Definition of a system for harvesting in an ultra-scaled CMOS sensing device

The future improvements of medicine require personalized drug therapies, direct consequence of a precise health parameters analysis. In order to achieve a continuous real-time monitoring, the optimal solution is represented by implantable medical devices (IMD), typically composed by a sensor, a battery and a circuital control module. However, the presence of a battery reduces the autonomy of IMDs, requiring periodical maintenance, causing heat dissipation and increasing the total occupation area. Enzymatic biofuel cells (EBFCs) partially solve problems typical of a lithium-ion battery: an EBFC generates electrical power converting chemical energy obtained from the reaction between an enzyme present in the human blood and electrolytes, with a reduced generation of by-products. Reducing furtherly the total area allows the possibility to obtain a system usable in a body-dust context: a challenging microelectronic emerging diagnostic technology based on artificial cells able to cross the gut-wall barrier. Thus, it has been investigated the reliability of a device that exploits the presence of a EBFC to perform at the same time both harvesting and sensing, without the necessity of a dedicated sensor. The glucose has been chosen among different fuels accessible in the human body, considering the huge presence of previous cases of study. Once the glucose fills the space between electrodes, the cell starts to store energy in a capacitor, supplying power to the circuit; a comparator senses the voltage drop on the capacitor, allowing the finite state machine that operates on the output of comparator to alternate a charge phase with a discharge phase, during which the EBFC and the capacitor are disconnected, the available energy is used to transmit the information outside of the body by a Bluetooth low energy transmitter. The desired value of glucose concentration is extrapolated from the charge-discharge characteristic of the capacitor. As a first item of the analysis, various possibilities of glucose EBFCs have been described, in terms of materials, geometries and theoretical models necessary to determine the physical behaviour, pointing out the ideal configuration to obtain the maximum power in output maintaining the occupation area under control and identifying the physical ideal configuration of the sensor. An electrical circuit corresponding to the ideal EBFC cell has been defined, in order to perform SPICE simulations. Afterward, basing on output values from experiments previously described in literature and simulated in laboratory, an analog schematic of the circuitry has been designed, working with ideal parameter for initial analysis and taking into account both ultra-low power and area requirements. Finally, further improvements are suggested: to obtain a testable device, each block of the schematic has to be designed and simulated at a CMOS level, considering leakages, frequency variations with temperature, ease of fabrication of the sensor. The last point is particularly critical, considering the complexity of CMOS manufacturing at a nanoscale that might impose restrictions on the EBFC design and on the choice of components.