Novel Two-Terminal Synapse/Neuron Based on an Antiferroelectric Hafnium Zirconium Oxide Device for Neuromorphic Computing.

Functionally diverse devices with artificial neuron and synapse properties are critical for neuromorphic systems. We present a two-terminal artificial leaky-integrate-fire (LIF) neuron based on 6 nm Hf0.1Zr0.9O2 (HZO) antiferroelectric (AFE) thin films and develop a synaptic device through work function (WF) engineering. LIF neuron characteristics, including integration, firing, and leakage, are achieved in W/HZO/W devices due to the accumulated polarization and spontaneous depolarization of AFE HZO films. By engineering the top electrode with asymmetric WFs, we found that Au/Ti/HZO/W devices exhibit synaptic weight plasticity, such as paired-pulse facilitation and long-term potentiation/depression, achieving >90% accuracy in digit recognition within constructed artificial neural network systems. These findings suggest that AFE HZO capacitor-based neurons and WF-engineered artificial synapses hold promise for constructing efficient spiking neuron networks and artificial neural networks, thereby advancing neuromorphic computing applications based on emerging AFE HZO devices.

Impact of Chamber/Annealing Temperature on the Endurance Characteristic of Zr:HfO2 Ferroelectric Capacitor.

The endurance characteristic of Zr-doped HfO2 (HZO)-based metal-ferroelectric-metal (MFM) capacitors fabricated under various deposition/annealing temperatures in the atomic layer deposition (ALD) process was investigated. The chamber temperature in the ALD process was set to 120 °C, 200 °C, or 250 °C, and the annealing temperature was set to 400 °C, 500 °C, 600 °C, or 700 °C. For the given annealing temperature of 700 °C, the remnant polarization (2Pr) was 17.21 µC/cm2, 26.37 µC/cm2, and 31.8 µC/cm2 at the chamber temperatures of 120 °C, 200 °C, and 250 °C, respectively. For the given/identical annealing temperature, the largest remnant polarization (Pr) was achieved when using the chamber temperature of 250 °C. At a higher annealing temperature, the grain size in the HZO layer becomes smaller, and thereby, it enables to boost up Pr. It was observed that the endurance characteristics for the capacitors fabricated under various annealing/chamber temperatures were quite different. The different endurance characteristics are due to the oxygen and oxygen vacancies in ferroelectric films, which affects the wakeup/fatigue behaviors. However, in common, all the capacitors showed no breakdown for an externally applied pulse (up to 108 cycles of the pulse).

High-Performance Ferroelectric Thin Film Transistors with Large Memory Window Using Epitaxial Yttrium-Doped Hafnium Zirconium Gate Oxide.

Preventing ferroelectric materials from losing their ferroelectricity over a low thickness of several nanometers is crucial in developing multifunctional nanoelectronics. Epitaxially grown 5 at. % yttrium-doped Hf0.5Zr0.5O2 (YHZO) thin films exhibit an atomically smooth surface, an ability to maintain ferroelectricity even at a thickness of 10 nm, and excellent insulating properties, making them suitable for use as gate oxides in ferroelectric thin film transistors (FeTFTs). Through the epitaxial growth of a YHZO/La0.67Sr0.33MnO3 (LSMO)/SrTiO3 (STO) heterostructure, YHZO effectively retains its ferroelectricity and orthorhombic single phase, leading to enhancing electron mobility (∼19.74 cm2 V-1 s-1) and memory window (3.7 V) in the amorphous InGaZnO4 (a-IGZO)/YHZO/LSMO/STO FeTFTs. These FeTFTs demonstrate a consistent memory function with remarkable endurance (∼106 cycles) and retention (∼104 s). Furthermore, they sustain a constant memory window even under ±6 V bias stress for 104 s and exhibit excellent stability even under ±6 V/1 ms pulse cycling for 107 cycles. For comparison, a transistor with the same structure was fabricated using epitaxial nonferroelectric LaAlO3 (LAO) and epitaxial undoped Hf0.5Zr0.5O2 (HZO) as alternatives to YHZO. This study presents a novel approach to exploit the potential of YHZO in FeTFTs, contributing to the development of next-generation logic-in-memory.

Mapping Ferroelectric Fields Reveals the Origins of the Coercivity Distribution.

Better techniques for imaging ferroelectric polarization would aid the development of new ferroelectrics and the refinement of old ones. Here we show how scanning transmission electron microscope (STEM) electron beam-induced current (EBIC) imaging reveals ferroelectric polarization with obvious, simply interpretable contrast. Planar imaging of an entire ferroelectric hafnium zirconium oxide (Hf0.5Zr0.5O2, HZO) capacitor shows an EBIC response that is linearly related to the polarization determined in situ with the positive-up, negative-down (PUND) method. The contrast is easily calibrated in MV/cm. The underlying mechanism is magnification-independent, operating equally well on micrometer-sized devices and individual nanoscale domains. Coercive-field mapping reveals that individual domains are biased "positive" and "negative", as opposed to being "easy" and "hard" to switch. The remanent background E-fields generating this bias can be isolated and mapped. Coupled with STEM's native capabilities for structural identification, STEM EBIC imaging provides a revolutionary tool for characterizing ferroelectric materials and devices.

Ferroelectric Al0.85Sc0.15N and Hf0.5Zr0.5O2 Domain Switching Dynamics.

The capability to reliably program partial polarization states with nanosecond programming speed and femtojoule energies per bit in ferroelectrics makes them an ideal candidate to realize multibit memory elements for high-density crossbar arrays, which could enable neural network models with a large number of parameters at the edge. However, a thorough understanding of the domain switching dynamics involved in the polarization reversal is required to achieve full control of the multibit capability. Transient current integration measurements are adopted to investigate the domain dynamics in aluminum scandium nitride (Al0.85Sc0.15N) and hafnium zirconium oxide (Hf0.5Zr0.5O2). The switching dynamics are correlated to the crystal structure of the films. The contributions of domain nucleation and domain wall motion are decoupled by analyzing the rate of change of the time-dependent normalized switched polarization. Thermally activated creep domain wall motion characterizes the Al0.85Sc0.15N switching dynamics. The statistics of independently nucleating domains and the domain wall creep motion in Hf0.5Zr0.5O2 are associated with the spatially inhomogeneous distribution of local switching field due to polymorphism, absence of preferential crystallite orientation, as well as defects and charges that can be located at the grain boundaries. The c-axis texture, single-phase nature, and strong likelihood of less fabrication process-induced defects contribute to the homogeneity of the local switching field in Al0.85Sc0.15N. Nonetheless, defects generated and redistributed upon bipolar electric field switching cycling result in Al0.85Sc0.15N domain wall pinning. The wake-up effect in Hf0.5Zr0.5O2 is explained thorough the continuous addition of switchable regions associated with two independent distributions of characteristic switching times.

Nanoscale Phase and Orientation Mapping in Multiphase Polycrystalline Hafnium Zirconium Oxide Thin Films Using 4D-STEM and Automated Diffraction Indexing.

Ferroelectric hafnium zirconium oxide (HZO) holds promise for nextgeneration memory and transistors due to its superior scalability and seamless integration with complementary metal-oxide-semiconductor processing. A major challenge in developing this emerging ferroelectric material is the metastable nature of the non-centrosymmetric polar phase responsible for ferroelectricity, resulting in a coexistence of both polar and non-polar phases with uneven grain sizes and random orientations. Due to the structural similarity between the multiple phases and the nanoscale dimensions of the thin film devices, accurate measurement of phase-specific information remains challenging. Here, the application of 4D scanning transmission electron microscopy is demonstrated with automated electron diffraction pattern indexing to analyze multiphase polycrystalline HZO thin films, enabling the characterization of crystallographic phase and orientation across large working areas on the order of hundreds of nanometers. This approach offers a powerful characterization framework to produce a quantitative and statistically robust analysis of the intricate structure of HZO films by uncovering phase composition, polarization axis alignment, and unique phase distribution within the HZO film. This study introduces a novel approach for analyzing ferroelectric HZO, facilitating reliable characterization of process-structure-property relationships imperative to accelerating the growth optimization, performance, and successful implementation of ferroelectric HZO in devices.

Improvement of the Ferroelectric Response of La-Doped Hafnium Zirconium Oxide Employing Tungsten Oxide Interfacial Layer with Back-End-of-Line Compatibility.

In this work, the impact of a tungsten oxide (WO3) seed and capping layer for ferroelectric La-doped (Hf, Zr)O2 (La:HZO) based capacitors, designed with back-end-of-line (BEOL) compatibility, is systematically investigated. The WO3 capping layer supplies oxygen to the La:HZO layer throughout the fabrication process and during device cycling. This facilitates the annihilation of oxygen vacancies (Vo) within the La:HZO layer, thereby stabilizing its ferroelectric orthorhombic phase and resulting in an increase of the remanent polarization (Pr) value in the capacitor. Moreover, the effectiveness of the WO3 capping layer depends on the seed layer of the HZO film, suggesting that proper combination of the seed and capping layers should be employed to maximize the ferroelectric response. Finally, a TiN/TiO2 seed layer/La:HZO/WO3 capping layer/TiN capacitor is successfully fabricated and optimized by a complete set of atomic layer deposition (ALD) processes, achieving a superior 2Pr value and endurance value of more than 109 cycles at an electric field of 2.5 MV/cm. The WO3 capping layer is anticipated to offer a viable solution for doped HZO capacitors with reduced thickness, addressing the challenge of elevated Vo levels that favor the tetragonal phase and result in low 2Pr values.

Low-Temperature Solution-Processed HfZrO Gate Insulator for High-Performance of Flexible LaZnO Thin-Film Transistor.

Metal-oxide-semiconductor (MOS)-based thin-film transistors (TFTs) are gaining significant attention in the field of flexible electronics due to their desirable electrical properties, such as high field-effect mobility (µFE), lower IOFF, and excellent stability under bias stress. TFTs have widespread applications, such as printed electronics, flexible displays, smart cards, image sensors, virtual reality (VR) and augmented reality (AR), and the Internet of Things (IoT) devices. In this study, we approach using a low-temperature solution-processed hafnium zirconium oxide (HfZrOx) gate insulator (GI) to improve the performance of lanthanum zinc oxide (LaZnO) TFTs. For the optimization of HfZrO GI, HfZrO films were annealed at 200, 250, and 300 °C. The optimized HfZrO-250 °C GI-based LaZnO TFT shows the µFE of 19.06 cm2V-1s-1, threshold voltage (VTH) of 1.98 V, hysteresis voltage (VH) of 0 V, subthreshold swing (SS) of 256 mV/dec, and ION/IOFF of ~108. The flexible LaZnO TFT with HfZrO-250 °C GI exhibits negligible ΔVTH of 0.25 V under positive-bias-temperature stress (PBTS). The flexible hysteresis-free LaZnO TFTs with HfZrO-250 °C can be widely used for flexible electronics. These enhancements were attributed to the smooth surface morphology and reduced defect density achieved with the HfZrO gate insulator. Therefore, the HfZrO/LaZnO approach holds great promise for next-generation MOS TFTs for flexible electronics.

A MoS2 Hafnium Oxide Based Ferroelectric Encoder for Temporal-Efficient Spiking Neural Network.

Spiking neural network (SNN), where the information is evaluated recurrently through spikes, has manifested significant promises to minimize the energy expenditure in data-intensive machine learning and artificial intelligence. Among these applications, the artificial neural encoders are essential to convert the external stimuli to a spiking format that can be subsequently fed to the neural network. Here, a molybdenum disulfide (MoS2 ) hafnium oxide-based ferroelectric encoder is demonstrated for temporal-efficient information processing in SNN. The fast domain switching attribute associated with the polycrystalline nature of hafnium oxide-based ferroelectric material is exploited for spike encoding, rendering it suitable for realizing biomimetic encoders. Accordingly, a high-performance ferroelectric encoder is achieved, featuring a superior switching efficiency, negligible charge trapping effect, and robust ferroelectric response, which successfully enable a broad dynamic range. Furthermore, an SNN is simulated to verify the precision of the encoded information, in which an average inference accuracy of 95.14% can be achieved, using the Modified National Insitute of Standards and Technology (MNIST) dataset for digit classification. Moreover, this ferroelectric encoder manifests prominent resilience against noise injection with an overall prediction accuracy of 94.73% under various Gaussian noise levels, showing practical promises to reduce the computational load for the neural network.

Trench FinFET Nanostructure with Advanced Ferroelectric Nanomaterial HfZrO2 for Sub-60-mV/Decade Subthreshold Slope for Low Power Application.

Ferroelectric fin field-effect transistors with a trench structure (trench Fe-FinFETs) were fabricated and characterized. The inclusion of the trench structures improved the electrical characteristics of the Fe-FinFETs. Moreover, short channel effects were suppressed by completely surrounding the trench channel with the gate electrodes. Compared with a conventional Fe-FinFET, the fabricated trench Fe-FinFET had a higher on-off current ratio of 4.1 × 107 and a steep minimum subthreshold swing of 35.4 mV/dec in the forward sweep. In addition, the fabricated trench Fe-FinFET had a very low drain-induced barrier lowering value of 4.47 mV/V and immunity to gate-induced drain leakage. Finally, a technology computer-aided design simulation was conducted to verify the experimental results.

Impact of Pt grain size on ferroelectric properties of zirconium hafnium oxide by chemical solution deposition.

The effects of the grain size of Pt bottom electrodes on the ferroelectricity of hafnium zirconium oxide (HZO) were studied in terms of the orthorhombic phase transformation. HZO thin films were deposited by chemical solution deposition on the Pt bottom electrodes with various grain sizes which had been deposited by direct current sputtering. All the samples were crystallized by rapid thermal annealing at 700 °C to allow a phase transformation. The crystallographic phases were determined by grazing incidence X-ray diffraction, which showed that the bottom electrode with smaller Pt grains resulted in a larger orthorhombic phase composition in the HZO film. As a result, capacitors with smaller Pt grains for the bottom electrode showed greater ferroelectric polarization. The smaller grains produced larger in-plane stress which led to more orthorhombic phase transformation and higher ferroelectric polarization.

Determination of Hafnium Zirconium Oxide Interfacial Band Alignments Using Internal Photoemission Spectroscopy and X-ray Photoelectron Spectroscopy.

Doped ferroelectric HfO2 is highly promising for integration into complementary metal-oxide semiconductor (CMOS) technology for devices such as ferroelectric nonvolatile memory and low-power field-effect transistors (FETs). We report the direct measurement of the energy barriers between various metal electrodes (Pt, Au, Ta, TaN, Ti/Pt, Ni, Al) and hafnium zirconium oxide (Hf0.58Zr0.42O2, HZO) using internal photoemission (IPE) spectroscopy. Results are compared with valence band offsets determined using the three-sample X-ray photoelectron spectroscopy (XPS) as well as the two-sample hard X-ray photoelectron spectroscopy (HAXPES) techniques. Both XPS and IPE indicate roughly the same dependence of the HZO barrier on metal work function with a slope of 0.8 ± 0.5. XPS and HAXPES-derived barrier heights are on average about 1.1 eV smaller than barrier heights determined by IPE, suggesting the presence of negative charge in the HZO.

Back-End, CMOS-Compatible Ferroelectric Field-Effect Transistor for Synaptic Weights.

Neuromorphic computing architectures enable the dense colocation of memory and processing elements within a single circuit. This colocation removes the communication bottleneck of transferring data between separate memory and computing units as in standard von Neuman architectures for data-critical applications including machine learning. The essential building blocks of neuromorphic systems are nonvolatile synaptic elements such as memristors. Key memristor properties include a suitable nonvolatile resistance range, continuous linear resistance modulation, and symmetric switching. In this work, we demonstrate voltage-controlled, symmetric and analog potentiation and depression of a ferroelectric Hf0.57Zr0.43O2 (HZO) field-effect transistor (FeFET) with good linearity. Our FeFET operates with low writing energy (fJ) and fast programming time (40 ns). Retention measurements have been performed over 4 bit depth with low noise (1%) in the tungsten oxide (WOx) readout channel. By adjusting the channel thickness from 15 to 8 nm, the on/off ratio of the FeFET can be engineered from 1 to 200% with an on-resistance ideally >100 kΩ, depending on the channel geometry. The device concept is using earth-abundant materials and is compatible with a back end of line (BEOL) integration into complementary metal-oxide-semiconductor (CMOS) processes. It has therefore a great potential for the fabrication of high-density, large-scale integrated arrays of artificial analog synapses.

Indium-Tin-Oxide Transistors with One Nanometer Thick Channel and Ferroelectric Gating.

In this work, we demonstrate high-performance indium-tin-oxide (ITO) transistors with a channel thickness down to 1 nm and ferroelectric Hf0.5Zr0.5O2 as gate dielectric. An on-current of 0.243 A/mm is achieved on submicron gate-length ITO transistors with a channel thickness of 1 nm, while it increases to as high as 1.06 A/mm when the channel thickness increases to 2 nm. A raised source/drain structure with a thickness of 10 nm is employed, contributing to a low contact resistance of 0.15 Ω·mm and a low contact resistivity of 1.1 × 10-7 Ω·cm2. The ITO transistor with a recessed channel and ferroelectric gating demonstrates several advantages over 2D semiconductor transistors and other thin-film transistors, including large-area wafer-size nanometer thin-film formation, low contact resistance and contact resistivity, an atomic thin channel being immune to short channel effects, large gate modulation of high carrier density by ferroelectric gating, high-quality gate dielectric and passivation formation, and a large bandgap for the low-power back-end-of-line complementary metal-oxide-semiconductor application.