Most important for our bio detection experiments is the vari

Most important for our bio-detection experiments is the variation of the transmittance with sheet conductance-a quantity we call the sensitivity factor and estimate as the first derivative . By visual inspection of Fig. 2 we expect the sensitivity factor to be much greater near the peaks of T than near the valleys. This is quantified graphically in Fig. 2(b) where we plot the computed vs and parametrized by frequency around the first-nonzero-frequency peak (fundamental resonance) in 2(a) centered at 112GHz. For , we see drops by 5× between 112 and 145GHz. This behavior is consistent with the fact that the peaks in T are points of perfect constructive interference of successive internal partial waves, so are most sensitive to small changes in the reflectivity of the graphene interface.
Experiments Fig. 2(b) guided our design of the experimental sensor apparatus shown schematically in Fig. 3. It consists of a waveguide-mounted Gunn-oscillator operating at a fixed frequency of 101GHz just 11GHz below the fundamental resonance of the graphene–silicon etalon. It is square-wave amplitude modulated with a power-MOSFET circuit, radiated horizontally through a pyramidal horn antenna that feeds an off-axis paraboloid which directs and focuses the beam downward. The focused beam is mode-matched to a second feedhorn that collects the radiation into a Schottky-rectifier receiver. The GFET structure is then located in the beam path just above the receive feedhorn where the spot size is approximately 5mm. The output signal from the Schottky rectifier is fed to a 1000×-gain low-noise voltage amplifier, and then demodulated with a lock-in amplifier synchronized to the square wave. A waveguide attenuator between the Gunn oscillator and the feedhorn allows the received signal to be increased to the maximum possible output signal-to-noise ratio (SNR) before the onset of p53 tumor suppressor and other nonlinearity. Typically, this background SNR was ∼60dB.
Results and discussions The GFET structure was operated with the backgate bias of from a low-noise power supply, and a drain-source constant-voltage bias of from a Keithley 2400 source meter. This backgate bias was chosen because of its proximity to the Dirac-point bias of ∼30V, which was attainable but only with a large degree of fluctuation in the drain-source current. The 25V backgate bias creates a significant background electron sheet concentration in the graphene, and the is just high enough to allow accurate measurement of the sheet conductance without excessive drift and 1/f noise. Because the graphene geometry between the S and D electrodes is approximately square (area ), the absolute DC sheet conductance is . For the specific GFET tested we measured an of 0.2132mA, or sheet conductance of 2.132mS. The corresponding background 101-GHz transmittance signal was where the last digit is significant given the high SNR. To validate the measurement technique and assess its accuracy, we used the backgate to induce a known-change in graphene DC sheet conductance and compare this with the derived change of the 101-GHz quantity. We applied backgate voltages of and to straddle the nominal +25V and allow for mean-value estimation. The DC current values at the two gated voltages were 0.2017 and 0.2255mA, respectively, corresponding to . The lock-in signals for the same backgate voltages were and corresponding to a transmittance difference of . Division by then yields in good agreement with . The biodetection protocol was to apply 13-mer single-stranded DNA solutions of three different molarities (0.01, 1.0, and 100nM) sequential at 900-s intervals, starting with the 0.01nM solution. A drop of each was placed directly on the graphene with a syringe, allowed to settle for 300s, and then blown dry with an oil-free air gun. The Keithley-2400 DC current was recorded simultaneous with the THz transmitted signal via the output from the lock-in amplifier. The experimental results for DC current are shown in Fig. 4(a) where we see an initial value of 222μA between ≈600 and 900s, corresponding to a sheet conductance of 2.22mS. Then a large fall in DC current occurs with the application of all three drops, and a lesser fall upon blow drying 300s later. Both effects are most pronounced with the 0.01nM solution and become progressively weaker with the other two. The 1.0 and 100nM drops have a significant effect in their aqueous form but little further change occurs upon drying. In all cases, however, at constant bias voltage the DNA decreases the DC sheet conductance of the graphene.