Micromechanical Bolometers for Subterahertz Detection at Room Temperature

Fast room-temperature imaging at terahertz (THz) and subterahertz (sub-THz) frequencies is an interesting technique that could unleash the full potential of plenty of applications in security, healthcare, and industrial production. In this Letter, we introduce micromechanical bolometers based on silicon nitride trampoline membranes as broad-range detectors down to sub-THz frequencies. They show, at the longest wavelengths, room-temperature noise-equivalent powers comparable to those of state-of-the-art commercial devices (∼100 pW Hz–1/2), which, along with the good operation speed and the easy, large-scale fabrication process, could make the trampoline membrane the next candidate for cheap room-temperature THz imaging and related applications.


Mechanical, thermal and optical parameters of Si3N4 and Cr/Au layers
The following tables provides a list of material parameters which are relevant for this work. As explained in the Methods section of the main text, the COMSOL simulations were conducted with a single 35 nm thick Au film, rather than a 5 Cr-30 Au two-layer structure. For this reason, Table S1 contains mechanical and thermal parameters of Si3N4 and Au alone, while Table S2 the optical absorption of the complete Cr/Au film. The value of the Intrinsic Stress displayed in Table S1 was not known a-priori, and was therefore adjusted to match simulated and experimental resonance frequency of the membranes. The radiation absorptions reported in Table S2 were calculated from the real and imaginary parts of the refractive indices of the materials. Si3N4 was assumed to be completely transparent for all frequencies under consideration 1 .

S2
Calculation of thermal conductance, heat capacity and total mass Using the material parameters of Table S1, we calculated the individual contribution of the Si3N4 and the Au layers to the thermal conductance, heat capacity and total mass of both devices. It clearly emerges how the largest contribution to heat dissipation is given by the metal layer, whereas the Si3N4 layer mostly provides thermal mass to the device. The ratio between the thermal conductance and the heat capacity gives an approximate value for the thermal response frequency of the devices, resulting in fM1~12 Hz and fM2~22 Hz.

Infrared laser responsivity for device M2
The following figure shows the linear responsivity of device M2, as function of the impinging infrared laser power, similarly to Figure 2 of the main text. As stated in the main text, the responsivity is 187 kHz/W. Given the resonance frequency at zero incident power f0=91.26 kHz, the normalized responsivity is 2.05 W -1 .

Infrared laser responsivity for device M2 without Cr/Au metal layer
The following figure shows responsivity of device M2, measured before the deposition of the Cr/Au metal layer, as function of the impinging infrared laser power. Here, the responsivity is 8 kHz/W, approximately 23 times smaller than the responsivity after metal deposition.

Green laser responsivity
The linearity of device responsivity was also tested using the green laser (532 nm), adding Neutral Density filters with increasing opacity directly in front of the source. The shift of the resonance peak was measured similarly to the infrared case, sweeping the excitation frequency sent the piezo actuator and reading the self-mixing signal via a single lock-in channel. The infrared laser used as probe for the self-mixing readout was kept slightly above lasing threshold (1.5 mW) during the measurement. The obtained green laser responsivities are 520 kHz/W and 600 kHz/W for device M1 and M2, respectively. The latter value (M2) is compatible with the measurement of phase responsivity described in the main text.

Spot size measurement of the sub-THz source
We measured the spot size of the sub-THz source, after the reflection in the parabolic mirror, mapping the 2D intensity of the beam with a calibrated pyrometer. The map was then fitted with a 2D Gaussian profile, convoluted with the active area of the pyrometer, given by a disk of radius 5 mm. The beam shape is very asymmetric in the membrane plane, elongated along the x-direction, because of the reflection in the parabolic mirror. In particular, 2σx= 7 .1 mm and 2σy= 4.3 mm.

Spot size measurements of green and infrared lasers
We measured the spot size of both green and infrared laser (intended as twice the standard deviation 2σ of a Gaussian beam profile) using the knife edge method. Self-mixing signal amplitude as function of the piezo actuator voltage We verified the linear response of the self-mixing signal, read after the lock-in amplifier, as function of the piezo driving voltage. The following figure refers to the measurements performed on device M1.