In-body sensors, also known as implantable sensors or bio-implants, are devices designed to be inserted into the human body to monitor various physiological parameters or deliver therapeutic treatments. These sensors can provide real-time data about the body’s internal conditions, allowing for continuous monitoring and analysis. They are typically used in medical applications and can be used for diagnostics, disease management, or research purposes. In-body sensors offer unique advantages in the context of chronic disease management, including:
– Spatial resolution: access to local environment, direct and sensitive measure of specific parameters
– Temporal resolution: sense onset of events and conditions, enabling timely interventions
– Closed-loop : adaptive, personalized, targeted and automated therapy
Our research focuses on the integration of CMOS, MEMS technologies to develop multi-analyte, multimodality and long-term injectable and insertable sensors. We take a holistic approach and consider the interaction of all the primary components of a wireless sensor system, including the biointerface, the readout electronics and the wireless units. We study novel ways to integrate and miniaturize these components with focus on power efficiency and cost-effective microfabrication processes.
Our main research areas include:
– Ultra-low-power (ULP) CMOS sensor readout circuits
– Wireless powering and energy harvesting
– MEMS/CMOS sensor and circuit integration
Ultra-Low-Power (ULP) Wireless Sensor Readouts
Wireless sensors are poised to play a pivotal role in future disease management. By developing mm-scale smart insertable/injectible medical sensors, we can enable minimally invasive remote monitoring of patient’s chronic conditions and, in turn, drive more efficient therapies. The need for sensor miniaturization however precludes the use of batteries, resorting to alternative forms of energy, either harvested from the local environment or external sources. In either case the power available at the sensor node is in the mW range and below. To retain useful sensor functionality, innovative solutions are needed to push the boundaries of power consumption and energy-efficiency of the sensor readout units. The WiS Lab addresses these challenges by developing sensor interfaces based on the integration of a number of ULP design strategies, including:
o Analog-to-time conversion (ATC) and time-based signal processing.
o Dynamic amplifiers and comparators
o Multi-sensor reconfigurable front-ends
o Pulse modulation (PM) wireless schemes for power efficient high-speed data transfer
Wireless Powering and Energy Harvesters
Inductive and ultrasonic wireless power transfer (WPT) are well establish methods to power battery-free miniaturized implants. WPT can be adopted to either recharge a small energy-storage element (supercapacitor) or continuously power the implant. Further device miniaturization results, however, in smaller energy storing capacity and lower power transfer efficiencies. In addition, WPT systems require a reliable external energy source, worn and frequently operated by the patient. Energy autonomy is therefore the ultimate ambition in miniaturized wireless sensors. The critical components of energy-autonomous sensors are the energy harvester and the power management unit. Inertial, electrostatic and thermal harvesters are the most commonly adopted solutions in integrated systems, thanks to their relatively high power output and compatibility with standard CMOS technology. To further increase the available power, we are currently investigating approaches to integrate hybrid MEMS-CMOS harvesting solutions with smart power managements circuits that can support reliable and continuous autonomous sensing. In addition, we are developing hybrid (multi-tier) powering methods for distributed sensor networks and to support sensor-to-sensor communication.
Integrated CMOS-MEMS Sensors
MEMS sensors benefit from high-performance, small size, reproducibility and low cost. Sensors that monolithically integrate MEMS structures and CMOS circuits offers unmatched performance in terms of SNR, scalability and functionality. MEMS can be integrated with CMOS circuits in mainly 3 ways: In MEMS-first processes, MEMS components are integrated in a silicon substrate which is then planarized to then support processing steps for CMOS electronics. MEMS-first approaches are often not compatible with foundries design rules thus limiting the acceptance by foundries of pre-processed wafers. In MEMS-last processes, MEMS components are integrated after full completion of the CMOS chip fabrication. In MEMS-CMOS processes, MEMS structures are developed from an existing CMOS metallization stack (back end of line – BEOL). By leveraging the electromechanical properties of CMOS layers, high-performing CMOS-MEMS devices can be developed, with the major benefit of ease of integration, CMOS compatibility and scalability. Currently we are exploring a number of avenues for the CMOS-MEMS monolithic integration of high-density multimodal sensor arrays and microfluidic structures. Combinations of BEOL-embedded and bulk-micromachined MEMS structures shows great promise in terms of electromechanical performance, process compatibility and potential cost.
Wireless Sensing Lab 