In recent years, there has been tremendous progress in large-scale

mechanically flexible electronics, where electrical components are

fabricated on non-crystalline substrates such as plastics and glass. These devices are currently serving as the basis for various applications such as flat-panel displays, smart cards, and wearable electronics.

In this thesis, a promising approach using chemically synthesized

nanomaterials is explored to overcome various obstacles current technology faces in this field. Here, we use chemically synthesized semiconducting nanowires (NWs) including group IV (Si, Ge), III-V (InAs) and II-IV (CdS, CdSe) NWs, and semiconductor-enriched SWNTs (99 % purity), and developed reliable, controllable, and more importantly uniform assembly methods on 4-inch wafer-scale flexible substrates in the form of either parallel NW arrays or SWNT random networks, which act as the active components in thin film transistors (TFTs).

Thusly obtained TFTs composed of nanomaterials show respectable

electrical and optical properties such as 1) cut-off frequency, ft ~ 1 GHz and maximum frequency of oscillation, fmax ~ 1.8 GHz from InAs parallel NW array TFTs with channel length of ~ 1.5 μm, 2) photodetectors covering visible wavelengths (500-700 nm) using compositionally graded CdSxSe1-x (0 < x < 1) parallel NW arrays, and 3) carrier mobility of ~ 20 cm2/Vs, which is an order of magnitude larger than conventional TFT materials such as a-Si and organic semiconductors, without sacrificing current on/off ratio (Ion/Ioff ~ 104) from SWNT network TFTs.

The capability to uniformly assemble nanomaterials over large-scale

flexible substrates enables us to use them for more sophisticated

applications. Artificial electronic skin (e-skin) is demonstrated by laminating pressure sensitive rubber on top of nanomaterial-based active matrix backplanes. Furthermore, an x-ray imaging device is also achieved by combining organic photodiodes with this backplane technology.

Ever since the invention of the transistor, aggressive channel length scaling has been pursued to achieve higher performance and greater packing density. In order to preserve gate control at short channel lengths, the transistor channel has evolved from bulk to low dimensional substrates, such as 2D thin films and 1D nanowires. For scaling to continue, it is vital that we understand the processing and physics of low dimensional materials.

Chapter 2 focuses on quasi-2D ultrathin body InAsSb-on-insulator n-FETs. III-V materials offer high mobilities for excellent on-state currents, and by using a thin film platform we could potentially obtain good off-state characteristics. Previously we have demonstrated high performance InAs-on-insulator n-FETs. In this study we implement InAsSb transistors on SiO₂ and achieve a ~2x enhancement in effective mobility over analogous InAs devices. Top-gated devices are demonstrated with an I_ON/I_OFF of 10²-10³ and an intrinsic conductance of ~0.56 mS/μm.

1D InAs nanowire (NW) n-FETs are explored in chapter 3. In particular, the nanowire transistors are used to study ballistic transport, the theoretical current density upper limit. We experimentally observe ~ 60nm channel length devices reaching ~80% of the ballistic limit. Length dependent studies on the same nanowire are used to extract a mean free path of ~150nm for the 1st and 2nd electron subbands. We find the mean free path to be independent of temperature, suggesting that surface roughness scattering is the dominant scattering mechanism.

Chapter 4 explores 2D transition metal dichalcogenide (TMDC) thin films. TMDC thin films offer the physical limit of scaling, and ohmic contacts to its conduction and valence bands are required for it to realize low power complementary logic. Previous studies show that elemental metal energy levels are pinned near the conduction band of TMDCs and do not offer effective hole injection. To address this we explore a high work function transition metal oxide, substoichiometric molybdenum trioxide (MoO_x, x<3), as a hole injection layer to MoS₂ and WSe₂. MoS₂ diodes and p-FETs are demonstrated with MoO_x contacts. WSe₂ p-FETs with MoO_x contacts show a ~10x on-current improvement over devices with Pd contacts.

In chapter 5 we present heterojunction diodes formed by thin films of InAs and WSe₂. In traditional epitaxial heterojunctions, the number of possible material combinations are limited by lattice constraints. In this study we overcome this restraint by transferring one layer upon another to form a heterojunction. Specifically, InAs/WSe₂ heterojunction diodes are fabricated and measured. A forward/reverse current ratio >10⁶, reverse bias current <10^-12μm², and ideality factor of 1.1 are observed.

Since the inception of solid state semiconductors and device fabrication techniques, continuous scaling has been implemented as a key driver behind realizing faster electronics while optimizing for power consumption, improving the field in an exponential fashion (i.e. Moore’s Law) and facilitating modern technological advances that otherwise would have been impossible. In recent years, transistor gate length has entered into the single nanometer regime, encountering significant engineering and cost challenges. While efforts at mitigating these challenges have extended the lifetime of silicon-based semiconductors (all-around gate FETs for example), a more fundamental overhaul of the transistor is needed for long term progress if Moore’s law is to be upheld in terms of power reduction and performance improvement.

2D materials serve as an ideal candidate for addressing scaling issues on various fronts. Possessing atomic scale smoothness, varying band alignment, and desirable bandgap in the single layer limit, 2D transition metal dichalcogenides (TMD) material can serve as the active channel layer for transistors geared towards various different applications. Given their atomic smoothness and interlayer van der Waal interactions, 2D TMDs provides the intrinsic scaling advantage in the vertical axis while the ideal uniformity allows for predictability of carrier behavior across lateral areas. However, realistic integration of 2D TMDs into devices have been far from ideal, and the existence of both surface and edge defects on the system becomes the current bottleneck prohibiting any realistic integration of 2D TMDs into modern devices. In this report, we examine both surface and edge defects of 2D TMDs, their effects on carrier movements and recombination, establish an analytical model of defect analysis, and introducing a new approach to TMD patterning considering effects on the resultant edges.

Specifically, the surface defects on tungsten diselenide (WSe2) will be discussed in relation to traditional field effect transistor applications, and how we can take advantage of such surface defects and engineer them into sites of p-type doping via covalent functionalization. In this way, we demonstrate a 5 order of reduction in the contact resistance tunable degenerate doping.

Additionally, edge defects of tungsten disulfide (WS2) are characterized by ways of experimentally measuring generated carrier recombination. A universal metric applicable to all 2D semiconductor is introduced here towards describing the edge defect quality called Edge Recombination Velocity (ERV). A qualitative discussion of the edge defect will also be presented here with respect to edge etching methods, differences in the 2D material chosen, and intrinsic edge orientation.

Finally, we demonstrate scanning probe lithography (SPL) as a reliable top down method towards nanoscale patterning of 2D materials, and expand the ERV characterization platform to MoS2, MoSe2, WS2, and WSe2. Additionally, we demonstrate that through SPL, different materials possess drastically different ERVs, highlighting the lessened impact of the process induced edge defects, establishing a platform for material-based edge passivation experiments.

Integrated circuits consists of building blocks called transistors. A transistor is a switch that enables logic operations to perform computing. Since the invention of the first integrated circuit, transistors have been scaled down in their dimensions to increase the density of transistors per unit area to enable more functionality. Transistor scaling is continued by introducing novel device structures and materials at each technology node. Due to the challenges such as short channel effects and the power consumption issues, novel materials are investigated as a candidate for next generation transistors. In this thesis, 2-dimensinal layered semiconductors, namely transition metal dichalcogenides (TMDCs) are studied to assess their electronic material properties as a candidate channel material for next generation electronic devices.

Chapter one, introduces the challenges in the state of the art MOSFET devices. Then the motivation for the use of TMDCs in MOSFETs is explained. In chapter two, doping of the TMDCs is studied to be able to probe the intrinsic electronic properties of the devices fabricated using these materials. Contact resistance can be decreased by doping and TMDC MOSFETs with near-ideal performance metrics are demonstrated. In chapter three the CMOS integration of the devices using TMDCs are examined. Logic operations are conducted by fabricating WSe2 n-FETs and p-FETs on the same flake. Then vertical 3-dimensional integration of n-FETs and p-FETs are demonstrated using the same gate. These transistors are connected as a CMOS inverter and logic operations are performed. Chapter four presents the band structure engineering study using TMDCs. Mono-multilayer MoS2 junctions are found to have a type-I heterojunction. Optoelectronic properties of this junction are investigated and the junction is shown to have a photoresponse that dominates the photoresponse coming from the contacts. In chapter five, the tunneling devices using TMDCs are studied. Dual-gated heterojunctions and homojunctions are studied. Gate-tunable diode characteristics of these structures are shown. Then p-TFETs are examined using black phosphorus homojunction devices.

Since the beginnings of the semiconductor revolution, device engineering and material development have been deeply interconnected. Innovations in one invariably spurred development in the other. One of the most important current challenges are developing techniques which enable deposition of materials on arbitrary substrates and the design constraints of devices fabricated via these techniques. This thesis focuses on furthering the materials development-device design cycle for three techniques: (i) epitaxial lift-off, (ii) nanowire growth via the vapor-liquid-solid growth mode, and (iii) the thin-film vapor-liquid-solid growth technique. The physics of devices made via (i) and (ii) are discussed in chapters two and three, the material science of (iii) is discussed in chapter four, and a method to engineer the physical and electronic properties of individual nanowires made via (ii) is discussed in chapter

Chapter two deals with the device physics and performance of ultra-thin compound semiconductor on insulator field effect transistors. Due to its excellent electron transport properties, InAs is used as the material of choice for the transistors, termed XOI FETs. These devices are fabricated utilizing an epitaxial layer transfer technique, enabling highly lattice mismatched single-crystalline layers of InAs to be deposited on Si wafers. Allowing devices which combine the excellent transport properties of compound-semiconductors and the established processing infrastructure for silicon. Chapter three discusses the design constraints and guidelines for nanopillar photovoltaics, specifically those fabricated utilizing the CdS/CdTe material system. Critically, the materials parameters that are favorable to non-planar cells are discussed, and the performance expectations for CdS/CdTe nanopillar photovoltaics are discussed.

Chapter four focuses on a growth technique developed to enable growth of ultra-large grain III-V semiconductors on non-epitaxial substrates, termed thin-film vapor-liquid-solid growth. This growth technique takes advantage of the vapor liquid solid growth mode to enable films with thickness down to 500 nm and grain sizes greater than 50 m. This represents a significant advance over all conventional growth techniques, which produce films with grain sizes 10-100x smaller. Furthermore, the optoelectronic quality of these materials is within a few percent of single crystal, as determined by quantitative luminescence efficiency measurements. Finally, Chapter five discusses the use of nanowires as physical templates, allowing engineering of Van der Waals forces to fabricate self-selective electrical connectors. By utilizing nanowire forests as starting templates, successive layers of material are deposited, each playing a specific role in the final structure, enabling composite material properties otherwise not possible.

Unusual device architectures may open up unexpected opportunities for optoelectronic devices. For example, the range of luminescent materials that can be easily integrated in light-emitting devices is often limited by material processing and band alignment issues. This hinders the development of electroluminescent devices at extreme wavelengths and the use of electroluminescence spectroscopy as an analytical technique. In this thesis, I explore an unconventional scheme of electrically-driven light emission based on pulsed metal-oxide-semiconductor capacitor devices. This concept can be used to produce electroluminescence from a wide range of materials including bulk semiconductors, two-dimensional semiconductors, colloidal quantum dots, organic molecules, and more. The performance of this light-emitting device is studied through simulation and experiment, and practical characterization is undertaken with different alternating-current driving schemes. By scaling the device, low voltage operation of these devices is further demonstrated. The device not only achieves emission spanning the near-ultraviolet to near-infrared energy spectrum, but also offers a straightforward avenue towards achieving electroluminescence from materials with different physical compositions and length scales spanning single particles to thick films. As a result of the scalable fabrication approach, miniaturized arrays of many different color light-emitting devices can be easily created on a monolithic substrate. The potential of this device platform for spectroscopic applications is illustrated by spectral imaging demonstrations which combine compressive sensing-based algorithms with highly multicolored and multiplexed light-emitting arrays.

Smart materials initiate drastic physical property changes in response to external stimuli such as chemical environment, light illumination, temperature change, electrochemical reaction and electrostatic field etc. The physical property changes normally include shapes, colors, and structural properties. Moreover, the responsive properties can be programmed in a controllable manner by controlling the external stimuli which are useful for various applications such as smart textiles and apparel, chemical sensors, and intelligent medical instruments. Smart materials can be categorized by the external stimuli they respond to such as electro active polymers, electrostrictive polymers, light responsive materials, shape memory alloys, and thermal responsive materials etc.. The most commonly seen smart materials are based on polymers and in a few cases they can be inorganic materials. Therefore, it can be well expected that the responsive mechanism are very diversified. In response to the external stimuli, the physical property change can be induced by a phase transformation, mass transport, microstructure change or as simple as physical stretch.

In particular, smart materials that can initiate large strain change or large deformation have very useful applications in artificial muscle, medical surgery, smart curtains and camouflage etc.. While the applications are diversified, there are no single criteria to describe the figure of merit. In general, to demonstrate useful applications, it requires smart materials to have fast responsive, large force, large strain and easy operation. My research centers on the development of smart materials that have fast response, large force, and a simple actuation mechanism. The general strategy for my research is to utilize the superior mechanical, thermal and optical properties of carbon nanotubes to make nanocomposites.

In the first part of my work, a simple approach is described to fabricate reversible, thermally- and optically responsive actuators utilizing composites of poly(N-isopropylacrylamide) (pNIPAM) loaded with single-walled carbon nanotubes. With nanotube loading at concentrations of 0.75 mg/mL, I demonstrated up to 5 times enhancement to the thermal response time of the nanotube-pNIPAM hydrogel actuators caused by the enhanced mass transport of water molecules. Additionally, I demonstrated the ability to obtain ultrafast near-infrared optical response in nanotube-pNIPAM hydrogels under laser excitation enabled by the strong absorption properties of nanotubes. The work opens the framework to design complex and programmable self-folding materials, such as cubes and flowers, with advanced built-in features, including tunable response time as determined by the nanotube loading.

In the second part of my work, I demonstrate a very easy and low-cost fabrication route for making single-walled carbon nanotube/polycarbonate based bilayer thin film materials that can be operated solely by solar light. The as-made nanocomposite materials possess superb light responsive properties such as fast response (less than 1 second), extreme solar light sensitivity (as low as 10 mW/cm2), very large deformation (90 degrees), and reversible actuation. The underlying mechanism is found to be the thermal expansion mismatch between the carbon nanotubes and polycarbonate, and the fast actuation response and large deflection can be attributed to the excellent light absorbing property, high thermal conductivity, very mechanically coherent interface, large Young's modulus, and small thermal expansion coefficient of the carbon nanotubes. I have also shown that by utilizing nanotubes with different chiralities, chromatic actuators that are selected responsive to specific wavelength band can be achieved. As a benchmark for fast and reversible light responsive material, a fast moving light-driven motor has been demonstrated. The easy-to-make and low-cost materials open the ways for many advanced photomechanical as well as biological applications.

In the third part of my work, theoretical and experimental results are presented for low voltage and large deformation electrostatic nanowire actuators. The design utilizes nanowire forest to enhance the electrostatic field. The goal of this work is to address the high operation voltage challenges of conventional dielectric elastomer actuators. Although there are still experimental challenges, my work sheds light on the validity of the design concept. Further progress can leads to huge breakthroughs in artificial muscles and microrobots.

With the innovation of technologies that will impact our daily life such as the internet of things (IoT), wearable health care systems, digitally extended reality, and autonomous vehicles, the development of next-generation optoelectronic devices that possess advanced functionality and high performance has emerged as a substantial component for the evolution of future technology. Two-dimensional (2D) semiconductors have been extensively studied in the exploration of new phenomena and properties that are not seen in conventional bulk semiconductors, especially for optoelectronic applications such as light-emitting diodes (LEDs) and photodetectors. While the majority of existing technology relies on bulk crystalline semiconductors, unprecedented possibilities can be opened up for developing next-generation optoelectronics owing to the unique properties and the tuning capabilities that 2D semiconductors offer.Chapter 2 presents the material processing approaches that are applied in atomically-thin semiconductor monolayers to enhance their luminescence efficiency. A combination of optimized doping and growth conditions enables extraordinarily bright luminescence in 2D transition-metal dichalcogenide (TMDC) monolayers with large scale and high stability. Chapter 3 describes the first critical step towards highly efficient LEDs at all brightness using monolayer semiconductors. The fundamental limitations of degraded luminescence efficiency in 2D semiconductors at high photocarrier densities have been identified and overcome by applying small mechanical strain to the materials, leading to the complete suppression of nonradiatvie recombination at all exciton densities. Chapter 4 demonstrates the room-temperature infrared optoelectronic devices whose operating wavelengths can be widely modulated by utilizing strain-tunable bandgap in black phosphorus. Mid-wavelength infrared (MWIR) LEDs with large spectrum tunability are developed and applied in non-dispersive infrared gas sensing technology. Highly-responsive photodetectors exhibit their detectivity exceeding those of state-of-the-art photodetectors with their detection wavelengths spanning from short-wavelength infrared (SWIR) to MWIR range.

A fundamental figure of merit for high performance optoelectronic devices is their photoluminescence (PL) quantum yield (QY). This value, which accounts for the fraction of absorbed photons are re-emitted by the material governs the ultimate performance of optoelectronic devices such as light emitting diodes, photodetectors/photovoltaics, and solar cells. Two-dimensional (2D) semiconductors have emerged as a promising material system for high-performance optoelectronic and electronic applications. This class of materials possess out-of-plane van der Waals bonding, and as a result has naturally terminated surfaces, even when the material thickness is scaled to the monolayer limit. As a result, surface recombination, a dominant non-radiate pathway in most materials, can be mediated due to the naturally terminated surfaces of van der Waals materials. This offers a potentially pathway to enable near-unity PL QY at the monolayer limit.

Despite this, the room-temperature PL QY for most monolayer 2D semiconductors is extremely poor. The prototypical 2D material, MoS2 has a maximum QY of 0.6% which indicates a considerable defect density. In the first section of this thesis we will address approaches of obtaining 2D materials with high PL QY, primarily focusing on an air-stable solution-based chemical treatment by an organic superacid which uniformly enhances the photoluminescence and minority carrier lifetime of MoS2 monolayers by over two orders of magnitude. The treatment defect-mediated non-radiative recombination pathways, thus resulting in a final QY of over 95% with a longest observed lifetime of 10.8±0.6 nanoseconds. This treatment methodology was also extended to other 2D semiconductors such as WS2 as well as scaled to large area films prepared by chemical vapor deposition.

Leveraging 2D materials with high brightness, the second section of this work will then discuss the development of light emitting devices. Firstly, we discuss the development of a transient electroluminescent device structure for use with monolayer 2D materials. One of the primary challenges for 2D materials scaled down to the monolayer limit has been the efficient simultaneous injection of electrons and holes. To overcome this crucial limitation, I will discuss the development of a device which uses a single metal-semiconductor contact and a gate. Utilizing an AC potential applied to the gate, this device achieves electron and hole injection via tunneling. This results in excess electrons and holes present during the AC transient, which subsequently recombine to emit light. This device is then combined with the passivation methodologies presented in the first section of this thesis. Next, I will demonstrate light emitting diodes based on black phosphorous (bP). Unlike the majority of 2D semiconductors which are direct gap only at the monolayer limit, bP maintains a direct gap over all thickness. Moreover, this bandgap is thickness tunable and can be varied from 1.8 eV for a monolayer down to 0.3 eV for thick films. This property is directly leveraged to fabricate devices with an emission wavelength which can tuned over range of 960 µm to 4.1 µm with external quantum efficiencies as high as 4.4% at room temperature. Notably, bP LEDs offer high performance for emission at infrared wavelengths where the quantum efficiency traditional III-V and II-VI LEDs are limited by auger recombination. The final section of this work I will discuss the development of infrared detectors based on 2D semiconductors.

Gas detection has wide-ranging applications, from industrial process control and greenhouse gas monitoring to disease diagnosis and pollution tracking, which are vital to support out civil life in many respects. To meet the needs of these applications, gas sensor technologies should enable scalable production and high performance. The current dominant technology is the micro-hotplate-based resistive ceramic sensor which suffers from high power consumption and poor selectivity. One of promising sensor technology, the chemical-sensitive field-effect transistors (CS-FET) have the advantage of miniature size, low-power consumption, and scalability. In order to improve the applicability of CS-FET sensors for real-world scenario, it is imperative to solve the problem of its response to ambient humidity. This can be achieved using integrated microheaters to locally heat the sensor, and evaporate the adsorbed water molecules.Another challenge for CS-FET devices is they cannot be made sensitive to gases such as methane, CO2, etc. at room temperature. Optical gas sensors are useful for these applications because their sensing mode does not depend on the interaction of the gas with the active material. However, optical sensors need high-performance infrared light emitters to detect gases at low concentrations. Black phosphorus (bP) has proven its ability to outperform the state-of-the-art compound semiconductors with its high photoluminescence quantum yield. Non-dispersive infrared (NDIR) gas sensors with LEDs based on bP are documented which can detect trace concentrations of greenhouse gases. However, optoelectronics based on bP have scalability issues due to its current deposition and processing methods. An ink-based approach to make uniform bP films by spin-coating has been demonstrated, which is then used to fabricate LEDs and photodetectors with high performance.