Mechanism of interstitial oxygen diffusion in hafnia

Multilayer redox-based HfOx/Al2O3/TiO2 memristive structures for neuromorphic computing

Redox-based memristive devices have shown great potential for application in neuromorphic computing systems. However, the demands on the device characteristics depend on the implemented computational scheme and unifying the desired properties in one stable device is still challenging. Understanding how and to what extend the device characteristics can be tuned and stabilized is crucial for developing application specific designs. Here, we present memristive devices with a functional trilayer of HfO_x/Al₂O₃/TiO₂ tailored by the stoichiometry of HfO_x (x = 1.8, 2) and the operating conditions. The device properties are experimentally analyzed, and a physics-based device model is developed to provide a microscopic interpretation and explain the role of the Al₂O₃ layer for a stable performance. Our results demonstrate that the resistive switching mechanism can be tuned from area type to filament type in the same device, which is well explained by the model: the Al₂O₃ layer stabilizes the area-type switching mechanism by controlling the formation of oxygen vacancies at the Al₂O₃/HfO_x interface with an estimated formation energy of ≈ 1.65 ± 0.05 eV. Such stabilized area-type devices combine multi-level analog switching, linear resistance change, and long retention times (≈ 10⁷-10⁸ s) without external current compliance and initial electroforming cycles. This combination is a significant improvement compared to previous bilayer devices and makes the devices potentially interesting for future integration into memristive circuits for neuromorphic applications.

Afterglow-Suppressed Lu2O3:Eu3+ Nanoscintillators for High-Resolution and Dynamic Digital Radiographic Imaging

Lu_2(1-x)Eu_2xO₃ nanoscintillators (x = 0.005, 0.01, 0.03, 0.05, 0.07, and 0.10) with red emission were synthesized by a coprecipitation method. It is found that their photo- and radioluminescence intensities increase with increasing Eu³⁺ concentration until x = 0.05. According to their concentration-dependent luminescence intensity ratios (I₆₁₀(C₂)/I₅₈₂(S₆)), the existing energy transfer from Eu³⁺(S₆) (occupying S₆ sites) to Eu³⁺(C₂) (occupying C₂ sites) can be confirmed. Based on the spectral data and density functional theory (DFT) calculations, the origin of Lu₂O₃:Eu³⁺ persistent luminescence at low concentration might be related to the tunneling processes between Eu³⁺ (occupying C₂ and S₆ sites) and oxygen interstitials (O_i^×). After dispersing afterglow-suppressed Lu₂O₃:Eu³⁺ nanoscintillators into polymethyl methacrylate (PMMA) polymer-acetone solution, flexible PMMA-Lu₂O₃:Eu³⁺ composite films with high thermal stability and radiation resistance were fabricated by a doctor blade method. As the flexible composite film was used as an imaging plate, static X-ray images with high spatial resolution (5.5 lp/mm) under an extremely low dose of ∼1.1 μGy_air can be acquired. When a watch with a moving second hand was used as an object, the dynamic X-ray imaging can be realized under a dose rate of 55 μGy_air·s^-1. Our results demonstrate that Lu₂O₃:Eu³⁺ nanoscintillators can be regarded as candidate materials for dynamic digital radiographic imaging.

Mechanism of creation and destruction of oxygen interstitial atoms by nonpolar zinc oxide(101[combining macron]0) surfaces

Oxygen vacancies (VO) influence many properties of ZnO in semiconductor devices, yet synthesis methods leave behind variable and unpredictable VO concentrations. Oxygen interstitials (Oi) move far more rapidly, so post-synthesis introduction of Oi to control the VO concentration would be desirable. Free surfaces offer such an introduction mechanism if they are free of poisoning foreign adsorbates. Here, isotopic exchange experiments between nonpolar ZnO(101[combining macron]0) and O2 gas, together with mesoscale modeling and first-principles calculations, point to an activation barrier for injection only 0.1-0.2 eV higher than for bulk site hopping. The modest barrier for hopping in turn enables diffusion lengths of tens to hundreds of nanometers only slightly above room temperature, which should facilitate defect engineering under very modest conditions. In addition, low hopping barriers coupled with statistical considerations lead to important qualitative manifestations in diffusion via an interstitialcy mechanism that does not occur for vacancies.

Enhancing Magneto-Ionic Effects in Magnetic Nanostructured Films via Conformal Deposition of Nanolayers with Oxygen Acceptor/Donor Capabilities

Effective manipulation of the magnetic properties of nanostructured metallic alloys, exhibiting intergrain porosity (i.e., channels) and conformally coated with insulating oxide nanolayers, with an electric field is demonstrated. Nanostructured Co-Pt films are grown by electrodeposition (ED) and subsequently coated with either AlO_x or HfO_x by atomic layer deposition (ALD) to promote magneto-ionic effects (i.e., voltage-driven ion migration) during electrolyte gating. Pronounced variations in coercivity (H_C) and magnetic moment at saturation (m_S) are found at room temperature after biasing the heterostructures. The application of a negative voltage results in a decrease of H_C and an increase of m_S, whereas the opposite trend is achieved for positive voltages. Although magneto-ionic phenomena are already observed in uncoated Co-Pt films (because of the inherent presence of oxygen), the ALD oxide nanocoatings serve to drastically enhance the magneto-ionic effects because of partially reversible oxygen migration, driven by voltage, across the interface between AlO_x or HfO_x and the nanostructured Co-Pt film. Co-Pt/HfO_x heterostructures exhibit the most significant magneto-electric response at negative voltages, with an increase of m_S up to 76% and a decrease of H_C by 58%. The combination of a nanostructured magnetic alloy and a skinlike insulating oxide nanocoating is shown to be appealing to enhance magneto-ionic effects, potentially enabling electrolyte-gated magneto-ionic technology.

Multiscale Modeling for Application-Oriented Optimization of Resistive Random-Access Memory

Memristor-based neuromorphic systems have been proposed as a promising alternative to von Neumann computing architectures, which are currently challenged by the ever-increasing computational power required by modern artificial intelligence (AI) algorithms. The design and optimization of memristive devices for specific AI applications is thus of paramount importance, but still extremely complex, as many different physical mechanisms and their interactions have to be accounted for, which are, in many cases, not fully understood. The high complexity of the physical mechanisms involved and their partial comprehension are currently hampering the development of memristive devices and preventing their optimization. In this work, we tackle the application-oriented optimization of Resistive Random-Access Memory (RRAM) devices using a multiscale modeling platform. The considered platform includes all the involved physical mechanisms (i.e., charge transport and trapping, and ion generation, diffusion, and recombination) and accounts for the 3D electric and temperature field in the device. Thanks to its multiscale nature, the modeling platform allows RRAM devices to be simulated and the microscopic physical mechanisms involved to be investigated, the device performance to be connected to the material’s microscopic properties and geometries, the device electrical characteristics to be predicted, the effect of the forming conditions (i.e., temperature, compliance current, and voltage stress) on the device’s performance and variability to be evaluated, the analog resistance switching to be optimized, and the device’s reliability and failure causes to be investigated. The discussion of the presented simulation results provides useful insights for supporting the application-oriented optimization of RRAM technology according to specific AI applications, for the implementation of either non-volatile memories, deep neural networks, or spiking neural networks.

Metamaterial emitter for thermophotovoltaics stable up to 1400 °C

High temperature stable selective emitters can significantly increase efficiency and radiative power in thermophotovoltaic (TPV) systems. However, optical properties of structured emitters reported so far degrade at temperatures approaching 1200 °C due to various degradation mechanisms. We have realized a 1D structured emitter based on a sputtered W-HfO₂ layered metamaterial and demonstrated desired band edge spectral properties at 1400 °C. To the best of our knowledge the temperature of 1400 °C is the highest reported for a structured emitter, so far. The spatial confinement and absence of edges stabilizes the W-HfO₂ multilayer system to temperatures unprecedented for other nanoscaled W-structures. Only when this confinement is broken W starts to show the well-known self-diffusion behavior transforming to spherical shaped W-islands. We further show that the oxidation of W by atmospheric oxygen could be prevented by reducing the vacuum pressure below 10⁻⁵ mbar. When oxidation is mitigated we observe that the 20 nm spatially confined W films survive temperatures up to 1400 °C. The demonstrated thermal stability is limited by grain growth in HfO₂, which leads to a rupture of the W-layers, thus, to a degradation of the multilayer system at 1450 °C.

Modelling of oxygen vacancy aggregates in monoclinic HfO2: can they contribute to conductive filament formation?

Formation of metal rich conductive filaments and their rearrangements determine the switching characteristics in HfO2 based resistive random access memory (RRAM) devices. The initiation of a filament formation process may occur either via aggregation of pre-existing vacancies randomly distributed in the oxide or via generation of new oxygen vacancies close to the pre-existing ones. We evaluate the feasibility of vacancy aggregation processes by calculating the structures and binding energies of oxygen vacancy aggregates consisting of 2, 3 and 4 vacancies in bulk monoclinic (m)-HfO2 using density functional theory (DFT). We demonstrate that formation of neutral oxygen vacancy aggregates is accompanied by small energy gain, which depends on the size and shape of the aggregate. In the most strongly bound configurations, vacancies are unscreened by Hf cations and form voids within the crystal, with the larger aggregates having larger binding energy per vacancy (-0.11 to  -0.18 eV). The negatively charged di-vacancy was found to have similar binding energies to the neutral one, while the positively charged di-vacancy was found to be unstable. Thus aggregation process of either neutral or negatively charged oxygen vacancies is energetically feasible.

Oxygen self-diffusion in HfO2 studied by electron spectroscopy

High-resolution measurement of the energy of electrons backscattered from oxygen atoms makes it possible to distinguish between (18)O and (16)O isotopes as the energy of elastically scattered electrons depends on the mass of the scattering atom. Here we show that this approach is suitable for measuring oxygen self-diffusion in HfO2 using a Hf(16)O2 (20 nm)/Hf(18)O2 bilayers (60 nm). The mean depth probed (for which the total path length equals the inelastic mean free path) is either 5 or 15 nm in our experiment, depending on the geometry used. Before annealing, the elastic peak from O is thus mainly due to electrons scattered from (16)O in the outer layer, while after annealing the signal from (18)O increases due to diffusion from the underlying Hf(18)O2 layer. For high annealing temperatures the observed interdiffusion is consistent with an activation energy of 1 eV, but at lower temperatures interdiffusion decreases with increasing annealing time. We interpret this to be a consequence of defects, present in the layers early on and enhancing the oxygen diffusivity, disappearing during the annealing process.

Kinetics of Frenkel Defect Formation in TiO2 from First Principles

Motivated by the unusual behavior of TiO2 films seen in electrical stress and defect annealing experiments, we studied the energy profile for forming a Frenkel defect in rutile TiO2, using first-principles calculations with a nudged-elastic-band method. We found strongly asymmetric diffusion barriers. The Frenkel pairs with small separation are exceedingly short-lived: the Ti interstitial position nearest to the the Ti vacancy is separated by only a 0.15eV barrier, and the next-nearest interstitial position is dynamically unstable. The formation enthalpies of Frenkel pairs with larger separation gradually vary between 4.2 and 5.0 eV, separated by 0.3-0.4eV barriers along the (001) direction. Contrary to some previous studies, we do not find Frenkel configurations with tetrahedrally bonded Ti interstitials. The very low barriers for Frenkel defect evolution are consistent with the observations from the electrical stress damage annealing experiments.

Theoretical Analysis of Oxygen Diffusion in monoclinic HfO2

In this report, we present the detailed analysis of the interstitial oxygen (O2+, O0, O2-) diffusion in monoclinic HfO2 (hafnia) using the first principles calculations. The interstitial oxygen atom kicks out the oxygen atom at the 3-fold-site and occupies the 3-fold-site. And then the newly kicked-out interstitial oxygen atom jumps to the nearest neighbor site and couples again with the atoms at the crystal sites. This kick-out- mechanism is valid for all charge states of the interstitial oxygen in monoclinic HfO2. In hafnia, the interstitial oxygen atom can take 3 charge states (+2, 0, -2) depending on the chemical potential (Ef), whereas the oxygen-vacancy in hafnia can get +2 or 0 charge state being dependent on Ef. In the lower range of Ef, O2+ and O0 might contribute. In the middle range of Ef, the O2- does not contribute to the diffusion process in hafnia because of the pair annihilation process between O2- and oxygen vacancy (V2+) defect pair. We can simulate such a pair annihilation process in hafnia. In the higher range, O2- might contribute the diffusion process.