Fig. 6: Long-term stability of sensing performance in humid synthetic air.
From: Neural network enabled nanoplasmonic hydrogen sensors with 100 ppm limit of detection in humid air
a Long-term stability test protocol run at 80 °C sensor temperature and comprised of (i) the standard initialization section, (ii) a first dry section of H2 pulses, (iii) the humidity test sequence comprised of 10 regular and 9 randomized H2 concentration pulse groups (PGs) executed at 80% RH and (iv) a dry control section at the end of the entire protocol. b Correspondingly obtained λpeak sensor readout. c Correspondingly obtained Transformer-based readout, \({{{{{{\rm{c}}}}}}}_{{{{{{{\rm{H}}}}}}}_{2},{{{{{\rm{NN}}}}}}}\). d Extracted Δλpeak as a function \({{{{{{\rm{c}}}}}}}_{{{{{{{\rm{H}}}}}}}_{2}}\) for the 10 regular and 9 random pulse groups, respectively. e Transformer-based readout, \({{{{{{\rm{c}}}}}}}_{{{{{{{\rm{H}}}}}}}_{2},{{{{{\rm{NN}}}}}}}\), as a function \({{{{{{\rm{c}}}}}}}_{{{{{{{\rm{H}}}}}}}_{2}}\), for the 10 regular and 9 random pulse groups, respectively. f LoD evolution along the entire long-term stability test, assessed as a comparison of the LoD values for the initial regular pulse groups in dry (PG1) and humid (PG2) conditions, the last regular pulse group in humid conditions (PG3) and the two final pulse groups in dry conditions (PG4 and PG5). All values are close to an LoD of 0.06% H2 and show no sign of sensor degradation along the entire measurement sequence.
