The recent discovery of ferroelectricity in thin hafnium oxide films has led to a resurgence of interest in ferroelectric memory devices. Although both experimental and theoretical studies on this new ferroelectric system have been undertaken, much remains to be unveiled regarding its domain landscape and switching kinetics. Here we demonstrate that the switching of single domains can be directly observed in ultrascaled ferroelectric field effect transistors. Using models of ferroelectric domain nucleation we explain the time, field and temperature dependence of polarization reversal. A simple stochastic model is proposed as well, relating nucleation processes to the observed statistical switching behavior. Our results suggest novel opportunities for hafnium oxide based ferroelectrics in nonvolatile memory devices.

With the discovery of ferroelectric hafnium oxide (FE-HfO2), the ferroelectric field effect transistor (FeFET), a long-term contender for non-volatile data storage, has finally managed to scale to the 2x nm technology node. Here for the first time, we correlate the thickness dependent ferroelectric properties of Si:HfO2 with the memory characteristics of small (56bit) FeFET arrays. First, an electrical and structural analysis of metal-ferroelectric-metal capacitors is given. Even though possessing room-temperature deposited top electrodes, TiN / Si:HfO2 (20nm) / TiN capacitors are showing deteriorated polarization characteristics as compared to their 10nm Si:HfO2 counterparts. This could be attributed to an increased monoclinic phase fraction, as indicated by small-signal capacitance voltage and grazing incidence X-ray diffraction measurements. Identical Si:HfO2 thin films with thicknesses of 10nm and 20nm respectively, were utilized in a 28nm high-k metal-gate CMOS flow to form small FeFET memory arrays of AND architecture. After extracting the most suitable operating conditions from erase matrix, single cell evaluation was performed by standard V-P/3 program and a novel V-P/3 positive-source drain erase scheme. Array cells incorporating 10nm Si:HfO2 films showed a maximum memory window of 1.03 V whereas cells incorporating 20nm Si:HfO2 films could reach up to 1.57V. Moreover, in accordance to the basic material properties, the previously observed increased monoclinic phase fraction in 20nm Si:HfO2 thin films correlate well with a reduced number of functional FeFET cells.

System on chip (SoC) embedded memory solutions promise small form factors and high operating speed, as well as a high energy and cost efficiency. Single cell scalability and basic memory parameters such as data retention, cycling endurance and disturb characteristics on array level are important aspects in stand-alone memory development and can serve as a guideline for embedded solutions. However, one of the key aspects in embedded memory development is compatibility of the memory technology to its underlying complementary metal oxide semiconductor (CMOS) platform. This includes the voltage requirements for logic and memory operation, the need for additional lithographic steps and minimally CMOS invasive integration efforts, as well as the introduction of new materials and related contamination concerns. Especially in state of the art high-k metal gate (HKMG) CMOS technologies at minimum feature size (F) these aspects proof rather challenging when searching for a suitable embedded memory solution. As a consequence most approaches result in large memory cells of multiple F2 or leave a BEoL integration of the memory cell as the only viable option. With the introduction of ferroelectric hafnium oxide, however, a scalable one-transistor (1T) memory solution derived from the conventional HKMG transistor was presented for the 2X nm node. The therewith close resemblance of the memory and logic transistor appears ideally suited for combining nonvolatile data storage and logic circuitry on the same chip. Nevertheless, in order to fulfill these expectations and to ease manufacturing issues this resemblance has to be as close as possible. In the context of a minimally invasive memory integration strategy this means that ideally the ferroelectric hafnium oxide based memory transistor has to adapt to the HKMG transistor in terms of thermal budget and post treatments, vertical and lateral dimensions, the use of stress engineering, as well as metal gate and work function engineering. Based on experimental gate first transistor and metal insulator metal (MIM) capacitor data these aspects together with embedded memory requirements will be analyzed and critically discussed.

Ferroelectrics are very interesting materials for nonvolatile data storage due to the fact that they deliver very low power programming operation combined with nonvolatile retention. For 60 years researchers have been inspired by these fascinating possibilities and have tried to build ferroelectric memory devices that can compete with mainstream technologies in their respective time. The progress of the current concepts is limited by the low compatibility of ferroelectrics like PZT with CMOS processing. Therefore, PZT or SBT based 1T1C ferroelectric memories are not scaling below 130 nm and 1T ferroelectric FETs based on the same materials are still struggling with low retention and very thick memory stacks. Hafnium oxide, a standard material in sub 45 nm CMOS, can show ferroelectric hysteresis with promising characteristics. By adding a few percent of silicon and annealing the films in a mechanically confined manner. Boescke et al. demonstrated ferroelectric hysteresis in hafnium oxide for the first time. Recently, a large number of dopants including Y, Al, Gd and Sr have been used to induce ferroelectricity in HfO2. This paper reviews the current status of hafnium oxide based ferroelectrics, its application to field effect transistors and puts this approach into a wider context of earlier developments in the field.