Search

The semiconductor-metal junction is one of the most critical factors for high performance electronic devices. In two-dimensional (2D) semiconductor devices, minimizing the voltage drop at this junction is particularly challenging and important. Despite numerous studies concerning contact resistance in 2D semiconductors, the exact nature of the buried interface under a three-dimensional (3D) metal remains unclear. Herein, we report the direct measurement of electrical and optical responses of 2D semiconductor-metal buried interfaces using a recently developed metal-assisted transfer technique to expose the buried interface which is then directly investigated using scanning probe techniques. We characterize the spatially varying electronic and optical properties of this buried interface with < 20 nm resolution. To be specific, potential, conductance and photoluminescence at the buried metal/MoS$_2$ interface are correlated as a function of a variety of metal deposition conditions as well as the type of metal contacts. We observe that direct evaporation of Au on MoS$_2$ induces a large strain of ~5% in the MoS$_2$ which, coupled with charge transfer, leads to degenerate doping of the MoS$_2$ underneath the contact. These factors lead to improvement of contact resistance to record values of 138 kohm-um, as measured using local conductance probes. This approach was adopted to characterize MoS$_2$-In/Au alloy interfaces, demonstrating contact resistance as low as 63 kohm-um. Our results highlight that the MoS$_2$/Metal interface is sensitive to device fabrication methods, and provides a universal strategy to characterize buried contact interfaces involving 2D semiconductors., Comment: 6 figures + supplement

The discovery of intrinsic magnetic topological order in $\rm MnBi_2Te_4$ has invigorated the search for materials with coexisting magnetic and topological phases. These multi-order quantum materials are expected to exhibit new topological phases that can be tuned with magnetic fields, but the search for such materials is stymied by difficulties in predicting magnetic structure and stability. Here, we compute over 27,000 unique magnetic orderings for over 3,000 transition metal oxides in the Materials Project database to determine their magnetic ground states and estimate their effective exchange parameters and critical temperatures. We perform a high-throughput band topology analysis of centrosymmetric magnetic materials, calculate topological invariants, and identify 18 new candidate ferromagnetic topological semimetals, axion insulators, and antiferromagnetic topological insulators. To accelerate future efforts, machine learning classifiers are trained to predict both magnetic ground states and magnetic topological order without requiring first-principles calculations., Comment: 9 pages, 6 figures

Deep learning models have shown impressive performance across a spectrum of computer vision applications including medical diagnosis and autonomous driving. One of the major concerns that these models face is their susceptibility to adversarial attacks. Realizing the importance of this issue, more researchers are working towards developing robust models that are less affected by adversarial attacks. Adversarial training method shows promising results in this direction. In adversarial training regime, models are trained with mini-batches augmented with adversarial samples. Fast and simple methods (e.g., single-step gradient ascent) are used for generating adversarial samples, in order to reduce computational complexity. It is shown that models trained using single-step adversarial training method (adversarial samples are generated using non-iterative method) are pseudo robust. Further, this pseudo robustness of models is attributed to the gradient masking effect. However, existing works fail to explain when and why gradient masking effect occurs during single-step adversarial training. In this work, (i) we show that models trained using single-step adversarial training method learn to prevent the generation of single-step adversaries, and this is due to over-fitting of the model during the initial stages of training, and (ii) to mitigate this effect, we propose a single-step adversarial training method with dropout scheduling. Unlike models trained using existing single-step adversarial training methods, models trained using the proposed single-step adversarial training method are robust against both single-step and multi-step adversarial attacks, and the performance is on par with models trained using computationally expensive multi-step adversarial training methods, in white-box and black-box settings., Comment: CVPR 2020

As humans, we inherently perceive images based on their predominant features, and ignore noise embedded within lower bit planes. On the contrary, Deep Neural Networks are known to confidently misclassify images corrupted with meticulously crafted perturbations that are nearly imperceptible to the human eye. In this work, we attempt to address this problem by training networks to form coarse impressions based on the information in higher bit planes, and use the lower bit planes only to refine their prediction. We demonstrate that, by imposing consistency on the representations learned across differently quantized images, the adversarial robustness of networks improves significantly when compared to a normally trained model. Present state-of-the-art defenses against adversarial attacks require the networks to be explicitly trained using adversarial samples that are computationally expensive to generate. While such methods that use adversarial training continue to achieve the best results, this work paves the way towards achieving robustness without having to explicitly train on adversarial samples. The proposed approach is therefore faster, and also closer to the natural learning process in humans., Comment: CVPR 2020

Two-dimensional (2D) transition metal dichalcogenides (TMDCs) have been the subject of sustained research interest due to their extraordinary electronic and optical properties. They also exhibit a wide range of structural phases because of the different orientations that the atoms can have within a single layer, or due to the ways that different layers can stack. Here we report the first study of direct-visualization of structural transformations in atomically-thin layers under highly non-equilibrium thermodynamic conditions. We probe these transformations at the atomic scale using real-time, aberration corrected scanning transmission electron microscopy and observe strong dependence of the resulting structures and phases on both heating rate and temperature. A fast heating rate (25 C/sec) yields highly ordered crystalline hexagonal islands of sizes of less than 20 nm which are composed of a mixture of 2H and 3R phases. However, a slow heating rate (25 C/min) yields nanocrystalline and sub-stoichiometric amorphous regions. These differences are explained by different rates of sulfur evaporation and redeposition. The use of non-equilibrium heating rates to achieve highly crystalline and quantum-confined features from 2D atomic layers present a new route to synthesize atomically-thin, laterally confined nanostrucutres and opens new avenues for investigating fundamental electronic phenomena in confined dimensions.

The progress in the last decade has enabled machine learning models to achieve impressive performance across a wide range of tasks in Computer Vision. However, a plethora of works have demonstrated the susceptibility of these models to adversarial samples. Adversarial training procedure has been proposed to defend against such adversarial attacks. Adversarial training methods augment mini-batches with adversarial samples, and typically single-step (non-iterative) methods are used for generating these adversarial samples. However, models trained using single-step adversarial training converge to degenerative minima where the model merely appears to be robust. The pseudo robustness of these models is due to the gradient masking effect. Although multi-step adversarial training helps to learn robust models, they are hard to scale due to the use of iterative methods for generating adversarial samples. To address these issues, we propose three different types of regularizers that help to learn robust models using single-step adversarial training methods. The proposed regularizers mitigate the effect of gradient masking by harnessing on properties that differentiate a robust model from that of a pseudo robust model. Performance of models trained using the proposed regularizers is on par with models trained using computationally expensive multi-step adversarial training methods.

Vertically-stacked monolayers of graphene and other atomically-thin 2D materials have attracted considerable research interest because of their potential in fabricating materials with specifically-designed properties. Chemical vapor deposition has proved to be an efficient and scalable fabrication method. However, a lack of mechanistic understanding has hampered efforts to control the fabrication process beyond empirical trial-and-error approaches. In this paper, we develop a general multiscale Burton-Cabrera-Frank (BCF) type model of the vertical growth of 2D materials to predict the necessary growth conditions for vertical versus in-plane (monolayer) growth of arbitrarily-shaped layers. This extends previous work where we developed such a model assuming the layers were fully-faceted (Ye et al., ACS Nano, 11, 12780-12788, 2017). To solve the model numerically, we reformulate the system using the phase-field/diffuse domain method that enables the equations to be solved in a fixed regular domain. We use a second-order accurate, adaptive finite-difference/nonlinear multigrid algorithm to discretize and solve the discrete system. We investigate the effect of parameters, including the van der Waals interaction energies between the layers, the kinetic attachment rates, the edge-energies and the deposition flux, on layer growth and morphologies. While the conditions that favor vertical growth generally follow an analytic thermodynamic criterion we derived for circular layers, the layer boundaries may develop significant curvature during growth, consistent with experimental observations. Our approach provides a mechanistic framework for controlling and optimizing the growth multilayered 2D materials.

Though Deep Neural Networks (DNN) show excellent performance across various computer vision tasks, several works show their vulnerability to adversarial samples, i.e., image samples with imperceptible noise engineered to manipulate the network's prediction. Adversarial sample generation methods range from simple to complex optimization techniques. Majority of these methods generate adversaries through optimization objectives that are tied to the pre-softmax or softmax output of the network. In this work we, (i) show the drawbacks of such attacks, (ii) propose two new evaluation metrics: Old Label New Rank (OLNR) and New Label Old Rank (NLOR) in order to quantify the extent of damage made by an attack, and (iii) propose a new adversarial attack FDA: Feature Disruptive Attack, to address the drawbacks of existing attacks. FDA works by generating image perturbation that disrupt features at each layer of the network and causes deep-features to be highly corrupt. This allows FDA adversaries to severely reduce the performance of deep networks. We experimentally validate that FDA generates stronger adversaries than other state-of-the-art methods for image classification, even in the presence of various defense measures. More importantly, we show that FDA disrupts feature-representation based tasks even without access to the task-specific network or methodology. Code available at: https://github.com/BardOfCodes/fda, Comment: Accepted in ICCV;19. Code Available at https://github.com/BardOfCodes/fda

Bilayer two-dimensional (2D) van der Waals (vdW) materials are attracting increasing attention due to their predicted high quality electronic and optical properties. Here we demonstrate dense, selective growth of WSe2 bilayer flakes by chemical vapor deposition with the use of a 1:10 molar mixture of sodium cholate and sodium chloride as the growth promoter to control the local diffusion of W-containing species. A large fraction of the bilayer WSe2 flakes showed a 0 and 60o twist between the two layers, while moire 15 and 30o-twist angles were also observed. Well-defined monolayer-bilayer junctions were formed in the as-grown bilayer WSe2 flakes, and these interfaces exhibited p-n diode rectification and an ambipolar transport characteristic. This work provides an efficient method for the layer-controlled growth of 2D materials, in particular, 2D transition metal dichalcogenides and promotes their applications in next-generation electronic and optoelectronic devices.

Adversarial samples are perturbed inputs crafted to mislead the machine learning systems. A training mechanism, called adversarial training, which presents adversarial samples along with clean samples has been introduced to learn robust models. In order to scale adversarial training for large datasets, these perturbations can only be crafted using fast and simple methods (e.g., gradient ascent). However, it is shown that adversarial training converges to a degenerate minimum, where the model appears to be robust by generating weaker adversaries. As a result, the models are vulnerable to simple black-box attacks. In this paper we, (i) demonstrate the shortcomings of existing evaluation policy, (ii) introduce novel variants of white-box and black-box attacks, dubbed gray-box adversarial attacks" based on which we propose novel evaluation method to assess the robustness of the learned models, and (iii) propose a novel variant of adversarial training, named Graybox Adversarial Training" that uses intermediate versions of the models to seed the adversaries. Experimental evaluation demonstrates that the models trained using our method exhibit better robustness compared to both undefended and adversarially trained model, Comment: Accepted to ECCV 2018

Only about 8 percent of patients with pancreatic ductal adenocarcinoma, PDAC, live up to 5 years following diagnosis; by 2020 PDAC will become the second most lethal cancer in the United States. PDAC includes an anisotropic fibrous-like stroma, desmoplasia, encompassing most of the tumor mass. Desmoplasia is produced by cancer-associated fibroblasts, CAFs, and their cell-derived extracellular matrices, CDMs. Since elimination of CAFs is detrimental to patients, CDM reprogramming, as opposed to desmoplasia ablation, is therapeutically desirable. In this study we used a human mimetic three-dimensional CAF producing CDM system and proceeded to study its dynamic architectural modifications, following underlying substrate stiffness alterations, using digital imaging analyses, atomic force microscopy, mathematical modeling, cell biology, biochemistry and human tissue quantitative simultaneous multiplex immunofluorescence. Results suggested that the architecture of CDMs can be manipulated to render a tumor-suppressive microenvironment. We posit that perhaps treatments that could reprogram desmoplasia to become tumor-restrictive or that could target tumoral ERK2 might provide future new means for treating PDAC patients., Comment: Main text with 6 figures, 1 table and references, including 5 supplemental figures and legends as well as one supplemental table and a movie legend with corresponding link

Quantum-based molecular dynamics (QMD) is a highly accurate and transferable method for material science simulations. However, the time scales and system sizes accessible to QMD are typically limited to picoseconds and a few hundred atoms. These constraints arise due to expensive self-consistent ground-state electronic structure calculations that can often scale cubically with the number of atoms. Linearly scaling methods depend on computing the density matrix P from the Hamiltonian matrix H by exploiting the sparsity in both matrices. The second-order spectral projection (SP2) algorithm is an O(N) algorithm that computes P with a sequence of 40-50 matrix-matrix multiplications. In this paper, we present task-based implementations of a recently developed data-parallel graph-based approach to the SP2 algorithm, G-SP2. We represent the density matrix P as an undirected graph and use graph partitioning techniques to divide the computation into smaller independent tasks. The partitions thus obtained are generally not of equal size and give rise to undesirable load imbalances in standard MPI-based implementations. This load-balancing challenge can be mitigated by dynamically scheduling parallel computations at runtime using task-based programming models. We develop task-based implementations of the data-parallel G-SP2 algorithm using both Intel's Concurrent Collections (CnC) as well as the Charm++ programming model and evaluate these implementations for future use. Scaling and performance results of our implementations are investigated for representative segments of QMD simulations for solvated protein systems containing more than 10,000 atoms.