Experimental demonstration of an on-chip p-bit core based on stochastic magnetic tunnel junctions and 2D MoS2 transistors.

Probabilistic computing is a computing scheme that offers a more efficient approach than conventional complementary metal-oxide-semiconductor (CMOS)-based logic in a variety of applications ranging from optimization to Bayesian inference, and invertible Boolean logic. The probabilistic bit (or p-bit, the base unit of probabilistic computing) is a naturally fluctuating entity that requires tunable stochasticity; by coupling low-barrier stochastic magnetic tunnel junctions (MTJs) with a transistor circuit, a compact implementation is achieved. In this work, by combining stochastic MTJs with 2D-MoS2 field-effect transistors (FETs), we demonstrate an on-chip realization of a p-bit building block displaying voltage-controllable stochasticity. Supported by circuit simulations, we analyze the three transistor-one magnetic tunnel junction (3T-1MTJ) p-bit design, evaluating how the characteristics of each component influence the overall p-bit output. While the current approach has not reached the level of maturity required to compete with CMOS-compatible MTJ technology, the design rules presented in this work are valuable for future experimental implementations of scaled on-chip p-bit networks with reduced footprint.

Tailoring amorphous boron nitride for high-performance two-dimensional electronics.

Two-dimensional (2D) materials have garnered significant attention in recent years due to their atomically thin structure and unique electronic and optoelectronic properties. To harness their full potential for applications in next-generation electronics and photonics, precise control over the dielectric environment surrounding the 2D material is critical. The lack of nucleation sites on 2D surfaces to form thin, uniform dielectric layers often leads to interfacial defects that degrade the device performance, posing a major roadblock in the realization of 2D-based devices. Here, we demonstrate a wafer-scale, low-temperature process (<250 °C) using atomic layer deposition (ALD) for the synthesis of uniform, conformal amorphous boron nitride (aBN) thin films. ALD deposition temperatures between 125 and 250 °C result in stoichiometric films with high oxidative stability, yielding a dielectric strength of 8.2 MV/cm. Utilizing a seed-free ALD approach, we form uniform aBN dielectric layers on 2D surfaces and fabricate multiple quantum well structures of aBN/MoS2 and aBN-encapsulated double-gated monolayer (ML) MoS2 field-effect transistors to evaluate the impact of aBN dielectric environment on MoS2 optoelectronic and electronic properties. Our work in scalable aBN dielectric integration paves a way towards realizing the theoretical performance of 2D materials for next-generation electronics.

High-Performance Complementary Circuits from Two-Dimensional MoTe2.

Two-dimensional (2D) materials hold great promise for future complementary metal-oxide semiconductor (CMOS) technology. However, the lack of effective methods to tune the Schottky barrier poses a challenge in constructing high-performance complementary circuits from the same material. Here, we reveal that the polarity of pristine MoTe2 field-effect transistors (FETs) with minimized air exposure is n-type, irrespective of the metal contact type. The fabricated n-FETs with palladium contact can reach electron currents up to 275 µA/µm at VDS = 2 V. For p-FETs, we introduce a novel nitric oxide doping strategy, allowing a controlled transition of MoTe2 FETs from n-type to unipolar p-type. By doping only in the contact region, we demonstrate hole currents up to 170 µA/µm at VDS= -2 V with preserved Ion/Ioff ratios of 105. Finally, we present a complementary inverter circuit comprising the high-performance n- and p-type FETs based on MoTe2, promoting the application of 2D materials in future electronic systems.

Statistical Assessment of High-Performance Scaled Double-Gate Transistors from Monolayer WS2.

Scaling of monolayer transition metal dichalcogenide (TMD) field-effect transistors (FETs) is an important step toward evaluating the application space of TMD materials. Although some work on ultrashort channel monolayer (ML) TMD FETs has been published, there exist no comprehensive studies that assess their performance in a statistically relevant manner, providing critical insights into the impact of the device geometry. Part of the reason for the absence of such a study is the substantial variability of TMD devices when processes are not carefully controlled. In this work, we show a statistical study of ultrashort channel double-gated ML WS2 FETs exhibiting excellent device performance and limited device-to-device variations. From a detailed analysis of cross-sectional scanning transmission electron microscopy (STEM) images and careful technology computer aided design (TCAD) simulations, we evaluated, in particular, an unexpected deterioration of the subthreshold characteristics for our shortest devices. Two potential candidates for the observed behavior were identified, i.e., buckling of the TMD on the substrate and loss of gate control due to the source geometry and the high-k dielectric between the metal gate and the metal source electrode.

Mobility Extraction in 2D Transition Metal Dichalcogenide Devices-Avoiding Contact Resistance Implicated Overestimation.

Schottky barrier (SB) transistors operate distinctly different from conventional metal-oxide semiconductor field-effect transistors, in a unique way that the gate impacts the carrier injection from the metal source/drain contacts into the channel region. While it has been long recognized that this can have severe implications for device characteristics in the subthreshold region, impacts of contact gating of SB in the on-state of the devices, which affects evaluation of intrinsic channel properties, have been yet comprehensively studied. Due to the fact that contact resistance (RC ) is always gate-dependent in a typical back-gated device structure, the traditional approach of deriving field-effect mobility from the maximum transconductance (gm ) is in principle not correct and can even overestimate the mobility. In addition, an exhibition of two different threshold voltages for the channel and the contact region leads to another layer of complexity in determining the true carrier concentration calculated from Q = COX * (VG -VTH ). Through a detailed experimental analysis, the effect of different effective oxide thicknesses, distinct SB heights, and doping-induced reductions in the SB width are carefully evaluated to gain a better understanding of their impact on important device metrics.

Artificial Sub-60 Millivolts/Decade Switching in a Metal-Insulator-Metal-Insulator-Semiconductor Transistor without a Ferroelectric Component.

Negative capacitance field-effect transistors (NC-FETs) have attracted wide interest as promising candidates for steep-slope devices, and sub-60 millivolts/decade (mV/decade) switching has been demonstrated in NC-FETs with various device structures and material systems. However, the detailed mechanisms of the observed steep-slope switching in some of these experiments are under intense debate. Here we show that sub-60 mV/decade switching can be observed in a WS2 transistor with a metal-insulator-metal-insulator-semiconductor (MIMIS) structure without any ferroelectric component. This structure resembles an NC-FET with internal gate, except that the ferroelectric layer is replaced by a leaky dielectric layer. Through simulations of the charging dynamics during the device characterization using a resistor-capacitor network model, we show that the observed steep-slope switching in our "ferroelectric-free" transistors can be attributed to the internal gate voltage response to the chosen varying gate voltage scan rates. We further show that a constant gate voltage scan rate can also lead to transient sub-60 mV/decade switching in an MIMIS structure with voltage-dependent internal gate capacitance. Our results indicate that the observation of sub-60 mV/decade switching alone is not sufficient evidence for the successful demonstration of a true steep-slope switching device and that experimentalists need to critically assess their measurement setups to avoid measurement-related artifacts.

Hardware implementation of Bayesian network building blocks with stochastic spintronic devices.

Bayesian networks are powerful statistical models to understand causal relationships in real-world probabilistic problems such as diagnosis, forecasting, computer vision, etc. For systems that involve complex causal dependencies among many variables, the complexity of the associated Bayesian networks become computationally intractable. As a result, direct hardware implementation of these networks is one promising approach to reducing power consumption and execution time. However, the few hardware implementations of Bayesian networks presented in literature rely on deterministic CMOS devices that are not efficient in representing the stochastic variables in a Bayesian network that encode the probability of occurrence of the associated event. This work presents an experimental demonstration of a Bayesian network building block implemented with inherently stochastic spintronic devices based on the natural physics of nanomagnets. These devices are based on nanomagnets with perpendicular magnetic anisotropy, initialized to their hard axes by the spin orbit torque from a heavy metal under-layer utilizing the giant spin Hall effect, enabling stochastic behavior. We construct an electrically interconnected network of two stochastic devices and manipulate the correlations between their states by changing connection weights and biases. By mapping given conditional probability tables to the circuit hardware, we demonstrate that any two node Bayesian networks can be implemented by our stochastic network. We then present the stochastic simulation of an example case of a four node Bayesian network using our proposed device, with parameters taken from the experiment. We view this work as a first step towards the large scale hardware implementation of Bayesian networks.

Demonstration of a pseudo-magnetization based simultaneous write and read operation in a Co60Fe20B20/Pb(Mg1/3Nb2/3)0.7Ti0.3O3 heterostructure.

Taking advantage of the magnetoelectric and its inverse effect, this article demonstrates strain-mediated magnetoelectric write and read operations simultaneously in Co60Fe20B20/Pb(Mg1/3Nb2/3)0.7Ti0.3O3 heterostructures based on a pseudo-magnetization µ ≡ mx2 - my2. By applying an external DC-voltage across a (011)-cut PMN-PT substrate, the ferroelectric polarization is re-oriented, which results in an anisotropic in-plane strain that transfers to the CoFeB thin film and changes its magnetic anisotropy Hk. The change in Hk in-turn results in a 90° rotation of the magnetic easy axis for sufficiently high voltages. Simultaneously, the inverse effect is employed to read changes of the magnetic properties. The change of magnetization in ferromagnetic (FM) layer induces an elastic stress in the piezoelectric (PE) layer, which generates a PE potential that can be used to readout the magnetic state of the FM layer. The experimental results are in excellent qualitative agreement with an equivalent circuit model that considers how magnetic properties are electrically controlled in such a PE/FM heterostructure and how a back-voltage is generated due to changing magnetic properties in a self-consistent model. We demonstrated that a change of easy axis of magnetization due to an applied voltage can be directly used for information processing, which is essential for future ME based devices.

Efficient Spin-Orbit Torque Switching of the Semiconducting Van Der Waals Ferromagnet Cr2 Ge2 Te6.

Being able to electrically manipulate the magnetic properties in recently discovered van der Waals ferromagnets is essential for their integration in future spintronics devices. Here, the magnetization of a semiconducting 2D ferromagnet, i.e., Cr2 Ge2 Te6 , is studied using the anomalous Hall effect in Cr2 Ge2 Te6 /tantalum heterostructures. The thinner the flakes, hysteresis and remanence in the magnetization loop with out-of-plane magnetic fields become more prominent. In order to manipulate the magnetization in such thin flakes, a combination of an in-plane magnetic field and a charge current flowing through Ta-a heavy metal exhibiting giant spin Hall effect-is used. In the presence of in-plane fields of 20 mT, charge current densities as low as 5 × 105 A cm-2 are sufficient to switch the out-of-plane magnetization of Cr2 Ge2 Te6 . This finding highlights that current densities required for spin-orbit torque switching of Cr2 Ge2 Te6 are about two orders of magnitude lower than those required for switching nonlayered metallic ferromagnets such as CoFeB. The results presented here show the potential of 2D ferromagnets for low-power memory and logic applications.

MoS2 for Enhanced Electrical Performance of Ultrathin Copper Films.

Copper nanowires are widely used as on-chip interconnects due to their superior conductivity. However, with aggressive Cu interconnect scaling, surface scattering of electrons drastically increases the electrical resistivity. In this work, we have studied the electrical performance of Cu thin films deposited on different materials. By comparing the thickness dependence of Cu films' resistivity on MoS2 and SiO2, we have demonstrated that MoS2 can be used to enhance the electrical performance of ultrathin Cu films due to improved specular surface scattering by up to 40%. By fitting the experimental data with the theoretical Fuchs-Sondheimer (FS) model, we have determined the specularity parameter at the Cu/MoS2 interface to be p ≈ 0.4 at room temperature. Furthermore, first principle calculations based on density functional theory (DFT) indicate that the localized density of states (LDOS) at the Cu/amorphous SiO2 interface is larger than the LDOS at the Cu/MoS2 interface, which is believed to be responsible for the higher resistivity in the Cu thin films that are deposited on SiO2 substrates. Our results suggest that MoS2 may serve as a performance enhancer for future generations of Cu interconnects.

Electric-field induced structural transition in vertical MoTe2- and Mo1-xWxTe2-based resistive memories.

Transition metal dichalcogenides have attracted attention as potential building blocks for various electronic applications due to their atomically thin nature and polymorphism. Here, we report an electric-field-induced structural transition from a 2H semiconducting to a distorted transient structure (2Hd) and orthorhombic Td conducting phase in vertical 2H-MoTe2- and Mo1-xWxTe2-based resistive random access memory (RRAM) devices. RRAM programming voltages are tunable by the transition metal dichalcogenide thickness and show a distinctive trend of requiring lower electric fields for Mo1-xWxTe2 alloys versus MoTe2 compounds. Devices showed reproducible resistive switching within 10 ns between a high resistive state and a low resistive state. Moreover, using an Al2O3/MoTe2 stack, On/off current ratios of 106 with programming currents lower than 1 µA were achieved in a selectorless RRAM architecture. The sum of these findings demonstrates that controlled electrical state switching in two-dimensional materials is achievable and highlights the potential of transition metal dichalcogenides for memory applications.

Complementary Black Phosphorus Tunneling Field-Effect Transistors.

Band-to-band tunneling field-effect transistors (TFETs) have emerged as promising candidates for low-power integration circuits beyond conventional metal-oxide-semiconductor field-effect transistors (MOSFETs) and have been demonstrated to overcome the thermionic limit, which results intrinsically in sub-threshold swings of at least 60 mV/dec at room temperature. Here, we demonstrate complementary TFETs based on few-layer black phosphorus, in which multiple top gates create electrostatic doping in the source and drain regions. By electrically tuning the doping types and levels in the source and drain regions, the device can be reconfigured to allow for TFET or MOSFET operation and can be tuned to be n-type or p-type. Owing to the proper choice of materials and careful engineering of device structures, record-high current densities have been achieved in 2D TFETs. Full-band atomistic quantum transport simulations of the fabricated devices agree quantitatively with the current-voltage measurements, which gives credibility to the promising simulation results of ultrascaled phosphorene TFETs. Using atomistic simulations, we project substantial improvements in the performance of the fabricated TFETs when channel thicknesses and oxide thicknesses are scaled down.

Spin-torque devices with hard axis initialization as Stochastic Binary Neurons.

Employing the probabilistic nature of unstable nano-magnet switching has recently emerged as a path towards unconventional computational systems such as neuromorphic or Bayesian networks. In this letter, we demonstrate proof-of-concept stochastic binary operation using hard axis initialization of nano-magnets and control of their output state probability (activation function) by means of input currents. Our method provides a natural path towards addition of weighted inputs from various sources, mimicking the integration function of neurons. In our experiment, spin orbit torque (SOT) is employed to "drive" nano-magnets with perpendicular magnetic anisotropy (PMA) -to their metastable state, i.e. in-plane hard axis. Next, the probability of relaxing into one magnetization state (+mi) or the other (-mi) is controlled using an Oersted field generated by an electrically isolated current loop, which acts as a "charge" input to the device. The final state of the magnet is read out by the anomalous Hall effect (AHE), demonstrating that the magnetization can be probabilistically manipulated and output through charge currents, closing the loop from charge-to-spin and spin-to-charge conversion. Based on these building blocks, a two-node directed network is successfully demonstrated where the status of the second node is determined by the probabilistic output of the previous node and a weighted connection between them. We have also studied the effects of various magnetic properties, such as magnet size and anisotropic field on the stochastic operation of individual devices through Monte Carlo simulations of Landau Lifshitz Gilbert (LLG) equation. The three-terminal stochastic devices demonstrated here are a critical step towards building energy efficient spin based neural networks and show the potential for a new application space.

Characterization of Single Defects in Ultrascaled MoS2 Field-Effect Transistors.

MoS2 has received a lot of attention lately as a semiconducting channel material for electronic devices, in part due to its large band gap as compared to that of other 2D materials. Yet, the performance and reliability of these devices are still severely limited by defects which act as traps for charge carriers, causing severely reduced mobilities, hysteresis, and long-term drift. Despite their importance, these defects are only poorly understood. One fundamental problem in defect characterization is that due to the large defect concentration only the average response to bias changes can be measured. On the basis of such averaged data, a detailed analysis of their properties and identification of particular defect types are difficult. To overcome this limitation, we here characterize single defects on MoS2 devices by performing measurements on ultrascaled transistors (∼65 × 50 nm) which contain only a few defects. These single defects are characterized electrically at varying gate biases and temperatures. The measured currents contain random telegraph noise, which is due to the transfer of charge between the channel of the transistors and individual defects, visible only due to the large impact of a single elementary charge on the local electrostatics in these small devices. Using hidden Markov models for statistical analysis, we extract the charge capture and emission times of a number of defects. By comparing the bias-dependence of the measured capture and emission times to the prediction of theoretical models, we provide simple rules to distinguish oxide traps from adsorbates on these back-gated devices. In addition, we give simple expressions to estimate the vertical and energetic positions of the defects. Using the methods presented in this work, it is possible to locate the sources of performance and reliability limitations in 2D devices and to probe defect distributions in oxide materials with 2D channel materials.

Vertical charge transport through transition metal dichalcogenides - a quantitative analysis.

Based on the careful design of two-terminal devices from multi-layer transition metal dichalcogenides (TMDs) such as MoS2 and WSe2, truly vertical transport has been experimentally evaluated and theoretically analyzed. By exploring, the electric field and temperature dependence of in total 28 TMD devices of various thicknesses, a model that describes vertical transport as Fowler Nordheim mediated at high electric fields and thermal injection dominated at low fields has been developed. Our approach is similar to the description chosen to capture gate leakage current levels through amorphous materials such as SiO2. Employing our quantitative analysis, an effective vertical transport mass of m*/m0 (MoS2) ≈ 0.18 and m*/m0 (WSe2) ≈ 0.14 has been extracted for the first time and barriers at the metal contact-to-TMD interface of heights similar to those extracted for lateral transport in TMD transistors have been confirmed.

Understanding contact gating in Schottky barrier transistors from 2D channels.

In this article, a novel two-path model is proposed to quantitatively explain sub-threshold characteristics of back-gated Schottky barrier FETs (SB-FETs) from 2D channel materials. The model integrates the "conventional" model for SB-FETs with the phenomenon of contact gating - an effect that significantly affects the carrier injection from the source electrode in back-gated field effect transistors. The two-path model is validated by a careful comparison with experimental characteristics obtained from a large number of back-gated WSe2 devices with various channel thicknesses. Our findings are believed to be of critical importance for the quantitative analysis of many three-terminal devices with ultrathin body channels.

Vertical versus Lateral Two-Dimensional Heterostructures: On the Topic of Atomically Abrupt p/n-Junctions.

The key appeal of two-dimensional (2D) materials such as graphene, transition metal dichalcogenides (TMDs), or phosphorene for electronic applications certainly lies in their atomically thin nature that offers opportunities for devices beyond conventional transistors. It is also this property that makes them naturally suited for a type of integration that is not possible with any three-dimensional (3D) material, that is, forming heterostructures by stacking dissimilar 2D materials together. Recently, a number of research groups have reported on the formation of atomically sharp p/n-junctions in various 2D heterostructures that show strong diode-type rectification. In this article, we will show that truly vertical heterostructures do exhibit much smaller rectification ratios and that the reported results on atomically sharp p/n-junctions can be readily understood within the framework of the gate and drain voltage response of Schottky barriers that are involved in the lateral transport.

Bandgap Extraction and Device Analysis of Ionic Liquid Gated WSe2 Schottky Barrier Transistors.

Through the careful study of ionic liquid gated WSe2 Schottky barrier field-effect transistors as a function of flake thickness-referred to in the following as body thickness, tbody-critical insights into the electrical properties of WSe2 are gained. One finding is that the inverse subthreshold slope shows a clear dependence on body thickness, i.e., an approximate square root dependent increase with tbody, that provides evidence that injection into the WSe2 channel is mediated by thermally assisted tunneling through the gate-controlled Schottky barriers at the source and drain. By employing our Schottky barrier model, a detailed experimental plot of the WSe2 bandgap as a function of body thickness is obtained. We will discuss why the analysis employed here is critically dependent on the use of the above-mentioned ionic liquid gate and how device characteristics are analyzed in detail.

Covalent Nitrogen Doping and Compressive Strain in MoS2 by Remote N2 Plasma Exposure.

Controllable doping of two-dimensional materials is highly desired for ideal device performance in both hetero- and p-n homojunctions. Herein, we propose an effective strategy for doping of MoS2 with nitrogen through a remote N2 plasma surface treatment. By monitoring the surface chemistry of MoS2 upon N2 plasma exposure using in situ X-ray photoelectron spectroscopy, we identified the presence of covalently bonded nitrogen in MoS2, where substitution of the chalcogen sulfur by nitrogen is determined as the doping mechanism. Furthermore, the electrical characterization demonstrates that p-type doping of MoS2 is achieved by nitrogen doping, which is in agreement with theoretical predictions. Notably, we found that the presence of nitrogen can induce compressive strain in the MoS2 structure, which represents the first evidence of strain induced by substitutional doping in a transition metal dichalcogenide material. Finally, our first principle calculations support the experimental demonstration of such strain, and a correlation between nitrogen doping concentration and compressive strain in MoS2 is elucidated.

Ligand-induced dependence of charge transfer in nanotube-quantum dot heterostructures.

As a model system to probe ligand-dependent charge transfer in complex composite heterostructures, we fabricated double-walled carbon nanotube (DWNT)-CdSe quantum dot (QD) composites. Whereas the average diameter of the QDs probed was kept fixed at ∼4.1 nm and the nanotubes analyzed were similarly oxidatively processed, by contrast, the ligands used to mediate the covalent attachment between the QDs and DWNTs were systematically varied to include p-phenylenediamine (PPD), 2-aminoethanethiol (AET), and 4-aminothiophenol (ATP). Herein, we have put forth a unique compilation of complementary data from experiment and theory, including results from transmission electron microscopy (TEM), near-edge X-ray absorption fine structure (NEXAFS) spectroscopy, Raman spectroscopy, electrical transport measurements, and theoretical modeling studies, in order to fundamentally assess the nature of the charge transfer between CdSe QDs and DWNTs, as a function of the structure of various, intervening bridging ligand molecules. Specifically, we correlated evidence of charge transfer as manifested by changes and shifts associated with NEXAFS intensities, Raman peak positions, and threshold voltages both before and after CdSe QD deposition onto the underlying DWNT surface. Importantly, for the first time ever in these types of nanoscale composite systems, we have sought to use theoretical modeling to justify and account for our experimental results. Our overall data suggest that (i) QD coverage density on the DWNTs varies, based upon the different ligand pendant groups used and that (ii) the presence of a π-conjugated carbon framework within the ligands themselves coupled with the electron affinity of their pendant groups collectively play important roles in the resulting charge transfer from QDs to the underlying CNTs.