Study of Silicon Nanowire Field Effect Transistor for Analog and Digital Biosensing

The advancement of semiconductor technology has popularized the low power, economical and small form-factor solid state devices, such as those highly integrated and interconnected as the fundamental infrastructure for the internet of things (IoT). Due to its CMOS-compatibility and electrical interface, the biosensor utilizing field effect transistor (FET) as transducer has become the perfect candidate to interface directly with the chemical and biological properties of the physical world. Especially, nanowire (NW) FET biosensor has received great attention as a highly sensitive biosensing platform, benefiting from its increased surface-to-volume ratio. In this work, several challenges and key aspects of existing NW FET biosensor were studied, and solutions were proposed to address these problems. For example, the hydrolytic stability of the surface sensing element was evaluated and improved by a hydrolysis process, which led to a significant increase in the overall biosensor performance. Another challenge is the noise in the electric potential of the sensing solutions. A secondary reference electrode was introduced in the biosensing system, and its potential was used to subtract the noise from the measured sensor output. Compared to a reference FET, this approach greatly reduced the system complexity and requirement, yet still improved the limit of detection (LOD) by 50 – 70%. This work also involved careful investigation into the analyte sensitivity, which can be considerably affected by the charge buffering effect from the surface hydroxyl groups. Analytical studies and numerical simulations were carried out, revealing that both low pH sensitivity and large analyte buffer capacity are required to achieve a reasonable analyte sensitivity. The most significant portion of this work was the experimental demonstration of the digital biosensing concept with single serpentine NW FET biosensor. The majority of existing FET biosensors utilized the device as an analog transducer, which measures the captured analyte density to generate an output, and suffers from various noise factors, especially the nonspecific changes of the sensing solutions than cannot be reduced by averaging. Digital biosensor no longer depends on the amplitude of the sensor output and is therefore better immune from these noise factors. Instead, the individual binding event of single analyte is counted and analyzed statistically to determine the analyte concentration. The single serpentine NW FET is the ideal device design to achieve digital biosensing. It maintains the low noise level with the equivalently long channel, yet achieves a small footprint enough for binding of only a single analyte. The binding of analyte to multiple segments of the NW results in both higher sensitivity and binding avidity. The small footprint also enables high integration density of the individual digital biosensors into an array format, which is a responsive, highly sensitive, and cost-effective future biosensing platform.