Transfer Learning of Full Molecular Weight Distributions via High-Throughput Computer-Controlled Polymerization.

The skew and shape of the molecular weight distribution (MWD) of polymers have a significant impact on polymer physical properties. Standard summary metrics statistically derived from the MWD only provide an incomplete picture of the polymer MWD. Machine learning (ML) methods coupled with high-throughput experimentation (HTE) could potentially allow for the prediction of the entire polymer MWD without information loss. In our work, we demonstrate a computer-controlled HTE platform that is able to run up to 8 unique variable conditions in parallel for the free radical polymerization of styrene. The segmented-flow HTE system was equipped with an inline Raman spectrometer and offline size exclusion chromatography (SEC) to obtain time-dependent conversion and MWD, respectively. Using ML forward models, we first predict monomer conversion, intrinsically learning varying polymerization kinetics that change for each experimental condition. In addition, we predict entire MWDs including the skew and shape as well as SHAP analysis to interpret the dependence on reagent concentrations and reaction time. We then used a transfer learning approach to use the data from our high-throughput flow reactor to predict batch polymerization MWDs with only three additional data points. Overall, we demonstrate that the combination of HTE and ML provides a high level of predictive accuracy in determining polymerization outcomes. Transfer learning can allow exploration outside existing parameter spaces efficiently, providing polymer chemists with the ability to target the synthesis of polymers with desired properties.

Self-Selective Multi-Terminal Memtransistor Crossbar Array for In-Memory Computing.

Two-terminal resistive switching devices are commonly plagued with longstanding scientific issues including interdevice variability and sneak current that lead to computational errors and high-power consumption. This necessitates the integration of a separate selector in a one-transistor-one-RRAM (1T-1R) configuration to mitigate crosstalk issue, which compromises circuit footprint. Here, we demonstrate a multi-terminal memtransistor crossbar array with increased parallelism in programming via independent gate control, which allows in situ computation at a dense cell size of 3-4.5 F2 and a minimal sneak current of 0.1 nA. Moreover, a low switching energy of 20 fJ/bit is achieved at a voltage of merely 0.42 V. The architecture is capable of performing multiply-and-accumulate operation, a core computing task for pattern classification. A high MNIST recognition accuracy of 96.87% is simulated owing to the linear synaptic plasticity. Such computing paradigm is deemed revolutionary toward enabling data-centric applications in artificial intelligence and Internet-of-things.

Deep learning-enabled prediction of 2D material breakdown.

Characterizing electrical breakdown limits of materials is a crucial step in device development. However, methods for repeatable measurements are scarce in two-dimensional materials, where breakdown studies have been limited to destructive methods. This restricts our ability to fully account for variability in local electronic properties induced by surface contaminants and the fabrication process. To tackle this, we implement a two-step deep-learning model to predict the breakdown mechanism and breakdown voltage of monolayer MoS2devices with varying channel lengths and resistances using current measured in the low-voltage regime as inputs. A deep neural network (DNN) first classifies between Joule and avalanche breakdown mechanisms using partial current traces from 0 to 20 V. Following this, a convolutional long short-term memory network (CLSTM) predicts breakdown voltages of these classified devices based on partial current traces. We test our model with electrical measurements collected using feedback-control of the applied voltage to prevent device destruction, and show that the DNN classifier achieves an accuracy of 79% while the CLSTM model has a 12% error when requiring only 80% of the current trace as inputs. Our results indicate that information encoded in the current behavior far from the breakdown point can be used for breakdown predictions, which will enable non-destructive and rapid material characterization for 2D material device development.

Layer Rotation-Angle-Dependent Excitonic Absorption in van der Waals Heterostructures Revealed by Electron Energy Loss Spectroscopy.

Heterostructures comprising van der Waals (vdW) stacked transition metal dichalcogenide (TMDC) monolayers are a fascinating class of two-dimensional (2D) materials. The presence of interlayer excitons, where the electron and the hole remain spatially separated in the two layers due to ultrafast charge transfer, is an intriguing feature of these heterostructures. The optoelectronic functionality of 2D heterostructure devices is critically dependent on the relative rotation angle of the layers. However, the role of the relative rotation angle of the constituent layers on intralayer absorption is not clear yet. Here, we investigate MoS2/WSe2 vdW heterostructures using monochromated low-loss electron energy loss (EEL) spectroscopy combined with aberration-corrected scanning transmission electron microscopy and report that momentum conservation is a critical factor in the intralayer absorption of TMDC vdW heterostructures. The evolution of the intralayer excitonic low-loss EEL spectroscopy peak broadenings as a function of the rotation angle reveals that the interlayer charge transfer rate can be about an order of magnitude faster in the aligned (or anti-aligned) case than in the misaligned cases. These results provide a deeper insight into the role of momentum conservation, one of the fundamental principles governing charge transfer dynamics in 2D vdW heterostructures.

Carrier control in 2D transition metal dichalcogenides with Al2O3 dielectric.

We report transport measurements of dual gated MoS2 and WSe2 devices using atomic layer deposition grown Al2O3 as gate dielectrics. We are able to achieve current pinch-off using independent split gates and observe current steps suggesting possible carrier confinement. We also investigated the impact of gate geometry and used electrostatic potential simulations to explain the observed device physics.

Direct n- to p-Type Channel Conversion in Monolayer/Few-Layer WS2 Field-Effect Transistors by Atomic Nitrogen Treatment.

We present a method for substitutional p-type doping in monolayer (1L) and few-layer (FL) WS2 using highly reactive nitrogen atoms. We demonstrate that the nitrogen-induced lattice distortion in atomically thin WS2 is negligible due to its low kinetic energy. The electrical characteristics of 1L/FL WS2 field-effect transistors (FETs) clearly show an n-channel to p-channel conversion with nitrogen incorporation. We investigate the defect formation energy and the origin of p-type conduction using first-principles calculations. We reveal that a defect state appears near the Fermi level, leading to a shallow acceptor level at 0.24 eV above the valence band maximum in nitrogen-doped 1L/FL WS2. This doping strategy enables a substitutional p-type doping in intrinsically n-type 1L/FL transition metal dichalcogenides (TMDCs) with tunable control of dopants, offering a method for realizing complementary metal-oxide-semiconductor FETs and optoelectronic devices on 1L/FL TMDCs by overcoming one of the major limits of TMDCs, that is, their n-type unipolar conduction.

Modification of Vapor Phase Concentrations in MoS2 Growth Using a NiO Foam Barrier.

Single-layer molybdenum disulfide (MoS2) has attracted significant attention due to its electronic and physical properties, with much effort invested toward obtaining large-area high-quality monolayer MoS2 films. In this work, we demonstrate a reactive-barrier-based approach to achieve growth of highly homogeneous single-layer MoS2 on sapphire by the use of a nickel oxide foam barrier during chemical vapor deposition. Due to the reactivity of the NiO barrier with MoO3, the concentration of precursors reaching the substrate and thus nucleation density is effectively reduced, allowing grain sizes of up to 170 µm and continuous monolayers on the centimeter length scale being obtained. The quality of the monolayer is further revealed by angle-resolved photoemission spectroscopy measurement by observation of a very well resolved electronic band structure and spin-orbit splitting of the bands at room temperature with only two major domain orientations, indicating the successful growth of a highly crystalline and well-oriented MoS2 monolayer.

Electronic properties of atomically thin MoS2 layers grown by physical vapour deposition: band structure and energy level alignment at layer/substrate interfaces.

We present an analysis of the electronic properties of an MoS2 monolayer (ML) and bilayer (BL) as-grown on a highly ordered pyrolytic graphite (HOPG) substrate by physical vapour deposition (PVD), using lab-based angle-resolved photoemission spectroscopy (ARPES) supported by scanning tunnelling microscopy (STM) and X-ray photoelectron spectroscopy (XPS) for morphology and elemental assessments, respectively. Despite the presence of multiple domains (causing in-plane rotational disorder) and structural defects, electronic band dispersions were clearly observed, reflecting the high density of electronic states along the high symmetry directions of MoS2 single crystal domains. In particular, the thickness dependent direct-to-indirect band gap transition previously reported only for MoS2 layers obtained by exfoliation or via epitaxial growth processes, was found to be also accessible in our PVD grown MoS2 samples. At the same time, electronic gap states were detected, and attributed mainly to structural defects in the 2D layers. Finally, we discuss and clarify the role of the electronic gap states and the interlayer coupling in controlling the energy level alignment at the MoS2/substrate interface.

Electronic Properties of a 1D Intrinsic/p-Doped Heterojunction in a 2D Transition Metal Dichalcogenide Semiconductor.

Two-dimensional (2D) semiconductors offer a convenient platform to study 2D physics, for example, to understand doping in an atomically thin semiconductor. Here, we demonstrate the fabrication and unravel the electronic properties of a lateral doped/intrinsic heterojunction in a single-layer (SL) tungsten diselenide (WSe2), a prototype semiconducting transition metal dichalcogenide (TMD), partially covered with a molecular acceptor layer, on a graphite substrate. With combined experiments and theoretical modeling, we reveal the fundamental acceptor-induced p-doping mechanism for SL-WSe2. At the 1D border between the doped and undoped SL-WSe2 regions, we observe band bending and explain it by Thomas-Fermi screening. Using atomically resolved scanning tunneling microscopy and spectroscopy, the screening length is determined to be in the few nanometer range, and we assess the carrier density of intrinsic SL-WSe2. These findings are of fundamental and technological importance for understanding and employing surface doping, for example, in designing lateral organic TMD heterostructures for future devices.

Functionalization of 2D transition metal dichalcogenides for biomedical applications.

Recent research has revealed a gamut of interesting properties present in layered two-dimensional (2D) transition metal dichalcogenides (TMDCs) such as photoluminescence, comparatively high electron mobility, flexibility, mechanical strength and relatively low toxicity. The large surface to area ratio inherent in these materials also allows easy functionalization and maximal interaction with the external environment. Due to its unique physical and chemical properties, much work has been done in tailoring TMDCs through chemical functionalization for use in a diverse range of biomedical applications as biosensors, drug delivery carriers or even as therapeutic agents. In this review, current progress on the different types of TMDC functionalization for various biological applications will be presented and its future outlook will be discussed.

Competition between Hexagonal and Tetragonal Hexabromobenzene Packing on Au(111).

Low-temperature scanning tunneling microscope investigations reveal that hexabromobenzene (HBB) molecules arrange in either hexagonally closely packed (hcp) [Formula: see text] or tetragonal [Formula: see text] structure on Au(111) dependent on a small substrate temperature difference around 300 K. The underlying mechanism is investigated by density functional theory calculations, which reveal that substrate-mediated intermolecular noncovalent C-Br···Br-C attractions induce hcp HBB islands, keeping the well-known Au(111)-22×√3 reconstruction intact. Upon deposition at 330 K, HBB molecules trap freely diffusing Au adatoms to form tetragonal islands. This enhances the attraction between HBB and Au(111) but partially reduces the intermolecular C-Br···Br-C attractions, altering the Au(111)-22×√3 reconstruction. In both cases, the HBB molecule adsorbs on a bridge site, forming a ∼15° angle between the C-Br direction and [112̅]Au, indicating the site-specific molecule-substrate interactions. We show that the competition between intermolecular and molecule-substrate interactions determines molecule packing at the subnanometer scale, which will be helpful for crystal engineering, functional materials, and organic electronics.

Incorporating isolated molybdenum (Mo) atoms into bilayer epitaxial graphene on 4H-SiC(0001).

The atomic structures and electronic properties of isolated Mo atoms in bilayer epitaxial graphene (BLEG) on 4H-SiC(0001) are investigated by low temperature scanning tunneling microscopy (LT-STM). LT-STM results reveal that isolated Mo dopants prefer to substitute C atoms at α-sites and preferentially locate between the graphene bilayers. First-principles calculations confirm that the embedding of single Mo dopants within BLEG is energetically favorable as compared to monolayer graphene. The calculated band structures show that Mo-incorporated BLEG is n-doped, and each Mo atom introduces a local magnetic moment of 1.81 µB into BLEG. Our findings demonstrate a simple and stable method to incorporate single transition metal dopants into the graphene lattice to tune its electronic and magnetic properties for possible use in graphene spin devices.

Spatially resolved electronic structures of atomically precise armchair graphene nanoribbons.

Graphene has attracted much interest in both academia and industry. The challenge of making it semiconducting is crucial for applications in electronic devices. A promising approach is to reduce its physical size down to the nanometer scale. Here, we present the surface-assisted bottom-up fabrication of atomically precise armchair graphene nanoribbons (AGNRs) with predefined widths, namely 7-, 14- and 21-AGNRs, on Ag(111) as well as their spatially resolved width-dependent electronic structures. STM/STS measurements reveal their associated electron scattering patterns and the energy gaps over 1 eV. The mechanism to form such AGNRs is addressed based on the observed intermediate products. Our results provide new insights into the local properties of AGNRs, and have implications for the understanding of their electrical properties and potential applications.

Trapping single polar molecules in SiC nanomesh via out-of-plane dipoles.

The self-assembly of nonplanar chloroaluminum phthalocyanine (ClAlPc) molecules as well-ordered single-molecule dipole arrays on the silicon carbide (SiC) nanomesh substrate was investigated using low temperature scanning tunneling microscopy. ClAlPc exclusively adsorbs in the center of the SiC nanomesh holes with its inherent dipole (from Cl to Al) pointing toward the substrate. The dipole can be inverted by a positively biased tip with a threshold tip voltage of 3.3 V. We deduce that the interaction between the intrinsic dipole of ClAlPc and the periodic out-of-plane component of the surface dipole on the SiC nanomesh plays a significant role in the dipole array formation.

Energy-gap opening in a Bi110 nanoribbon induced by edge reconstruction.

Scanning tunnelling microscopy and spectroscopy experiments complemented by first-principles calculations have been conducted to study the electronic structure of 4 monolayer Bi(110) nanoribbons on epitaxial graphene on silicon carbide [4H-SiC(0001)]. In contrast with the semimetal property of elemental bismuth, an energy gap of 0.4 eV is measured at the centre of the Bi(110) nanoribbons. Edge reconstructions, which can facilitate the edge strain energy release, are found to be responsible for the band gap opening. The calculated density of states around the Fermi level are decreased quickly to zero from the terrace edge to the middle of a Bi(110) nanoribbon potentially signifying a spatial metal-to-semiconductor transition. This study opens new avenues for room-temperature bismuth nanoribbon-based electronic devices.

Quasi-free-standing epitaxial graphene on SiC (0001) by fluorine intercalation from a molecular source.

We demonstrated a novel method to obtain charge neutral quasi-free-standing graphene on SiC (0001) from the buffer layer using fluorine from a molecular source, fluorinated fullerene (C(60)F(48)). The intercalated product is stable under ambient conditions and resistant to elevated temperatures of up to 1200 °C. Scanning tunneling microscopy and spectroscopy measurements are performed for the first time on such quasi-free-standing graphene to elucidate changes in the electronic and structural properties of both the graphene and interfacial layer. Novel structures due to a highly localized perturbation caused by the presence of adsorbed fluorine were produced in the intercalation process and investigated. Photoemission spectroscopy is used to confirm these electronic and structural changes.