZTO MFS-FeTFT, achieving 60F2 per cell (2C-FeFET). For adjusting the threshold voltage and improving the switching speed (100 ns) of n-type ferroelectric TFT, we employed a dual-gate configuration and a unique operation scheme, making it comparable to those of Si-based FeFETs. We performed array-level simulation with a 512 × 512 subarray size and a 3-bit flash ADC, demonstrating that the image recognition accuracies using the MNIST and CIFAR-10 data sets were increased by 3.17 and 14.07%, respectively, in comparison to other nonvolatile XNOR synapses. In addition, we performed system-level analysis on a 512 × 512 XNOR C-FeFET, exhibiting an outstanding throughput of 717.37 GOPS and an energy efficiency of 196.7 TOPS/W. We expect that our approach would contribute to the high-density memory systems, logic-in-memory technology, and hardware implementation of neural networks.
RESUMO
We present herewith a novel approach of equally thick AFE/FE (ZrO2/HZO) bilayer stack heterostructure films for achieving an equivalent oxide thickness (EOT) of 4.1 Å with a dielectric constant (κ) of 56 in complementary metal-oxide semiconductor (CMOS) compatible metal-ferroelectric-metal (MFM) capacitors using a high-pressure annealing (HPA) technique. The low EOT and high κ values were achieved by careful optimization of AFE/FE film thicknesses and HPA conditions near the morphotropic phase boundary (MPB) after field cycling effects. Stable leakage current density (J < 10-7 A/cm2 at ±0.8 V) was found at 3/3 nm bilayer stack films (κ = 56 and EOT = 4.1 Å) measured at room temperature. In comparison with previous work, our remarkable achievement stems from the interfacial coupling between FE and AFE films as well as a high-quality crystalline structure formed by HPA. Kinetically stabilized hafnia films result in a small grain size in bilayer films, leading to reducing the leakage current density. Further, a higher κ value of 59 and lower EOT of 3.4 Å were found at 333 K. However, stable leakage current density was found at 273 K with a high κ value of 53 and EOT of 3.85 Å with J < 10-7 A/cm2. This is the lowest recorded EOT employing hafnia and TiN electrodes that are compatible with CMOS, and it has important implications for future dynamic random access memory (DRAM) technology.