Ferroelectric field effect transistors (FeFETs) are among the most promising candidates for the implementation of artificial synapses for neuromorphic computing. However, their scaling poses big challenges, as multilevel behavior is hindered by the two-level or very-few-level polarization behavior reported for devices featuring an aggressive lateral scaling. Moreover, the switching and stabilization of ferroelectric polarization, and consequently the device operation, is inherently linked to high levels of charge trapping in the gate stack, limiting the reliability and increasing the variability of the FeFETs. In this work we investigate, by means of calibrated TCAD simulations, FeFETs integrating a metal interlayer inside the gate stack. FeFETs are designed so that, by properly biasing the device, the metal interlayer can get charged thanks to charge tunneling through the dielectric layer. We first investigate a conventional front-end-of-line implementation, consisting of an aggressively scaled FeFET featuring a ferroelectric hafnium zirconium oxide and a dielectric hafnium oxide in the gate stack, a single-crystal silicon channel, and highly-doped source and drain pockets. Then we consider a back-end-of-line compatible device structure featuring lightly– doped amorphous gallium-oxide channel material, with Schottky-type tungsten contacts. We show that, for both architectures, the charge in the metal interlayer can effectively stabilize the ferroelectric polarization, even for a negligible charge trapping in the dielectric stack. Moreover, the interplay between ferroelectric polarization and charge in the metal interlayer enables a multilevel operation even for a uniform ferroelectric polarization.

High contact resistance remains a central obstacle to the integration of two-dimensional (2D) semiconductors in electronic devices. Recent advances have demonstrated that contact performance can be dramatically improved through interface engineering, including the use of group-V semimetals and charge-transfer contacts based on strong interfacial doping. Here, we show that controlled interfacial oxidation provides an effective route to convert a semimetal contact into a charge-transfer contact that degenerately n-dopes single layer MoS_2. Using a combination of angle-resolved photoemission spectroscopy, X-ray photoelectron diffraction, low-energy electron diffraction and scanning tunnelling spectroscopy, we demonstrate that putting single layer MoS_2 in contact with a pristine Bi layer merely results in weak doping, whereas oxidation of the Bi layer leads to a pronounced occupation of the MoS_2 conduction band with an electron density on the order of 10^13 cm^-2. The cause of this strong electron doping is the fact that an ultrathin β-Bi_2O_3 layer forms below the MoS_2 and that this has a particularly low work function, thereby acting as an efficient electron donor to MoS_2. Interfacial oxidation thus emerges as a powerful design knob for engineering charge-transfer contacts to 2D semiconductors.

Multi-level operation, conventionally obtained in ferroelectric devices thanks to a domain-dependent inhomogeneous polarization, poses a big challenge for highly-scaled ferroelectric devices, where the number of ferroelectric domains is drastically reduced. In this work, we study a highly scaled back-end-of-line (BEOL) compatible, ferroelectric field-effect transistor (FeFET) that integrates a metal interlayer in the gate stack. Through analytical models and calibrated TCAD simulations, we show how this device can achieve a multi-level operation exploiting the interplay between the ferroelectric polarization and the charge in the metal interlayer. Such a working principle does not rely on a domain-dependent inhomogeneous polarization, and the device operation is thus ensured also for a homogeneous ferroelectric material. We also demonstrate that the charge in the interlayer can effectively stabilize the ferroelectric polarization even in the absence of a high concentration of trapped charges in the gate stack. The potentiation and depression curves for the readout conductance confirm that the proposed device can be operated as a memristor for neuromorphic computing applications. Moreover, we show how the choice of the dielectric in the metal-ferroelectric-dielectric-metal gate stack can be used as a design knob to reduce the device operation voltage.

We present an ab-inito transport study, based on the Density Functional Theory (DFT) and the non-equilibrium Green's Functions (NEGF) formalism, to assess the effect of the Image Force Barrier Lowering (IFBL) on the source-tochannel resistance ($R_{\mathbf{C}}$) in metal-MoS2 top contacts. To compute the IFBL potential profile, our model relies on the numerical solution of the 3D Poisson equation for the point charge in a heterogeneous dielectric environment. By considering Al-MoS2 and Bi-MoS2 van-der-Waals heterojunctions, representative of a Schottky and an Ohmic contact respectively, we show that, while for Schottky contacts the inclusion of the IFBL has a vast impact on the source-to-channel resistance, in quasi-Ohmic contacts the influence of the IFBL on the $R_{\mathbf{C}}$ becomes negligible, especially at large inversion densities.

We present a critical reexamination of the Landau-Ginzburg-Devonshire (LGD) model for ferroelectric materials that is based on intrinsic nucleation events. Theoretical considerations and a systematic comparison with experiments steered us towards a novel version and calibration of the LGD model relying instead on extrinsic nucleations. We show that the new model can not only improve the agreement with experiments, but also help reconcile the interpretation of polarization reversal in poly-crystalline and epitaxial ferroelectrics.

This work presents a simulation study of a ferroelectric field effect transistor (FeFET), which leverages a metal interlayer to achieve a multilevel operation thanks to the interplay between the ferroelectric polarization and the charge stored in the interlayer. We show that the metal interlayer can effectively stabilize the ferroelectric polarization even for a negligible charge trapping in the dielectric stack and, moreover, enable a multilevel operation even for a uniform ferroelectric polarization.

The progress of biologically inspired neuromorphic computing hardware in the last decade has been fostered also by the advancement in CMOS-compatible memristors, providing a nonvolatile storage of multiconductance states mimicking the synaptic weights in biological systems. This article is instead focused on the less-explored field of memcapacitors (MemCaps), which only very recently has attracted a renewed interest, and it is based on devices capable of tuning their capacitance. In particular, we present by means of extensive numerical simulations carefully calibrated against experimental data the operation of a two-terminal ferroelectric MemCap exhibiting multilevel, polarization-dependent capacitance values. The MemCap exploits a ferroelectric gated-diode structure, and it is thus fully compatible with CMOS processing. Our results show that multilevel operation is viable using properly shaped pulse trains at the gate terminal, and moreover, a nondestructive readout can be achieved by means of small-amplitude ac signals.

By using a modelling framework consisting of NEGF-based ab-initio simulations, we investigate the operation and design of Dirac-Source FETs (DSFETs). First, we discuss some methodological aspects that we found pivotal to observe the sub-60mV/dec swing in ab-initio simulations. Then, we propose the novel HGr-DSFET based on hydro-genated graphene (HGr). We show that the HGr-DSFET has a robust sub-60mV/dec operation and an ON current between $2\times$ and $4\times$ larger than the graphene-MoS2 DSFET counterpart. Our study addresses and elucidates several physical and design aspects of DSFETs, including the rethermalization effects due to phonon scattering.

By using an in-house nonequilibrium Green's function (NEGF)-based ab initio simulator, we investigate the physical mechanisms driving the Sb(01 (1) over bar2)-MoS2 system to exhibit the lowest reported contact resistance, RC = 42 Omega center dot mu m, to the 2-D semiconductor MoS2. We can find that the transport from the hybridized bands in the Sb-MoS2 heterojunction is quite ineffective and that the back-gate-induced doping of MoS2 in the contact region is crucial to explain the experiments. In fact, by accounting in our ab initio simulations for the presence of a back gate according to the experiments, it is possible to match the band structure of the MoS2 in the Sb-MoS2 heterojunction with that of the external MoS2 layer, which drastically increases the electronic transmission throughout the contact, and ultimately pushes RC close to the quantum limit. Furthermore, we extend the applicability of our previously demonstrated simulation methodology and thus investigate a field-effect transistors (FETs)-like device including an ab initio description of the carrier injection at the Sb-MoS2 contact.

??High contact resistance (RC) between 3D metallic conductors and single-layer 2D semiconductors poses major challenges toward their integration in nanoscale electronic devices. While in experiments the large RC values can be partly due to defects, ab initio simulations suggest that, even in defect-free structures, the interaction between metal and semiconductor orbitals can induce gap states that pin the Fermi level in the semiconductor band gap, increase the Schottky barrier height (SBH), and thus degrade the contact resistance. In this paper, we investigate, by using an in-house-developed ab initio transport methodology that combines density functional theory and nonequilibrium Green???s function (NEGF) transport calculations, the physical properties and electrical resistance of several options for ntype top metal contacts to monolayer MoS2, even in the presence of buffer layers, and for ptype contacts to monolayer WSe2. The delicate interplay between the SBH and tunneling barrier thickness is quantitatively analyzed, confirming the excellent properties of the Bi??? MoS2 system as an n-type ohmic contact. Moreover, simulation results supported by literature experiments suggest that the Au???WSe2 system is a promising candidate for p-type ohmic contacts. Finally, our analysis also reveals that a small modulation of a few angstroms of the distance between the (semi)metal and the transition-metal dichalcogenide (TMD) leads to large variations of RC. This could help to explain the scattering of RC values experimentally reported in the literature because different metal deposition techniques can result in small changes of the metal-to-TMD distance besides affecting the density of possible defects.

This paper provides a brief introduction to the phenomenological aspects of the polarization in ferrroelectric materials, and then an analysis of a few selected topics related to the modelling of ferroelectrics. The description of ferroelectric-based devices is quite challenging, particularly because the ferroelectric is frequently stacked with other dielectrics or with a semiconductor, as opposed to being placed between metal electrodes. Predictive modelling of ferroelectric devices is admittedly difficult, and thus the scrutiny and calibration of the models by comparison to sound experimental data is of paramount importance.

The interaction between a Ni(111) substrate covered by a complete Gr monolayer and H atoms occurs through two parallel routes leading to the hydrogen chemisorption on graphene and, at much lower rate but still with some ease, to the intercalation of H atoms below it. This latter reaction determines a direct interaction of the H atoms with the metal surface and eventually the H diffusion into the Ni bulk under the Gr cover. In this study we have combined high-resolution X-ray photoelectron spectroscopy, thermal programmed desorption, and density functional theory calculations to establish how the chemisorption and intercalation yields and their interplay depend on temperature and to find out how graphene affects the amount and the evolution of the hydrogen diffused in the Ni bulk. We found that between 150 and 320 K, hydrogen chemisorption on Gr is independent of temperature and that Gr lifting, which signals the H intercalation below it, does not occur below 180-200 K, being limited by an energy barrier of the order of 150 meV. For the heavily hydrogenated samples, when H atoms diffuse also into the Ni bulk, the Gr cover plays a key role for H storage because it strongly enhances the amount of H loaded in the interface with respect to the bare Ni(111) substrate. This behavior, possibly exhibited also by other graphene/metal interfaces provided that intercalation of H below graphene can readily occur, might foster the design of innovative materials to be applied for H storage.

This paper investigates and compares through a comprehensive TCAD analysis 2D and 3D simulations for ferroelectric based FETs. We provide clear evidence that the multiple read conductance values experimentally observed in FeFETs stem from source to drain percolation current paths, which are governed by the polarization patterns in the ferroelectric domains. Such a physical picture makes 3D simulations indispensable to capture even the qualitative features of the device behaviour, not to mention the quantitative aspects.

We here report a joint experimental and simulation analysis for large signal P-V and AC small-signal C-V curves in ferroelectric tunnel junctions. The attempt to reproduce both experimental data sets with the same model and material parameters challenges our understanding of the underlying physics, but it also helps develop a sound background for the device design.

An accurate estimate of the ferroelectric polarization in ferroelectric-dielectric stacks is important from a materials science perspective, and it is also crucial for the development of ferroelectric based electron devices.This paper revisits the theory and application of the PUND technique in Metal-Ferroelectric-Dielectric-Metal (MFDM) structures by using analytical derivations and numerical simulations.In an MFDM structure the results of the PUND technique may largely differ from the polarization actually switched in the stack, which in turn is different from the remnant polarization of the underlying ferroelectric.The main hindrances that prevent PUND measurements from providing a good estimate of the polarization switching in MFDM stacks are thus discussed.The inspection of the involved physical quantities, not always accessible in experiments, provides a useful insight about the main sources of the errors in the PUND technique, and clarifies the delicate interplay between the depolarization field and the charge injection and trapping in MFDM stacks with a thin dielectric layer.

Fermi level pinning (FLP) in metal-MoS2 contacts induces large Schottky barrier heights which in turn results large contact resistances. In this work, we made use of Density Functional Theory (DFT) to study the origin in MoS2 contacts with a variety of metals. We also reported how the Fermi level de-pinning could be attained controlling the distance between the metal and MoS2. In this respect, the metal-MoS2 contacts can be engineered by means of the insertion of proper buffer layers and the use of back-gated structures. This results in a practically zeroed Schottky barrier heights for some specific metal-MoS2 stacks, which it is crucial to attain Ohmic with low series resistances.

By combining DFT calculations and an in-house developed, ab-initio transport simulator, we investigate the transport properties of n-type contacts between 3D metallic conductors and monolayer MoS2. Moreover, the impact of buffer layers between the metal and the MoS2 is also analyzed in order to reduce the Schottky barrier and approach the Ohmic behavior of contacts.

We investigate several options for metal contacts to monolayer 2D semiconductors with an in-house developed, ab-initio transport methodology. We identify an optimum separation between the metal and the semiconductor resulting in minimum contact resistance $(R_{\mathrm{C}})$. Such a minimum $R_{\mathrm{C}}$ is fundamentally related to the trade-off between Schottky barrier height and tunneling barrier width. We examine quantitatively the effects of buffer layers and report, for the first time, ab-initio R calculations for the $\mathrm{b}\mathrm{i}\mathrm{s}\mathrm{m}\mathrm{u}\mathrm{t}\mathrm{h}-\mathrm{M}\mathrm{o}\mathrm{S}_{2}$ system, comprising the influence of the inelastic electron-phonon interaction. Finally, we investigate $\mathrm{A}\mathrm{u}-\mathrm{W}\mathrm{S}\mathrm{e}_{2}$ as a material system for a low $R_{\mathrm{C}}$ -type contact.

Hybrid materials composed by porous Ni foams coated by graphene appear appealing architectures to be applied in the field of hydrogen storage, where, due to the high surface-to-volume ratio of both components, are expected to be much more performant than their flat counterparts. In order to explore this possibility, in this study we have grown single layer graphene on nickel foams by chemical vapour deposition in ultra-high vacuum and have investigated the interaction with H atoms as a function of the temperature. By using high resolution C1s core level spectroscopy we found that nearly half of the graphene layer interacts with the support almost as strongly as with the flat ordered Ni substrate, whereas the other half is nearly free standing. Such dual electronic and structural coupling drives the hydrogenation of the graphene/foam interface. By using thermal programmed desorption combined with x-ray photoelectron spectroscopy we found that in the weakly interacting graphene regions, even at very low temperatures, H atoms easily intercalate below graphene, and enter in the bulk of the foam, from where they start to desorb around 180 K. This behaviour mimics what happens when dosing H atoms on the bare Ni foam. On the contrary, H intercalation below the strongly interacting graphene regions occurs only for temperatures around and above 200 K. The thermal desorption curves demonstrated that the presence of the graphene layer does not reduce the effectiveness of H loading in the bulk of the foam. On the other hand, it does not even increase significantly the stored amount with respect to the uncoated support, but contributes to stabilizing the stored hydrogen. Although the fundamental aspects of graphene/foam hydrogenation were here investigated in a regime far below the saturation of the bulk absorption, these measurements can be the starting point for further investigations aimed at establishing the ultimate storage capability of these hybrid nano -structured tanks.

Time-, k-, and energy-resolved photoemission data for C 1s in H-intercalated graphene on SiC (QFMLG). Here we use a free electron laser to investigate the effect of a transiently excited hot electron gas on the core level spectrum of graphene, showing that it leads to a substantial broadening of the C 1s line shape.