Energy-efficient and flexible electronics for military applications

To ensure the success of future military endeavors, soldiers will need to be equipped with new technologies that enable enhanced situational awareness, command and control, communication, and sensing. Furthermore, these advanced capabilities will require increased levels of data processing, computing, and networking. Technologies that are lightweight, flexible, and energy-efficient will improve maneuverability by reducing the weight and volume carried by individual soldiers. The logistics burden on these technologies can be reduced by addressing the needs for increased power supplies, energy harvesting, renewable sources, and improved energy storage densities. The design and incorporation of novel energy-efficient materials, combined with conventional electronics, can be used to form heterogeneous low-power devices to reduce power demands.

For about 50 years, the scaling of silicon for digital electronics has followed Moore's law, and the energy efficiency of individual transistors (defined by the power delay product) has seen 30% compounded annual growth.¹ It has been possible for analog devices to piggyback on these scaling trends, which has resulted in the integration of increasingly higher-frequency silicon devices and circuits. Gallium arsenide or gallium nitride (GaN) semiconductors can be used in place of silicon for even higher frequency or power needs. Heterogeneous computing processes are being used to integrate semiconductors (e.g., GaN) onto silicon for high-power and high-frequency applications.² Fundamental physical limitations to the scaling of silicon and related materials, however, are predicted. This will lead to a plateau in the trend of improving energy efficiency.

We have investigated the use of 2D materials as an alternative way to improve the energy efficiency of silicon devices. By exploiting the novel physics of these materials, energy-efficient electronic devices and sensors can be enabled and then integrated with conventional semiconductors or flexible substrates. 2D materials can be isolated in a few, or single, atomic layers. The electronic band structure of 2D materials can also be altered significantly, depending on the layer count, the relative orientation between layers, the electric field, and strain. By layering different 2D materials on top of one another, we can produce so-called van der Waals (vdW) heterostructures, which have novel electrical, optical, and thermal properties. The band structure tunability of 2D materials and the multitude of possible vdW heterostructures provide new variables that can be manipulated to produce novel energy-efficient devices (such as vertical tunneling or spin-based transistors). Our aim has been to produce a switch that can break the energy-efficiency limit of a silicon transistor. As such, we have theoretically and experimentally explored the electrical and thermal properties of single and multilayer molybdenum disulfide (MoS₂), i.e., a 2D material.⁶

Strain is often used to engineer the band structure (i.e., band gaps and effective masses) of silicon technologies. This approach can increase the electron mobility and performance of transistor devices, which leads to improvements in energy efficiency. We have used density functional theory to predict strain-induced variations in the electronic properties of MoS₂ bilayer sheets.³ We found that the band gap of the material will shrink significantly when it is put under biaxial in-plane strain. Compressive strain, however, will have less of an effect (see Figure 1). Figure 1 also shows that biaxial strain—more than uniaxial strain—decreases the conduction band effective mass. We can use these results to devise new methods for increasing electron mobility and device performance. Since flexible devices will operate under various types and magnitudes of strain, it will also be of vital importance to understand how varying strain levels can affect the operation of devices. In another set of calculations, we have used first-principle molecular dynamics simulations to study the temperature-dependent phonon shifts in monolayer MoS₂ (see Figure 2).⁴ We were able to reproduce qualitatively the temperature-dependent phonon shifts that we observed with Raman spectroscopy measurements.

Figure 1. Variations in band gap energies with (a) biaxial strain and (b) uniaxial strain.³∊: Strain. Γ, K, I, Σ_min, and A are special high-symmetry points in the ‘Brillouin zone.’

Figure 2. Calculated phonon density states (DOS) of monolayer MoS₂at temperatures of 50 and 300K.⁴ E¹_2g: In-plane mode. A_1g: Out-of-plane mode. Δω: Frequency shift.

For our experimental work, we have focused on growing large-area MoS₂ sheets through the process of chemical vapor deposition (CVD). We have also fabricated field-effect transistors in monolayer MoS₂ crystals (see Figure 3) to measure their electrical properties. We have measured the current-voltage characteristics of these crystals at various temperatures and have correlated the electrical data from a single crystal with a high-resolution Raman map of its surface. In addition, we have used Raman and photoluminescence mapping to provide further insights into the material properties of MoS₂.⁵

Figure 3. Maps of a typical molybdenum disulfide (MoS₂) crystal made using (a) atomic force microscopy, (b) high-resolution Raman spectroscopy (out-of-plane mode intensity), and (c) backscatter field emission scanning electron microscopy.⁵These images show variation in the quality of the film from the center to the edge of the crystal, as well as contamination between the substrate and the film. These imperfections act as scattering centers that decrease electron mobility.

The combination of our theoretical and experimental investigations has improved our understanding of temperature-dependent phonon shifts in monolayer MoS₂. We have also gained new information concerning the temperature dependence of electron mobility. Imperfections in CVD-grown materials (as illustrated in Figure 3) act as scattering centers that decrease electron mobility. The theoretical maximum level for electron mobility is thus difficult to achieve in real-world scenarios, and it is a challenge to increase the performance of 2D materials.

We have investigated the properties of novel 2D materials, with the eventual aim of breaking the energy-efficiency limit of silicon microelectronics. Our theoretical calculations show that the electronic properties of these materials can be engineered using strain. With our fabricated devices, however, we have only been able to achieve performances that are significantly below the theoretical limits. We are now planning to study vertical transport within vdW heterostructures (consisting of multiple 2D materials in contact with 3D materials) to overcome some of the performance limitations associated with in-plane transport.