02 / Research topics
Potential PhD Research Topics in Microelectronics
Explore suggested topics and the methods, tools and expertise that could support each investigation.
1
Thermally aware chiplet architectures for scalable heterogeneous artificial intelligence processors
As monolithic scaling becomes expensive, chiplets offer modular integration but create thermal hotspots, communication bottlenecks, and packaging challenges. The gap is the lack of jointly optimized thermal, architectural, and interconnect models. This research would develop placement algorithms, adaptive chiplet networks, and cooling-aware runtime policies, validated through simulation, prototypes, and benchmark workloads. Novelty lies in co-optimizing physical temperature constraints with AI performance and reliability rather than treating thermal management as a post-design task.
Applications and use cases
2
Monolithic three-dimensional integrated circuits using low-temperature semiconductor processing
Despite the density advantages of monolithic 3D integration, conventional fabrication temperatures damage lower-tier devices and limit material choices. A major gap remains in reliable low-temperature processing with acceptable mobility, variability, and yield. The proposed work would investigate compatible transistors, inter-tier vias, process sequencing, and compact models below thermal budgets suitable for sequential fabrication. Its contribution would be a validated process-design methodology linking materials, device characteristics, circuits, and system-level benefits.
Applications and use cases
3
Energy-efficient in-memory computing arrays for reliable edge artificial intelligence
Emerging in-memory computing reduces data movement, yet analog nonidealities, endurance limits, noise, and device variability prevent dependable deployment. The unresolved challenge is achieving accuracy and energy efficiency simultaneously under realistic manufacturing conditions. This project would design memory-cell arrays, calibration schemes, error-aware algorithms, and peripheral circuits, then evaluate them using measured device statistics. Novelty would come from cross-layer co-design that treats imperfections as design parameters instead of assuming ideal memory behavior.
Applications and use cases
4
Neuromorphic spiking processors with adaptive memristive synapses for autonomous sensing
Future autonomous sensors require ultra-low-power computation, but conventional processors waste energy moving data between memory and logic. Memristive neuromorphic systems are promising, although device drift, stochastic switching, limited endurance, and training instability remain poorly controlled. This research would create adaptive synapse circuits, online learning methods, and hardware-aware spiking architectures, with experiments on event-based sensing tasks. The novelty is an integrated approach connecting device physics, learning dynamics, and application-level robustness under changing environments.
Applications and use cases
5
Silicon photonic interconnects for energy-efficient high-bandwidth computing systems
Future processors face electrical interconnect bandwidth and energy barriers as data movement dominates system power. Silicon photonics can provide high bandwidth, but laser integration, thermal sensitivity, packaging loss, and control overhead remain significant gaps. This project would develop co-designed photonic links, compact electro-optic drivers, calibration techniques, and packaging strategies. Its novelty would be system-level optimization of optical devices, circuits, and workload communication patterns, supported by fabricated or experimentally characterized link components.
Applications and use cases
6
Radiation-tolerant microelectronics for dependable space and nuclear instrumentation
Meanwhile, space and nuclear systems require electronics that survive ionizing radiation, displacement damage, and single-event effects, while commercial technologies continue scaling rapidly. Existing approaches often trade substantial area, speed, or energy for reliability, leaving room for efficient adaptive protection. This research would combine radiation-aware device structures, redundant circuits, error monitoring, and predictive lifetime models. The contribution would be experimentally validated design rules that connect radiation exposure, degradation mechanisms, circuit behavior, and mission-level reliability.
Applications and use cases
7
Hardware security primitives based on physically unclonable semiconductor device variability
Secure, low-cost device identity and cryptographic key generation remain difficult in constrained connected electronics. Although physically unclonable functions exploit manufacturing variation, many designs suffer from environmental instability, modeling attacks, aging, and unreliable enrollment. This project would engineer new device-level primitives, develop lightweight error correction, and evaluate resistance against machine-learning and side-channel attacks. Novelty would arise from jointly optimizing entropy, reliability, attack resilience, and manufacturability across realistic process corners.
Applications and use cases
8
Flexible and stretchable microelectronics for continuous biomedical and human-machine interfaces
Sustainable, comfortable biomedical electronics must conform to moving tissue while maintaining electrical performance, biocompatibility, and long-term reliability. Current systems often rely on rigid islands, wired instrumentation, or fragile conductive traces, leaving integration and lifecycle gaps. This research would develop stretchable conductors, flexible sensors, low-power readout circuits, and encapsulation methods, followed by mechanical, electrical, and biological testing. The novelty would be a fully integrated platform that preserves signal quality during repeated deformation and realistic physiological exposure.
Applications and use cases
9
Carbon-aware power management circuits for ultra-low-power edge microcontrollers
Sustainable, always-on edge devices need lower energy consumption across sensing, computation, communication, and standby operation. Conventional power management rarely accounts for workload intermittency, battery aging, embodied carbon, or renewable energy availability. This research would develop adaptive converters, energy-harvesting interfaces, workload predictors, and carbon-aware control policies. Its novelty would be evaluating microelectronic designs using both operational energy and lifecycle metrics, producing practical circuit architectures that maximize useful computation per unit of environmental impact.
Applications and use cases
10
Cryogenic CMOS control electronics for scalable quantum computing systems
Translating quantum processors beyond laboratory prototypes requires control electronics close to cryogenic qubits, reducing wiring, latency, and thermal load. Current cryogenic CMOS solutions face limited power budgets, device-model uncertainty, noise coupling, and insufficient integration density. This research would design low-temperature amplifiers, multiplexers, digital controllers, and thermal-aware architectures, validated across temperature and frequency ranges. The novelty would be a co-designed cryo-electronic stack that links transistor behavior, control fidelity, cooling constraints, and scalable quantum-system requirements.
Applications and use cases