Search for: All records

The continued miniaturization of nanoelectronic devices approaches its fundamental physical limits due to power dissipation. Negative capacitance field-effect transistors using ferroelectric gate insulators are promising to overcome these limits, which would allow further device scaling. However, the microscopic details of negative capacitance are not well understood so far, since mainly Landau based mean-field theories are used to model these phenomena. Here we use an educational and simplified approach to better understand the basic microscopic origin of ferroelectric negative capacitance. Our “toy” model shows that negative capacitance originates from the thermodynamic instability of the ferroelectric polarization and is bounded by the saturation of microscopic dipole polarizability. This shows that negative capacitance is strongly connected to the origin of ferroelectricity itself. Furthermore, our microscopic model results in the same qualitative behavior as mean-field Landau based approaches. 

2019 marks the 11 th year since the concept of ferroelectric negative capacitance was first proposed [1]. It was proposed as a physics solution to an engineering problem: The power dissipation in electronics and computing. Yet, it was unique in that the technology required a fundamental scientific discovery in a field that was ~90 years old at that time-the discovery of negative capacitance or static negative permittivity in ferroelectrics. Initially, interests into negative capacitance were limited to the device researchers; over time, the topic brought together the “often-disjoint” communities of device engineers, condensed matter physicists and material scientists in exploring, understanding and finally discovering this phenomenon [2]-[7]. Different manifestations of the negative capacitance phenomenon, namely, capacitance enhancement, negative differential relation between charge and voltage as well as atomic scale mapping of the stabilized, negative capacitance states corroborated with phase field and density functional theory based calculations have established this concept on a solid ground. 

We present a simple, physical explanation of underlying microscopic mechanisms that lead to the emergence of the negative phenomena in ferroelectric materials. The material presented herein is inspired by the pedagogical treatment of ferroelectricity by Feynman and Kittel. In a toy model consisting of a linear one-dimensional chain of polarizable units (i.e., atoms or unit cells of a crystal structure), we show how simple electrostatic interactions can create a microscopic, positive feedback action that leads to negative capacitance phenomena. We point out that the unstable negative capacitance effect has its origin in the so called “polarization catastrophe” phenomenon which is essential to explain displacement type ferroelectrics. Furthermore, the fact that even in the negative capacitance state, the individual dipole always aligns along the direction of the local electrical field not opposite is made clear through the toy model. Finally, how the “ S”-shaped polarization vs. applied electric field curve emerges out of the electrostatic interactions in an ordered set of polarizable units is shown. 

We report the first experimental demonstration of ferroelectric field-effect transistor (FEFET) based spiking neurons. A unique feature of the ferroelectric (FE) neuron demonstrated herein is the availability of both excitatory and inhibitory input connections in the compact 1T-1FEFET structure, which is also reported for the first time for any neuron implementations. Such dual neuron functionality is a key requirement for bio-mimetic neural networks and represents a breakthrough for implementation of the third generation spiking neural networks (SNNs)-also reported herein for unsupervised learning and clustering on real world data for the first time. The key to our demonstration is the careful design of two important device level features: (1) abrupt hysteretic transitions of the FEFET with no stable states therein, and (2) the dynamic tunability of the FEFET hysteresis by bias conditions which allows for the inhibition functionality. Experimentally calibrated, multi-domain Preisach based FEFET models were used to accurately simulate the FE neurons and project their performance at scaled nodes. We also implement an SNN for unsupervised clustering and benchmark the network performance across analog CMOS and emerging technologies and observe (1) unification of excitatory and inhibitory neural connections, (2) STDP based learning, (3) lowest reported power (3.6nW) during classification, and (4) a classification accuracy of 93%.