TY - JOUR
T1 - Physical Virtualization of a GFET for a Versatile, High‐Throughput, and Highly Discriminating Detection of Target Gas Molecules at Room Temperature
AU - Zanotti, Michele
AU - Freddi, Sonia
AU - Sangaletti, Luigi Ermenegildo
PY - 2024
Y1 - 2024
N2 - An e-nose has been built on a single graphene field effect transistor (GFET), based on a graphene/Si3N4/Si stack of layers. Multichannel data acquisition, enabling to mimic the architecture of a sensor array, was achieved by steering the gate potential, thus yielding a virtual array of 2D chemiresistors on a single sensing layer. This setting allowed for the detection of volatile compounds with a remarkable discrimination capability, boosted by intensive machine learning analysis. Sensing of gas phase NH3 was tested, along with a set of possible interferents, and discrimination of NH3+NO2 mixtures was successfully probed. High throughput in terms of sensitivity was achieved by tracking the shift of the minimum of the GFET transfer curve vs. NH3 concentration. With this readout scheme, a 20-fold sensitivity increase over a 5-50 ppm range was registered with respect to the same layer used as a chemiresistor. High discrimination capability was probed by leveraging on machine learning algorithms, from PCA to u-MAP and, finally, to a deep NN where input neurons are the virtual sensors created by the gate voltage driving. For the tested case, the DNN maximum accuracy was achieved with 21 virtual sensors.
AB - An e-nose has been built on a single graphene field effect transistor (GFET), based on a graphene/Si3N4/Si stack of layers. Multichannel data acquisition, enabling to mimic the architecture of a sensor array, was achieved by steering the gate potential, thus yielding a virtual array of 2D chemiresistors on a single sensing layer. This setting allowed for the detection of volatile compounds with a remarkable discrimination capability, boosted by intensive machine learning analysis. Sensing of gas phase NH3 was tested, along with a set of possible interferents, and discrimination of NH3+NO2 mixtures was successfully probed. High throughput in terms of sensitivity was achieved by tracking the shift of the minimum of the GFET transfer curve vs. NH3 concentration. With this readout scheme, a 20-fold sensitivity increase over a 5-50 ppm range was registered with respect to the same layer used as a chemiresistor. High discrimination capability was probed by leveraging on machine learning algorithms, from PCA to u-MAP and, finally, to a deep NN where input neurons are the virtual sensors created by the gate voltage driving. For the tested case, the DNN maximum accuracy was achieved with 21 virtual sensors.
KW - Artificial intelligence
KW - Electronic nose
KW - GFET
KW - Gas sensing
KW - Artificial intelligence
KW - Electronic nose
KW - GFET
KW - Gas sensing
UR - http://hdl.handle.net/10807/294957
UR - https://onlinelibrary.wiley.com/doi/10.1002/admt.202400985
U2 - 10.1002/admt.202400985
DO - 10.1002/admt.202400985
M3 - Article
SN - 2365-709X
SP - N/A-N/A
JO - Advanced Materials Technologies
JF - Advanced Materials Technologies
ER -