Mechanism and Impact of Bipolar Current Voltage Asymmetry in Computational Phase-Change Memory

Nanoscale resistive memory devices are being explored for neuromorphic and in-memory computing. However, non-ideal device characteristics of read noise and resistance drift pose significant challenges to the achievable computational precision. Here, it is shown that there is an additional non-ideality that can impact computational precision, namely the bias-polarity-dependent current flow. Using phase-change memory (PCM) as a model system, it is shown that this "current-voltage" non-ideality arises both from the material and geometrical properties of the devices. Further, we discuss the detrimental effects of such bipolar asymmetry on in-memory matrix-vector multiply (MVM) operations and provide a scheme to compensate for it.

An integrated photonics engine for unsupervised correlation detection

With more and more aspects of modern life and scientific tools becoming digitized, the amount of data being generated is growing exponentially. Fast and efficient statistical processing, such as identifying correlations in big datasets, is therefore becoming increasingly important, and this, on account of the various compute bottlenecks in modern digital machines, has necessitated new computational paradigms. Here, we demonstrate one such novel paradigm, via the development of an integrated phase-change photonics engine. The computational memory engine exploits the accumulative property of Ge₂Sb₂Te₅ phase-change cells and wavelength division multiplexing property of optics in delivering fully parallelized and colocated temporal correlation detection computations. We investigate this property and present an experimental demonstration of identifying real-time correlations in data streams on the social media platform Twitter and high-traffic computing nodes in data centers. Our results demonstrate the use case of high-speed integrated photonics in accelerating statistical analysis methods.

Structural Assessment of Interfaces in Projected Phase-Change Memory

Non-volatile memories based on phase-change materials have gained ground for applications in analog in-memory computing. Nonetheless, non-idealities inherent to the material result in device resistance variations that impair the achievable numerical precision. Projected-type phase-change memory devices reduce these non-idealities. In a projected phase-change memory, the phase-change storage mechanism is decoupled from the information retrieval process by using projection of the phase-change material’s phase configuration onto a projection liner. It has been suggested that the interface resistance between the phase-change material and the projection liner is an important parameter that dictates the efficacy of the projection. In this work, we establish a metrology framework to assess and understand the relevant structural properties of the interfaces in thin films contained in projected memory devices. Using X-ray reflectivity, X-ray diffraction and transmission electron microscopy, we investigate the quality of the interfaces and the layers’ properties. Using demonstrator examples of Sb and Sb₂Te₃ phase-change materials, new deposition routes as well as stack designs are proposed to enhance the phase-change material to a projection-liner interface and the robustness of material stacks in the devices.

Phase-change memtransistive synapses for mixed-plasticity neural computations

In the mammalian nervous system, various synaptic plasticity rules act, either individually or synergistically, over wide-ranging timescales to enable learning and memory formation. Hence, in neuromorphic computing platforms, there is a significant need for artificial synapses that can faithfully express such multi-timescale plasticity mechanisms. Although some plasticity rules have been emulated with elaborate complementary metal oxide semiconductor and memristive circuitry, device-level hardware realizations of long-term and short-term plasticity with tunable dynamics are lacking. Here we introduce a phase-change memtransistive synapse that leverages both the non-volatility of the phase configurations and the volatility of field-effect modulation for implementing tunable plasticities. We show that these mixed-plasticity synapses can enable plasticity rules such as short-term spike-timing-dependent plasticity that helps with the modelling of dynamic environments. Further, we demonstrate the efficacy of the memtransistive synapses in realizing accelerators for Hopfield neural networks for solving combinatorial optimization problems.

State dependence and temporal evolution of resistance in projected phase change memory

Phase change memory (PCM) is being actively explored for in-memory computing and neuromorphic systems. The ability of a PCM device to store a continuum of resistance values can be exploited to realize arithmetic operations such as matrix-vector multiplications or to realize the synaptic efficacy in neural networks. However, the resistance variations arising from structural relaxation, 1/f noise, and changes in ambient temperature pose a key challenge. The recently proposed projected PCM concept helps to mitigate these resistance variations by decoupling the physical mechanism of resistance storage from the information-retrieval process. Even though the device concept has been proven successfully, a comprehensive understanding of the device behavior is still lacking. Here, we develop a device model that captures two key attributes, namely, resistance drift and the state dependence of resistance. The former refers to the temporal evolution of resistance, while the latter refers to the dependence of the device resistance on the phase configuration of the phase change material. The study provides significant insights into the role of interfacial resistance in these devices. The model is experimentally validated on projected PCM devices based on antimony and a metal nitride fabricated in a lateral device geometry and is also used to provide guidelines for material selection and device engineering.

Filamentary High-Resolution Electrical Probes for Nanoengineering

Confining electric fields to a nanoscale region is challenging yet crucial for applications such as high-resolution probing of electrical properties of materials and electric-field manipulation of nanoparticles. State-of-the-art techniques involving atomic force microscopy typically have a lateral resolution limit of tens of nanometers due to limitations in the probe geometry and stray electric fields that extend over space. Engineering the probes is the most direct approach to improving this resolution limit. However, current methods to fabricate high-resolution probes, which can effectively confine the electric fields laterally, involve expensive and sophisticated probe manipulation, which has limited the use of this approach. Here, we demonstrate that nanoscale phase switching of configurable thin films on probes can result in high-resolution electrical probes. These configurable coatings can be both germanium-antimony-tellurium (GST) as well as amorphous-carbon, materials known to undergo electric field-induced nonvolatile, yet reversible switching. By forming a localized conductive filament through phase transition, we demonstrate a spatial resolution of electrical field beyond the geometrical limitations of commercial platinum probes (i.e., an improvement of ∼48%). We then utilize these confined electric fields to manipulate nanoparticles with single nanoparticle precision via dielectrophoresis. Our results advance the field of nanomanufacturing and metrology with direct applications for pick and place assembly at the nanoscale.

Strong Opto-Structural Coupling in Low Dimensional GeSe3 Films

Chalcogenide glasses as nanoscale thin films have become leading candidates for several optical and photonic technologies, ranging from reflective displays and filters to photonic memories. Current material systems, however, show strong optical absorption which limits their performance efficiencies and complicates device level integration. Herein, we report sputter deposited thin films of GeSe₃, which are low loss and in which the flexible nature of the atomic structure results in thermally activated tunability in the refractive index as well as in the film's physical volume. Such changes, which occur beyond a threshold temperature are observed to be accumulative and directed toward a more equilibrium amorphous state of the film, instead of crystallization. Our results provide insight into a new type of configurability that is based on strong coupling in the material's opto-structural properties. The low optical losses in this material system combined with the tunability in the optical properties in the visible and near-infrared have direct application in higher performing optical coatings and in corrective optics.

Grain Boundaries as Electrical Conduction Channels in Polycrystalline Monolayer WS2

We show that grain boundaries (GBs) in polycrystalline monolayer WS₂ can act as conduction channels with a lower gate onset potential for field-effect transistors made parallel, compared to devices made in pristine areas and perpendicular to GBs. Localized doping at the GB causes photoluminescence quenching and a reduced Schottky barrier with the metal electrodes, resulting in higher conductivity at lower applied bias values. Samples are grown by chemical vapor deposition with large domains of ∼100 μm, enabling numerous devices to be made within single domains, across GBs and at many similar sites across the substrate to reveal similar behaviors. We corroborate our electrical measurements with Kelvin probe microscopy, highlighting the nature of the doping-type in the material to change at the grain boundaries. Molecular dynamics simulations of the GB are used to predict the atomic structure of the dislocations and meandering tilt GB behavior on the nanoscale. These results show that GBs can be used to provide conduction pathways that are different to transport across GBs and in pristine area for potential electronic applications.

Engineering Interface-Dependent Photoconductivity in Ge2Sb2Te5 Nanoscale Devices

Phase-change materials are increasingly being explored for photonics applications, ranging from high-resolution displays to artificial retinas. Surprisingly, our understanding of the underlying mechanism of light-matter interaction in these materials has been limited to photothermal crystallization because of its relevance in applications such as rewritable optical discs. Here, we report a photoconductivity study of nanoscale thin films of phase-change materials. We identify strong photoconductive behavior in phase-change materials, which we show to be a complex interplay of three independent mechanisms: photoconductive, photoinduced crystallization, and photoinduced thermoelectric effects. We find that these effects also congruously contribute to a substantial photovoltaic effect, even in notionally symmetric devices. Notably, we show that device engineering plays a decisive role in determining the dominant mechanism; the contribution of the photothermal effects to the extractable photocurrent can be reduced to <0.4% by varying the electrodes and device geometry. We then show that the contribution of these individual effects to the photoresponse is phase-dependent with the amorphous state being more photoactive than the crystalline state and that a reversible change occurs in the charge transport from thermionic to tunnelling during phase transformation. Finally, we demonstrate photodetectors with an order of magnitude tunability in photodetection responsivity and bandwidth using these materials. Our results provide insights to the photophysics of phase-change materials and highlight their potential in future optoelectronics.

Inhomogeneous Strain Release during Bending of WS2 on Flexible Substrates

Two-dimensional (2D) materials hold great promise in flexible electronics, but the weak van der Waals interlayer bonding may pose a problem during bending, where easy interlayer sliding can occur. Furthermore, thin films of rigid materials are often observed to delaminate from soft substrates during straining. Here, we study the influence of substrate strain on some of the heterostructure configurations we expect to find in devices, composed of three common 2D materials: graphene, tungsten disulfide, and boron nitride. We used photoluminescence (PL) spectroscopy to measure changes in the heterostructures with strain applied in situ. All heterostructures were fabricated directly on polymer substrates, using materials synthesized by chemical vapor deposition. We observed an inhomogeneous release of strain in all structures, leading to a nonrecoverable broadening of the PL peak and shift of the bandgap. This suggests the need for preconditioning devices before service to ensure stable behavior. A gradual time-dependent relaxation of strain between strain cycles was characterized using time-dependent measurements-an effect which could lead to drift of device behavior during operation. Furthermore, possible degradation was assessed by performing the strain and relax the cycle up to 200 times, where we found little further change after the initial shifts had stabilized. These results have important ramifications for devices fabricated from these and other 2D materials, as they suggest extra processing steps and considerations that must be taken to achieve consistent and stable properties.

Revealing Strain-Induced Effects in Ultrathin Heterostructures at the Nanoscale

Two-dimensional materials are being increasingly studied, particularly for flexible and wearable technologies because of their inherent thickness and flexibility. Crucially, one aspect where our understanding is still limited is on the effect of mechanical strain, not on individual sheets of materials, but when stacked together as heterostructures in devices. In this paper, we demonstrate the use of Kelvin probe microscopy in capturing the influence of uniaxial tensile strain on the band-structures of graphene and WS₂ (mono- and multilayered) based heterostructures at high resolution. We report a major advance in strain characterization tools through enabling a single-shot capture of strain defined changes in a heterogeneous system at the nanoscale, overcoming the limitations (materials, resolution, and substrate effects) of existing techniques such as optical spectroscopy. Using this technique, we observe that the work-functions of graphene and WS₂ increase as a function of strain, which we attribute to the Fermi level lowering from increased p-doping. We also extract the nature of the interfacial heterojunctions and find that they get strongly modulated from strain. We observe that the strain-enhanced charge transfer with the substrate plays a dominant role, causing the heterostructures to behave differently from two-dimensional materials in their isolated forms.

Large Dendritic Monolayer MoS2 Grown by Atmospheric Pressure Chemical Vapor Deposition for Electrocatalysis

The edge sites of MoS₂ are catalytically active for the hydrogen evolution reaction (HER), and growing monolayer structures that are edge-rich is desirable. Here, we show the production of large-area highly branched MoS₂ dendrites on amorphous SiO₂/Si substrates using an atmospheric pressure chemical vapor deposition and explore their use in electrocatalysis. By tailoring the substrate construction, the monolayer MoS₂ evolves from triangular to dendritic morphology because of the change of growth conditions. The rough edges endow dendritic MoS₂ with a fractal dimension down to 1.54. The highly crystalline basal plane and the edge of the dendrites are visualized at atomic resolution using an annular dark field scanning transmission electron microscope. The monolayer dendrites exhibit strong photoluminescence, which is indicative of the direct band gap emission, which is preserved after being transferred. Post-transfer sulfur annealing restores the structural defects and decreases the n-type doping in MoS₂ monolayers. The annealed MoS₂ dendrites show good and highly durable HER performance on the glassy carbon with a large exchange current density of 32 μA cm_geo^-2, demonstrating its viability as an efficient HER catalyst.

Scaling Limits of Graphene Nanoelectrodes

Graphene nanogap electrodes have been of recent interest in a variety of fields, ranging from molecular electronics to phase change memories. Several recent reports have highlighted that scaling graphene nanogaps to even smaller sizes is a promising route to more efficient and robust molecular and memory devices. Despite the significant interest, the operating and scaling limits of these electrodes are completely unknown. In this paper, we report on our observations of consistent voltage driven resistance switching in sub-5 nm graphene nanogaps. We find that such electrical switching from an insulating state to a conductive state occurs at very low currents and voltages (0.06 μA and 140 mV), independent of the conditions (room ambient, low temperatures, as well as in vacuum), thus portending potential limits to scaling of functional devices with carbon electrodes. We then associate this phenomenon to the formation and rupture of carbon chains. Using a phase change material in the nanogap as a demonstrator device, fabricated using a self-alignment technique, we show that for gap sizes approaching 1 nm the switching is dominated by such carbon chain formation, creating a fundamental scaling limit for potential devices. These findings have important implications, not only for fundamental science, but also in terms of potential applications.