Two-Dimensional materials: from large-area growth to the performance in radio-frequency range

This thesis presents a collection of experimental results about the large-area growth of single-layer graphene and MoS2 films. In addition, we present the results of electrical characterisation of two-dimensional (2D) materials in direct current (DC) and radio-frequency (RF) range. The thesis is divided into four parts. In the first part we describe the recipe for the large-area growth of single-layer graphene on Cu substrate with the chemical vapor deposition (CVD) technique. In addition, we present two different methods for the transfer of the grown films onto arbitrary substrates. In order to investigate the intrinsic properties of the grown film we perform the electrical characterization in DC and RF range. The second part is devoted to the large-area growth of MoS2 single-layer films. We describe two different recipes for the growth: two-step thermolysis process of ammonium tetrathiomolybdate (NH4)2MoS4 in S atmosphere and CVD reaction ofMoO3 in S atmosphere. Additionally, we perform the electrical characterisation of the fabricated FETs based on grown MoS2 films. Apart from the grown materials, we also analysed the electrical properties of exfoliated MoS2. In the third part, we investigate the behaviour of 240 nm length top-gated MoS2 transistors operating in RF range. We fabricated MoS2 RF devices based on 1L-MoS2, 2L-MoS2, 3L-MoS2 and as well as multilayer MoS2 film with the thickness of » 5 nm. They exhibit current, power and voltage gain. The best DC and RF characteristics were obtained for 3L-MoS2 device showing a cutoff frequency fT = 6 GHz and a maximum oscillation frequency fmax = 8.2 GHz after the deembedding procedure. In order to boost the RF performance of MoS2 devices, firstly, we fabricated top-gated FETs with the gate length as short as 40 nm, secondly, we introduced the "edge-contacted" injection technique to our devices. Thus, we were able to obtain the record cutoff frequency for MoS2 devices of 6 GHz before deembedding and fT = 25 GHz after the deembedding procedure.

Single-Layer MoS2 Electronics

Simone Bertolazzi, Andras Kis, Dominik Sebastian Lembke

CONSPECTUS: Atomic crystals of two-dimensional materials consisting of single sheets extracted from layered materials are gaining increasing attention. The most well-known material from this group is graphene, a single layer of graphite that can be extracted from the bulk material or grown on a suitable substrate. Its discovery has given rise to intense research effort culminating in the 2010 Nobel Prize in physics awarded to Andre Geim and Konstantin Novoselov. Graphene however represents only the proverbial tip of the iceberg, and increasing attention of researchers is now turning towards the veritable zoo of so-called "other 2D materials". They have properties complementary to graphene, which in its pristine form lacks a bandgap: MoS2, for example, is a semiconductor, while NbSe2 is a superconductor. They could hold the key to important practical applications and new scientific discoveries in the two-dimensional limit. This family of materials has been studied since the 1960s, but most of the research focused on their tribological applications: MoS2 is best known today as a high-performance dry lubricant for ultrahigh-vacuum applications and in car engines. The realization that single layers of MoS2 and related materials could also be used in functional electronic devices where they could offer advantages compared with silicon or graphene created a renewed interest in these materials. MoS2 is currently gaining the most attention because the material is easily available in the form of a mineral, molybdenite, but other 2D transition metal dichalcogenide (TMD) semiconductors are expected to have qualitatively similar properties.In this Account, we describe recent progress in the area of single-layer MoS2-based devices for electronic circuits. We will start with MoS2 transistors, which showed for the first time that devices based on MoS2 and related TMDs could have electrical properties on the same level as other, more established semiconducting materials. This allowed rapid progress in this area and was followed by demonstrations of basic digital circuits and transistors operating in the technologically relevant gigahertz range of frequencies, showing that the mobility of MoS2 and TMD materials is sufficiently high to allow device operation at such high frequencies.Monolayer MoS2 and other TMDs are also direct band gap semiconductors making them interesting for realizing optoelectronic devices. These range from simple phototransistors showing high sensitivity and low noise, to light emitting diodes and solar cells. All the electronic and optoelectronic properties of MoS2 and TMDs are accompanied by interesting mechanical properties with monolayer MoS2 being as stiff as steel and 30x stronger. This makes it especially interesting in the context of flexible electronics where it could combine the high degree of mechanical flexibility commonly associated with organic semiconductors with high levels of electrical performance. All these results show that MoS2 and TMDs are promising materials for electronic and optoelectronic applications.

2015

Transmission Electron Microscopy and Nanoindentation Studies of Ultrathin Transition Metal Dichalcogenide Semiconductor Membranes

Jacopo Brivio

Until now, only two intrinsically two-dimensional materials have been isolated: graphene and boron nitride (BN). As a consequence, the knowledge of two-dimensional systems is still limited and other classes of two-dimensional materials are needed to enrich our understanding of these systems. In this thesis, we studied the structural and mechanical properties of two-dimensional transition metal dichalcogenide (TMD) crystals. Apart from fundamental interest, transition metal dichalcogenides have great appeal due to their possible applications in electronic. Here, the focus is on MoS2, while the main employed techniques are transmission electron microscopy and nano-indentation. The thesis is divided into three parts. The first part is devoted to summarize the general properties of transition metal dichalcogenide materials, to illustrate the instrumentation, and to describe the deposition techniques used in sample preparation. In particular, the transfer technique used to transfer selected flakes among different substrates is described. In the second part, we investigate single-layer MoS2 flakes obtained by mechanical exfoliation. We describe two methods to distinguish single-layer MoS2 flakes from thicker ones by electron diffraction. Ripples are observed over the flake surface. Thanks to diffraction pattern interpretation and high resolution transmission electron microscopy, their amplitude is estimated to be 0.6 - 1 nm in height. Nano-indentation is used to determine the Young’s modulus and the breaking strength of single-layer MoS2 membranes. The Young’s modulus is EY =270±100GPa while the breaking strength is σmax =23±5 GPa. In order to fully take advantage of the promising electronic properties of transition metal dichalcogenide, it is necessary to grow them over large scale. In this way, the pla- nar technology developed for silicon can be implemented and a large number of devices with uniform electrical properties can be realized. When comparing different growths, methods to quantify the quality of the films are mandatory. In this context, we show that transmission electron microscopy is a valuable method to characterize large area MoS2 and MoSe2 films. We are able to estimate the grain size and to have a direct view of the grain boundaries. In fact, grain size and boundary quality constitute the important parameters to qualify two-dimensional poly-crystalline films. Finally, we describe a novel in-situ experiment which allows us to simultaneously perform transmission electron microscopy and electrical characterization on a device based on a suspended MoS2 flakes. The effect of the current flow on the MoS2 flake are considered. These two topic constitute the last part of the thesis.

EPFL2013

Graphene Field Effect Devices and Circuits

Clemens Benedict Nyffeler

Electronics today permeate our life and existence. It has become nearly impossible to evade any dependence on electronic devices that surround us every day - computers, phones, televisions. But also other objects become increasingly "smart" - watches, cars, coffee machines, even entire buildings. Without necessarily being aware of it, every person in our modern world is the owner of billions, probably tens of billions of transistors. These elementary switches are the primitive units, of which all electronic devices are made, much like biological cells constitute our bodies. Decades of technological progress at an incredible pace have been fueled by the constant improvement of semiconductor device technology. The original combination of silicon, silicon dioxide and aluminum required to build a transistor, has been complemented by a myriad of other materials. One of the last remainders of the original technology, the silicon channel, is now about to be replaced as well. While a short term remedy for the next performance bottleneck might be found in III-V compounds a more compelling alternative could be found in 2D materials. Graphene was not only the first 2D material to be discovered and isolated in 2004, but also has the most extraordinary electric properties, owing to the high symmetry of its lattice. The content of this thesis presents a broad examination of the graphene field effect device reaching from the fabrication over electrical characterization to data analysis, device modeling and finally simulation of a small circuit. In this thesis, we present practical considerations regarding the experimental examination of graphene field-effect devices. A fabrication flow tailored for top-gated graphene devices was developed, taking into account the particular requirements and sensitivities of the material. We also describe a set of versatile software tools that were developed for the design of devices, chips and wafers, their automated electric characterization and finally for browsing and visualizing the measurement results. The data analysis was performed with a very effective conductance-based model, which is based on semi-empirical models commonly used to describe graphene devices. We provide an overview of these models, the phenomena which they take into account and the steps that can be taken to improve their accuracy to obtain the model we finally utilized. A environment was created to use our model in a \textsc{Spice}-like circuit simulator in order to study possible topologies in which graphene devices could constitute an elementary circuit block. We focus our study on devices that operate as a differential pair, resembling the configuration that also enables very high-speed source-coupled and emitter-coupled logic circuits based on standard silicon transistors. Using analytical calculations we determine tuning parameters and their optimum values to maximize the transfer characteristics of our graphene-based differential circuit block. In order to achieve more accurate simulation and as a means to verify the empirical model, we worked on a more rigorous approach. Based on first principles, we construct a model building on the specific carrier statistics in graphene. These deviate from the usual Boltzmann statistics and lead to a an equation describing the device's charge-voltage relation, which is transcendental and cannot ordinarily be solved. By using asymptotic approximations, we obtain closed-form expressions for the device current as a function of bias conditions. Unlike many other models, we can discriminate both between electron and hole currents as well as between drift and diffusion currents, making the model well-suited for implementation as a compact model.

EPFL2016