is an order magnitude larger than previously thought, yet near the low end of known solidsolid interfaces. Our study also reveals unexpected insight into non-uniformities of the MoS2 transistors (small bilayer regions), which do not cause significant self-heating, suggesting that such semiconductors are less sensitive to inhomogeneity than expected. These results provide key insights into energy dissipation of 2D semiconductors and pave the way for the future design of energy-efficient 2D electronics. Keywords: Energy dissipation, 2D semiconductors, thermal boundary conductance, Raman thermometry, MoS2 2The performance of nanoelectronics is most often constrained by thermal challenges, 1, 2 memory bottlenecks, 3 and nanoscale contacts. 4 The former have become particular acute, with high integration densities leading to high power density, and numerous interfaces (e.g. between silicon, copper, SiO2) leading to high thermal resistance. New applications and new form-factors call for dense vertical integration into multi-layer "high-rise" processors for high-performance computing, 3 or integration with poor thermal substrates like flexible plastics (of thermal conductivity 5xlower than SiO2 and nearly 500x lower than silicon) for wearable computing. 5 These are the two most likely platforms for incorporating 2D semiconductors into electronics, yet very little is known about fundamental limits or practical implications of energy dissipation in these contexts.At its most basic level, energy dissipation begins in the ultra-thin transistor channel and is immediately limited by the insulating regions and thermal resistance with the interfaces surrounding it. Herbert Kroemer's observation 6 that "the interface is the device" is remarkably aptfor 2D semiconductors such as monolayer MoS2. These have no bulk, and are thus strongly limited by their interfaces. For instance, even some of the best electrical contacts known today add >50% parasitic resistance to MoS2 transistors when these are scaled to sub-100 nm dimensions. Similarly, thermal interfaces may be expected to limit energy dissipation from 2D electronics, and their understanding is essential. Nevertheless, a key challenge is the need to differentiate heating of the sub-nanometer thin 2D material from its environment. Here, Raman spectroscopy holds a unique advantage, 8, 9 as the temperature of even a monolayer semiconductor can be distinguished from the material directly under (or above) it, if the Raman signatures are distinct. 10Figure 1a shows our typical device structure and measurement setup. We utilize high-qual- Minor, randomly distributed non-uniformities in the temperature seen in Figure 2 are within the uncertainty of the measurement and are also visible in the reference map taken at VDS = 0 (on a hot stage), for which the temperature is known to be uniform, as shown in Supporting Information Figure S4. The uniform self-heating of transistors from CVD-grown MoS2 suggests that any change in energy dissipation around the 2L spots or other non-uniformit...
Heterogeneous integration of nanomaterials has enabled advanced electronics and photonics applications. However, similar progress has been challenging for thermal applications, in part due to shorter wavelengths of heat carriers (phonons) compared to electrons and photons. Here, we demonstrate unusually high thermal isolation across ultrathin heterostructures, achieved by layering atomically thin two-dimensional (2D) materials. We realize artificial stacks of monolayer graphene, MoS2, and WSe2 with thermal resistance greater than 100 times thicker SiO2 and effective thermal conductivity lower than air at room temperature. Using Raman thermometry, we simultaneously identify the thermal resistance between any 2D monolayers in the stack. Ultrahigh thermal isolation is achieved through the mismatch in mass density and phonon density of states between the 2D layers. These thermal metamaterials are an example in the emerging field of phononics and could find applications where ultrathin thermal insulation is desired, in thermal energy harvesting, or for routing heat in ultracompact geometries.
We present our latest advancements of machine-learned potentials (MLPs) based on the neuroevolution potential (NEP) framework introduced in [Fan et al., Phys. Rev. B 104, 104309 (2021)] and their implementation in the open-source package GPUMD.We increase the accuracy of NEP models both by improving the radial functions in the atomic-environment descriptor using a linear combination of Chebyshev basis functions and by extending the angular descriptor with some four-body and five-body contributions as in the atomic cluster expansion approach.We also detail our efficient implementation of the NEP approach in graphics processing units as well as our workflow for the construction of NEP models, and we demonstrate their application in large-scale atomistic simulations.By comparing to state-of-the-art MLPs, we show that the NEP approach not only achieves above-average accuracy but also is far more computationally efficient.These results demonstrate that the GPUMD package is a promising tool for solving challenging problems requiring highly accurate, large-scale atomistic simulations.To enable the construction of MLPs using a minimal training set, we propose an active-learning scheme based on the latent space of a pre-trained NEP model.Finally, we introduce three separate Python packages, GPYUMD, CALORINE, and PYNEP, which enable the integration of GPUMD into Python workflows.
Thermal properties of molybdenum disulfide (MoS2) have recently attracted attention related to fundamentals of heat propagation in strongly anisotropic materials, and in the context of potential applications to optoelectronics and thermoelectrics. Multiple empirical potentials have been developed for classical molecular dynamics (MD) simulations of this material, but it has been unclear which provides the most realistic results. Here, we calculate lattice thermal conductivity of singleand multi-layer pristine MoS2 by employing three different thermal transport MD methods: equilibrium, nonequilibrium, and homogeneous nonequilibrium ones. These methods allow us to verify the consistency of our results and also facilitate comparisons with previous works, where different schemes have been adopted. Our results using variants of the Stillinger-Weber potential are at odds with some previous ones and we analyze the possible origins of the discrepancies in detail. We show that, among the potentials considered here, the reactive empirical bond order (REBO) potential gives the most reasonable predictions of thermal transport properties as compared to experimental data. With the REBO potential, we further find that isotope scattering has only a small effect on thermal conduction in MoS2 and the in-plane thermal conductivity decreases with increasing layer number and saturates beyond about three layers. We identify the REBO potential as a transferable empirical potential for MD simulations of MoS2 which can be used to study thermal transport properties in more complicated situations such as in systems containing defects or engineered nanoscale features. This work establishes a firm foundation for understanding heat transport properties of MoS2 using MD simulations.
Transistors based on two-dimensional (2D) materials often exhibit hysteresis in their electrical measurements, i.e. a dependence of measured current on voltage sweep direction due to charge trapping. Here we demonstrate a simple pulsed measurement technique which reduces this hysteretic behavior, enabling more accurate characterization of 2D transistors. We compare hysteresis and charge trapping in four types of devices fabricated from both exfoliated and synthetic MoS 2 , with SiO 2 and HfO 2 insulators, using DC and pulsed voltage measurements at different temperatures. Applying modest voltage pulses (~1 ms) on the gate significantly reduces charge trapping and results in the elimination of over 80% of hysteresis for all devices. At shorter pulse widths (~1 µs), up to 99% of hysteresis is reduced for some devices. Our measurements enable the extraction of a unique value of field-effect mobility, regardless of voltage sweep direction, unlike measurements that rely on forward or backward DC measurements. This simple and reproducible technique is useful for studying the intrinsic properties of 2D transistors, and can be similarly applied to other nanoscale and emerging devices where charge trapping is of concern. LETTER RECEIVED
Understanding the thermal properties of two-dimensional (2D) materials and devices is essential for thermal management of 2D applications. Here we perform molecular dynamics simulations to evaluate both the specific heat of MoS2 as well as the thermal boundary conductance (TBC) between one to five layers of MoS2 with amorphous SiO2 and between single-layer MoS2 and crystalline AlN. The results of all calculations are compared to existing experimental data. In general, the TBC of such 2D interfaces is low, below ~20 MWm -2 K -1 , due to the weak van der Waals (vdW) coupling and mismatch of phonon density of states (PDOS) between materials. However, the TBC increases with vdW coupling strength, with temperature, and with the number of MoS2 layers (which introduce additional phonon modes). These findings suggest that the TBC of 2D materials is tunable by modulating their interface interaction, the number of layers, and finding a PDOSmatched substrate, with important implications for future energy-efficient 2D electronics, photonics, and thermoelectrics.
Electrical and thermal properties of atomically thin two-dimensional (2D) materials are affected by their environment, e.g. through remote phonon scattering or dielectric screening. However, while it is known that mobility and thermal conductivity (TC) of graphene are reduced on a substrate, these effects are much less explored in 2D semiconductors such as MoS 2. Here, we use molecular dynamics to understand TC changes in monolayer (1L) and bilayer (2L) MoS 2 by comparing suspended, supported, and encased structures. The TC of monolayer MoS 2 is reduced from ~117 Wm-1 K-1 when suspended, to ~31 Wm-1 K-1 when supported by SiO 2 , at 300 K. Encasing 1L MoS 2 in SiO 2 further reduces its TC down to ~22 Wm-1 K-1. In contrast, the TC of 2L MoS 2 is not as drastically reduced, being >50% higher than 1L both when supported and encased. These effects are due to phonon scattering with remote vibrational modes of the substrate, which are partly screened in 2L MoS 2. We also examine the TC of 1L MoS 2 across a wide range of temperatures (300 to 700 K) and defect densities (up to 5×10 13 cm-2), finding that the substrate reduces the dependence of TC on these factors. Taken together, these are important findings for all applications which will use 2D semiconductors supported or encased by insulators, instead of freely suspended.
The design of applications, especially those based on heterogeneous integration, must rely on detailed knowledge of material properties, such as thermal conductivity (TC). To this end, multiple methods have been developed to study TC as a function of vibrational frequency. Here, we compare three spectral TC methods based on velocity decomposition in homogenous molecular dynamics simulations: Green-Kubo modal analysis (GKMA), the spectral heat current (SHC) method, and a method we propose called homogeneous nonequilibrium modal analysis (HNEMA). First, we derive a convenient per-atom virial expression for systems described by general many-body potentials, enabling compact representations of the heat current, each velocity decomposition method, and other related quantities. Next, we evaluate each method by calculating the spectral TC for carbon nanotubes, graphene, and silicon. We show that each method qualitatively agrees except at optical phonon frequencies, where a combination of mismatched eigenvectors and a large density of states produces artificial TC peaks for modal analysis (MA) methods. Our calculations also show that the HNEMA and SHC methods converge much faster than the GKMA method, with the SHC method being the most computationally efficient. Finally, we demonstrate that our MA implementation in the Graphics Processing Units Molecular Dynamics code on a single graphics processing unit is over 1000 times faster than the existing implementation in the Large-scale Atomic/Molecular Massively Parallel Simulator code on one central processing unit.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.