Search

The talk will describe the defect-to-device performance correlations for optoelectronic and power devices on hetero-epitaxially grown III-nitride epi-layers on silicon and sapphire. While UV detectors and power electronic devices might seem far apart, we see similarities between the defect origins of dark current in the former and leakage currents in the latter. Both have their origins in crystal growth. The first part of the work focuses on the correlation between performance parameters of vertical deep-ultraviolet photodetectors on c-plane sapphire and the density of extended defects vis-a-vis screw and edge dislocations. Through a careful optimization of nucleation density on the growth surface, state-of-the-art crystalline quality AlN epi-layers were grown. This led to the realization of record performance 289 nm p-i-n photodiodes with zero-bias quantum efficiency of 92 %, leakage current below 1 nA at 100 V and breakdown field in excess of 6 MV/cm. Performance of III-nitride power devices on silicon, on the other hand, was found to be more sensitive to structural defects such as surface pits as compared to dislocations. By utilizing a two-temperature growth technique which involved an AlN layer grown at high temperatures, a significant reduction in the pit density could be achieved. This resulted in three-orders of magnitude reduction in lateral as well as vertical leakage in AlGaN/GaN high electron mobility transistors (HEMTs) on silicon. To achieve a further reduction in the reverse leakage current and an enhancement in the device breakdown voltage, point defect control also becomes important. Carbon doping to reduce the background n-carrier concentration in the GaN epilayers is discussed as one of the possible approaches. Reduction of pit density and C-incorporation cumulatively led to the realization of 2 MV/cm vertical breakdown field in AlGaN/GaN HEMTs on silicon.

Use of low work function metal to lower the contact resistance on unpinned MIS contacts is presented. Through TCAD simulations and experimental data, it is established that conducting interfacial layers along with low work function metals like Ytterbium instead of Titanium for MIS contacts results in lower contact resistivity as well as an improvement in ION in ultra-scaled 14 nm Germanium-on-Insulator (GOI) n-channel FinFET devices. The benefit is larger for resistance-limited scaled FinFETs at the 10 and 7 nm nodes. Mixed-mode TCAD simulation of a self-loaded inverter circuit is used to demonstrate an improvement in overall energy-delay performance.

For safe and precise lunar landing it is required to guide the landing module to a pre-specified target location within specified accuracy limits by bringing down the landing velocity to near zero level. To achieve this, the thrust vector needs to be steered in an optimum way to minimize the fuel consumption. In this paper, a set of lunar centered 3 dimensional cartesian coordinates are used as state variables to represent the landing module translational dynamics and to ensure the soft landing at a specified target location. The precise landing trajectory design problem is formulated as an optimal control problem with constant thrust level and it is transformed into a two point boundary value problem(TPBVP) using the indirect approach based on Pontryagin's principle. The solution procedure of this TPBVP uses Differential Evolution, an evolutionary optimization technique. For real-time trajectory planning, few guidance laws are evaluated against the optimal control law obtained. The soft landing trajectory is propagated with control angles computed by the guidance algorithm to achieve the target conditions. The optimum fuel requirements with constant thrust to achieve different specified target conditions using the optimal control law are compared with fuel requirements obtained from simulated trajectory with guidance algorithm. Also the trajectory design is carried out with different constant thrust levels and the performance of the guidance laws is analyzed.

We demonstrate the benefit of using a low work-function metal (Yb) to lower the barrier height and hence, the contact resistivity (ρ c ) in unpinned metal-interfacial layer-semiconductor (MIS) contacts on n-type Ge and Si. Metal/doped TiO 2−x /n-Ge and n-Si MIS diodes fabricated with varying TiO 2−x thickness (1–5 nm) exhibit higher (upto 100X) current densities and lower contact barrier heights (upto 30%) for Yb versus Ti metallization. A combination of low work-function Yb, doped TiO 2−x with low conduction band offsets w.r.t Si and Ge, and high substrate doping( n+-Si:2×1020 cm−3 and n+-Ge:2.5×1019 cm−3) is shown to result in ultra-low ρ c values of 2.1×10−8 Ω.cm2 (Si) and 1.4×10−8 Ω.cm2 (Ge) for Yb/TiO 2−x contacts, significantly lower than Ti/TiO 2−x control devices, and the best reported contact resistivity on n-Ge till date.

Flush Air Data sensing System (FADS) forms a mission critical sub system for future reusable space transportation systems. FADS makes use of surface pressure measurements from the nose cap of the vehicle for deriving the air data parameters of the vehicle like angle of attack, angle of sideslip, Mach number, Altitude, etc. These parameters are used by the control and guidance system of the vehicle for reentry flight management as well as for subsequent mission phases upto landing. The altitude information is used in re-entry vehicles for gain scheduling, terminal area energy management and for homing the vehicle to the runway for landing. The FADS algorithm computes free stream static pressure from the surface pressure measurements made at the nose-cap. This paper addresses the computation of altitude from the free stream static pressure computed by FADS algorithm. Using the International Standard Atmosphere, hydrostatic equation, perfect gas law and temperature lapse rate equation, estimate of the vehicle altitude is obtained. The accuracy of the estimated altitude is of the order of 55 meter upto an altitude of 2 km. For higher altitudes of 5 km 10 km and 15 km, the accuracy of 235 meters, 565 meters and 695 meters is achieved respectively. Simulation results show that the proposed method is suitable for real time implementation.