< Back to previous page

Project

Towards Single-Molecule Sensing with Silicon bioFETs: Modeling and Experiments

High-throughput single-molecule sensing is expected to entail major advances in the computer industry as well as in the life sciences. A single-molecule sensing platform can enable the widespread and low-cost accessibility to genomic and proteomic information which could start a new era in personalized medicine. In this thesis, we investigate whether such a high-throughput single-molecule sensing platform could be realized with silicon field-effect transistor FET molecular sensors which are closely related to the billions of nano-scaled transistors found on current computer chips. The molecular FET sensor utilizes an electrolyte solution with reference electrode rather than a metal gate to control the conductance of the FET channel. As such charged target molecules present in the electrolyte can be electrically measured as a conductance change or threshold voltage shift of the FET. Since the scalability and integration capabilities of the silicon FET technology are proven, we focus here on the realization of the single-molecule detection limit with a silicon FET sensor.

The main sensitivity-limiting effect for FET sensors is charge screening in the electrolyte. It significantly hinders the achievement of a sufficiently large single-molecule signal-to-noise ratio (SNR). We show both theoretically and experimentally that molecule sensing experiments are typically performed in non-optimal screening conditions. Typical highly charged and pH-sensitive FET oxide surfaces induce enhanced nonlinear electrolyte screening and a pH interference effect. Our simulations show that these effects can reduce the FET sensitivity with two orders of magnitude compared to ideal sensing conditions utilizing a pH insensitive and uncharged FET surface. We obtain a 3× boost in our DNA signal when we reduce the FET oxide surface charge and pH sensitivity by measuring in a lower pH. This confirms the strong impact of nonlinear screening and the pH interference effect experimentally.

Next, we theoretically show using our finite volume simulations that the single-molecule signal strongly benefits from downscaling all FET channel dimensions from the typical micrometer to nanometer range. Due to a channel series resistance effect, the single-molecule signal is inversely proportional to the gate length. However, for small gate lengths (< 50 nm) the inversely proportional trend saturates because of short-channel effects and enhanced nonlinear screening. Downscaling large channel cross-sections also results in a linear increase of the single-molecule signal. However, an optimal cross-section dimension arises for the side to which the molecule binds. This is explained by the increased impact of the convex FET corners which reduce the fraction of the mirrored molecule charge in the FET channel for smaller dimensions. Including calibrated FET noise, we obtain the highest single-molecule SNR for a gate-all-around suspended FET geometry with a gate length of 35 nm and a cross-section of 5×10 nm2. For this geometry and measured in 1.5 mM electrolyte solutions, we predict SNR values of 4.2 and 36 to be possible for a singularly charged and 15 base-pair DNA molecule respectively. This is a promising outlook to use nano-scaled silicon FETs as single-molecule sensors.

Finally, we measure a 17 mV 20-base DNA signal in a 15 mM electrolyte solution with a pH of 4.5. For our smallest measured FET with a gate length of 90 nm and 13 by 20 nm cross-section, this amounts to only ~37 molecules being detected. Assuming a gate length of 50 nm and a 1.5 mM electrolyte solution, the calibrated simulations predict single DNA sensing to become feasible since an SNR of ~6 is obtained.

Date:16 Aug 2017 →  21 Dec 2021
Keywords:Single molecule detection, bioFET sensor
Disciplines:Ceramic and glass materials, Materials science and engineering, Semiconductor materials, Other materials engineering, Sensors, biosensors and smart sensors, Other electrical and electronic engineering
Project type:PhD project