Researchers at the University of the Bundeswehr Munich, with contributing researchers from Friedrich-Alexander-Universität Erlangen-Nürnberg and other international collaborators, have developed a graphene-based transistor that detects the stress hormone cortisol directly in human saliva, with results matching a standard lab test.
The device is an electrolyte-gated graphene field-effect transistor (EG-GFET) built on CVD-grown graphene. The graphene channel is functionalized with perylene bisimide (PBI) linker molecules using the team's own functional layer-assisted transfer (FLaT) method, which both anchors the sensing chemistry and shields the graphene from ambient contamination. Antibody fragments (Fab) targeting cortisol are then covalently attached to the PBI layer, giving the sensor a label-free way to detect the hormone: when cortisol binds to the immobilized Fab fragments, it shifts the transistor's Dirac point, producing a direct electrical readout.
The sensor detected cortisol across a dynamic range of 100 fM to 1 μM in artificial sweat, and down to 1 pM in phosphate-buffered saline, with a calculated limit of detection of 30 fM. In saliva, results came back within about 7 minutes of sample application. To test selectivity, the researchers exposed the sensor to serotonin at 50 times the cortisol concentration and saw a negligible response, indicating the Fab-functionalized graphene channel responds specifically to cortisol rather than other biomolecules.
The team then cross-validated the sensor against conventional ELISA testing, using saliva samples from five participants. At low cortisol concentrations, the two methods agreed closely - in one sample, the graphene sensor measured 0.79 nM versus 0.80 nM by ELISA, a difference of about 1.5%. Agreement was weaker at the higher cortisol levels typical of morning samples, where the sensor tended to underestimate concentration relative to ELISA, which the researchers attribute to the calibration range used and possible saturation effects. The team describes this as the first direct benchmarking of an EG-GFET cortisol sensor against ELISA in real human saliva.
The researchers suggest the platform could be extended to other stress-related biomarkers and integrated into compact, portable readout units for real-time, out-of-lab stress monitoring.
