Based on our previous discussion of the microwave transport properties in GR-FET devices
], the possibility to utilize GR for THz detection has become a more practical goal. Following the previously discussed approach, a clear response to THz radiation has been observed using the setup shown in Figure
. The fluctuations in the response of the device can be explained by considering the influence of bolometric and nonlinearity effects within the GR material. Exposure to THz radiation will inevitably induce these effects depending on the nature of the sample, whether it is monolayer with semimetallic behavior or bilayer with semiconductor behavior, resulting in a change in the resistance. Referring back to the original resistance's room temperature dependence in Figure
, the outcome of Figure
can be understood to be the result of a strong bolometric response that increases the resistance in the metallic-type devices and decreases the resistance in the semiconductor-type devices. In addition, nonlinearity effects play an important role in influencing the response of semiconductor-type devices to THz radiation. Nonlinear response occurs because the band gap excitation energy matches the incident wave frequency.
Transitions between THz ON and OFF exposure states change the resistance values in a manner that can be explained by bolometric and nonlinearity effects for both monolayer and bilayer devices. The flat regions of the curves within the first four cycles for sample 3 and the first three cycles for sample 2 show the transitions in the responses between the expected bolometric response and occasionally the nonlinear response. After a short period of time, the response is completely dominated by bolometric effects. To clarify the real bolometric impact, the blue background is subtracted to show the absolute resistance change. Fluctuation amplitude can be clearly seen in Figure
]. The observed results show a clear distinction between the response of single- and bilayer devices in sensing THz radiation.
Figure 5 Resistance fluctuation and amplitude response for THz irradiation. Most cycles follow a bolometric response, while a few cases follow a nonlinear response. The nonlinear response arises from the excitation of extra carriers which is reflected as an opposite (more ...)
The main aspects of characterization were indicated by the small arrows in the previous response curves of Figure
; the arrows simply indicate two sets of information. The first aspect is the change in the average resistance value for the transition from the THz-OFF state to the THz-ON state. The second aspect is the instantaneous value of the resistance at the two moments where THz radiation starts and the moment where THz radiation is terminated.
Furthermore, looking into the data analysis, sample 3 (metallic type) and sample 2 (semiconductor type) started in the THz-OFF state for 3 min where the average fluctuation amplitude was estimated to be 0.03 and 0.15 KΩ, respectively. Pulsed THz radiation was applied for 3-min intervals, as indicated by the gray-shaded regions in Figure
. The devices' bolometric response to THz radiation is reflected by the correlating resistance amplitude fluctuations. Examining Figure
, the differences in fluctuation amplitudes show a clear variation between complete THz-OFF and THz OFF-ON states. Metallic characteristics are observed for sample 3 after three successive cycles of exposure with an amplitude increase of 0.05 KΩ. Conversely, sample 2 shows semiconductor characteristics after two successive cycles of exposure with an amplitude decrease of 0.40 KΩ. The fluctuation amplitudes increase by a factor of 2 relative to the original THz-OFF state. Cycle 4 for sample 3 and cycle 3 for sample 2 show opposite responses since the change due to THz-ON radiation does not fade out with the THz-OFF state. Consequently, the response shows a linear growth for the fluctuation amplitudes. The metallic sample's average fluctuation amplitude increases by 0.08 KΩ during the THz-ON state, while the semiconductor sample's average fluctuation amplitude decreases by 0.65 KΩ during the THz-ON state. The fluctuation amplitudes changed by a factor of 3 relative to the original THz-OFF state. These trends can be observed in comparison to the original fluctuation as shown in Figures
. Transitions in response occur in correspondence to the opposite response observed in cycle 4 of sample 3 and cycle 3 of sample 2, as shown in Figure
Figure 6 Comparison of the resistance response between THz OFF-ON states and the complete THz-OFF state. The THz-OFF measurement was taken for 10 min and plotted as the blue curve. The same measurement is also fitted on the OFF-ON state measurement to indicate (more ...)
Finally, the efficiency of inducing the thermal energy required to observe a bolometric response has been related to the sample's domain size at the core of the antenna structure. Figure
shows the response of sample 4, which was obtained by CVD. The domain size of sample 4 is 10 mm2 and is 4 orders of magnitude larger than that of the exfoliated samples. Following a similar approach as described previously, the sample started in the THz-OFF state for 5 min where the average fluctuation amplitude was estimated to be 10 Ω. The tendency for bolometric response is reflected by the observed fluctuation amplitudes of the resistance. The differences in fluctuation amplitudes show the variation between complete OFF and ON states. Sample 4 shows a metallic characteristic with a fluctuation amplitude of 20 Ω, which reflects an increase by a factor of 2 relative to the original THz-OFF state.
Figure 7 Response of sample 4 (CVD, monolayer GR) to THz radiation. Due to a large sample size domain of 10 mm2, higher thermal energy is required to induce a sufficient bolometric response. The red solid line shows the actual data. The blue solid line shows the (more ...)
Overall, this experiment reveals the interplay of different photoresponse mechanisms primarily involving rectification due to THz radiation in the presence of nonlinearity and bolometric heating effects on the transport properties of GR-FET devices. The observation of such bolometric responses, especially at ultrahigh frequencies, is a highly prized characteristic for a variety of device applications. Similarly, such a response has been observed for GaAs
], which confirms the bolometric behavior observed in the GR-FET device, even at ambient conditions.
Realizing the need to improve our measurement setup, several modifications to the sample box shown in Figure
were made in order to extend the detection limit of our device. Modifications, such as suspending the device using Cu/Au wires rather than having it rest on an insulating substrate, were found to greatly reduce parasitic capacitance and increase the detection limit of the device. As discussed previously
], using SMA connectors presented a major limitation in the previous setup and affected the total response cutoff. In our recent attempt, SMK connectors and cables were used which have a higher cutoff frequency at 40 GHz. Therefore, the device response was predominantly limited by surface wave resonance effects from the metal plate stage and the lead contacts as demonstrated in Figure
. The device response shows possible conduction modes for the GR device up to 50 GHz, indicating that the ‘yield’ has drastically increased. At higher frequency regimes, a greater gain in amplitude relative to the starting point is observed, showing that the transport modes dominate the device performance as shown in Figure
The GHz transmission setup. (a) Modified box with floating lead connections that have drastically reduced the surface resonance effects and parasitic capacitance. (b) Frequency response profile for the transmitted signal up to 40 GHz.