Recently, transistor scaling is approaching its physical limit, hindering the further development of the computing capability. In the post-Moore era, emerging logic and storage devices have been the fundamental hardware for expanding the capability of intelligent computing. In this article, the recent progress of ferroelectric devices for intelligent computing is reviewed. The material properties and electrical characteristics of ferroelectric devices are elucidated, followed by a discussion of novel ferroelectric materials and devices that can be used for intelligent computing. Ferroelectric capacitors, transistors, and tunneling junction devices used for low-power logic, high-performance memory, and neuromorphic applications are comprehensively reviewed and compared. In addition, to provide useful guidance for developing high-performance ferroelectric-based intelligent computing systems, the key challenges for realizing ultrascaled ferroelectric devices for high-efficiency computing are discussed.

/pdf/ferroelectric-devices-for-intelligent-computing-17ahh47k.pdf

Ferroelectric Devices for Intelligent Computing

HfO2-ZrO2 superlattice (SL) ferroelectric (FE) ultrathin films exhibit significant improvement in endurance performance compared with solid-solution HfxZr1−xO2 (HZO). Despite the experimental evidence, the underlying microscopic mechanisms of the enhanced reliability of SL remain elusive. This Letter explores the mechanism by performing first-principle calculations on SL and HZO systems. The enhanced endurance in the SL can be well explained by higher oxygen vacancy (Vo) migration energy barriers along the FE polarization direction, which slow down the increase in Vo. The suppression of Vo increase will potentially help maintain the stability of the FE phase and alleviate the fatigue. Based on this mechanism, we suggest that doping materials with higher Vo migration barriers can further improve the endurance of HfO2-based FE devices. This work facilitates the future development of HfO2-based FE devices with enhanced endurance and reliability.

Physical origin of the endurance improvement for HfO2-ZrO2 superlattice ferroelectric film

Modern microelectronic systems and applications demand an every increasing amount of non-volatile memories that are fast, reliable, and consume little power. Memory concepts based on ferroelectric HfO2 like the ferroelectric field effect transistor (FeFET) and the ferroelectric random access memory (FeRAM) are promising to satisfy these requirements. As a consequence, continuing high attention is given to improve the ferroelectric properties and the reliability characteristics of the ferroelectric HfO2 films –- for instance by using different dopant elements, dopant concentrations, and film thicknesses. Superlattices (i.e., a periodic structure of two materials stacked upon each other) are a promising alternative approach. Herein, [HfO2/ZrO2] superlattices of various sublayer thicknesses and a constant total thickness of 10 nm are embedded into metal-ferroelectric-metal (MFM) capacitors and then electrically as well as structurally characterized with special focus on remanent polarization, coercive field, endurance, and high temperature reliability. Compared to a 10 nm (Hf,Zr)O2 solid solution reference film, the use of superlattice stacks significantly improves the above mentioned parameters. In addition, most of these parameters depend on the sublayer thickness, which allows, for instance, tailoring the coercive field of the whole device. 

https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/apxr.202200108

Ferroelectric [HfO
 2
 /ZrO
 2
 ] Superlattices with Enhanced Polarization, Tailored Coercive Field, and Improved High Temperature Reliability

Antiferroelectric random access memory (AFERAM) is one of the newest alternative non-volatile memory technologies to emerge in recent years. ZrO 2 -based antiferroelectric films are exceptionally well-suited for memory applications with very high cycling endurance (>10 10 ) and low operating voltages (< 2 V). Lightly alloying ZrO 2 with HfO 2 is performed to assess AFERAM device performance with back-end-of-line compatible thin film Zr 1−x Hf x O 2 (x <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\le0.13$ </tex-math></inline-formula> ) capacitors. The transition fields associated with antiferroelectric behavior are reduced with more Hf incorporation, yielding a larger magnitude switching polarization and memory window. Cycling endurance beyond 10 10 cycles is conducted on thin film capacitors where wake-up in AFERAM first leads to an increase, then a decrease in the memory window at a cumulative cycle number found to be dependent on the amount of Hf-incorporation. Hf-incorporation into ZrO 2 is demonstrated to be a feasible way to improve the memory window in ZrO 2 -based AFERAM.

Memory Window Enhancement in Antiferroelectric RAM by Hf Doping in ZrO₂

The ferroelectric polarization stability and dielectric characteristics of ZrO 2 -HfO 2 superlattice (SL) ferroelectric layer Metal-Ferroelectric-Metal (MFM) capacitor fabricated with various post-metal annealing (PMA) temperatures were presented. It was demonstrated that ZrO 2 -HfO 2 SL MFM capacitors with increased post metallization PMA temperature (i.e., 600 °C) showed improved polarization stability with almost <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">wake-up free and polarization <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">fatigue free characteristics. In this work, we have shown that highly stable remanent polarization (2P r @ 3MV <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$/$ </tex-math></inline-formula> cm) of ~ <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$30 ~\mu \text{C}/$ </tex-math></inline-formula> cm 2 with small variation ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\Delta 2\text{P}_{r} \le 2 ~\mu \text{C}/$ </tex-math></inline-formula> cm 2 ; <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\Delta 2\text{P}_{r} / 2\text{P}_{r,pristine} \le9$ </tex-math></inline-formula> %) up to 10 11 cycles of field cycling were exhibited by the atomic layer deposition (ALD) deposited ZrO 2 -HfO 2 superlattice with 600 °C PMA. To our knowledge, the nearly “ <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">wake-up”-free and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">fatigue-free polarization behavior up to 10 11 cycles are among the best polarization stability for the hafnium zirconium oxide based MFM capacitors sustained endurance tests for more than 10 10 cycles.

ZrO2-HfO2 Superlattice Ferroelectric Capacitors With Optimized Annealing to Achieve Extremely High Polarization Stability

HfO2-ZrO2 superlattice (SL) ferroelectric (FE) capacitor is demonstrated to have improved endurance performance and higher fatigue recovery capability compared to the HfZrO<italic>x</italic> (HZO) device. During the cycling of polarization (<inline-formula> <tex-math notation="LaTeX">${P}$ </tex-math></inline-formula>) <italic>vs.</italic> voltage (<inline-formula> <tex-math notation="LaTeX">${V}$ </tex-math></inline-formula>) loops, the SL metal-FE-metal (MFM) capacitor exhibits the higher <inline-formula> <tex-math notation="LaTeX">${P}$ </tex-math></inline-formula> and the lower leakage current over the HZO device indicating the lower defect density in SL. The SL capacitor achieves an endurance of <inline-formula> <tex-math notation="LaTeX">${5}\times {10} ^{{12}}$ </tex-math></inline-formula> cycles, which is three orders of magnitude higher than the HZO device. The <inline-formula> <tex-math notation="LaTeX">${P}$ </tex-math></inline-formula> fatigue of the SL capacitor can be fully recovered through a ~30 s break, and that of HZO is only partially recovered utilizing the higher field cycling. This is because the trapping/detrapping process significantly decreases in HfO2-ZrO2 SL over HZO capacitor by the reduced defect density. These results prove that the HfO2-ZrO2 SL is a promising technology for endurance unlimited FE random access memory.

HfO2-ZrO2 Superlattice Ferroelectric Capacitor With Improved Endurance Performance and Higher Fatigue Recovery Capability

In this article, an analog synapse based on a nonvolatile field-effect transistor with amorphous ZrO2 dielectrics has been fabricated and demonstrated. The conductance modulation properties of the devices have been systematically evaluated. Due to the polarization switching dynamics of the ferroelectric-like amorphous ZrO2 dielectric, which is attributable to the voltage-driven oxygen vacancies and negative charges dipoles, the proposed ZrO2-based devices exhibit superior synaptic characteristics, including good symmetry and linearity for both potentiation and depression, with small cycle-to-cycle variations. The ratio of maximum conductance and minimum conductance ( ${G}_{max}/{G}_{min})$ of devices reaches 130, with conductance states over 30. Also, spike-timing-dependent plasticity (STDP) has been mimicked in the devices successfully. Furthermore, based on the experimental STDP characteristics and conductance modulation properties of potentiation and depression, a spiking neural network architecture constructed by the proposed ZrO2-based synapses has been simulated. High offline and online learning accuracy of 94% and 87%, respectively, on the handwritten digits dataset, has been achieved.

Xiao Yu

Papers

HfO2-ZrO2 Superlattice Ferroelectric Capacitor With Improved Endurance Performance and Higher Fatigue Recovery Capability

Analog Synapses Based on Nonvolatile FETs With Amorphous ZrO2 Dielectric for Spiking Neural Network Applications

Ferroelectric Devices for Intelligent Computing

Physical origin of the endurance improvement for HfO2-ZrO2 superlattice ferroelectric film

Hf0.5Zr0.5O2 1T−1C FeRAM arrays with excellent endurance performance for embedded memory