Resumen de Towards reliability enhancement of graphene FET biosensor in complex analyte: Artificial neural network approach
Joyeeta Basu, Nirmalya Samanta, Sukhendu Jana, Chirasree RoyChaudhuri
In this paper, reliable biomolecule sensing in serum has been reported with a detection limit of 0.1 fM using graphene nanogrid field effect transistor (FET) biosensor, coupled with artificial neural network. Graphene nanogrid biosensors have been earlier reported to sense down to 0.1 fM concentration of target antigen in buffer but there was no attempt of detection in any physiological analyte. It has been observed that the drain current sensitivity overlaps significantly amongst the different concentrations of Hep-B surface antigen in serum, especially near the lower range and hence quantification becomes difficult. To address this problem, capacitive mode of measurement has also been coupled and the notch frequency obtained from this measurement has been extracted. This frequency value along with the drain current sensitivity has been processed by neural network architecture with back propagation algorithm to quantify the concentration of Hep-B target antigen. It has been observed that the system is capable of quantifying Hep-B surface antigen in serum with an error of around 5% in the range of 0.1 fM to 1 pM. The detection limit achieved is an improvement by at least three orders of magnitude compared to the recent reports of graphene field effect sensing in serum.
