Towards the Fundamental Thermal Radiation Sensing Limit using Nanomechanical Resonators

Abstract

Temperature-sensitive nanomechanical resonators (NMRs) have been intensively investigated in recent years as promising alternatives to traditional electrical thermal radiation detectors. Unlike traditional electrical detectors which sense thermal radiation through temperature-dependent electrical effects, NMRs undergo thermally induced mechanical resonance frequency shifts upon absorbing thermal radiation. Relying on this mechanical sensing scheme, NMRs hold the promise of reaching the fundamental thermal sensing limit set by temperature fluctuation noise, due to their potential immunity to electrical Johnson-Nyquist noise - the primary limiting factor for traditional electrical detectors.

To date, proposed temperature-sensitive NMRs have primarily focused on optimizing the thermal aspects of the resonators, using methods such as suspending NMRs with ultra-thin structures (e.g., thin rods or tethers) to enhance thermal isolation. However, mechanical frequency noises originating from the ultra-small effective and thermal masses of these NMRs has not been prioritized in the design process or clearly understood in the context of thermal radiation sensing. Consequently, existing nanomechanical thermal detectors have systematically shown not to be limited by fundamental temperature fluctuation noise, nor have they demonstrated significant improvements over state-of-the-art traditional thermal detectors.

In this thesis, we first construct a model that incorporates common noise sources in temperature-sensitive NMRs while accounting for the filtering effects imposed by the closed-loop frequency tracking scheme (i.e., a phase-locked loop, PLL). During our preliminary experiments using SiN membrane NMRs, we observe signs that such NMRs can operate in a regime dominated by fundamental temperature fluctuation noise, but only within a very limited range of sampling frequencies. Later, through a more comprehensive series of experiments, we identify and correct inaccuracies in the initially proposed model. Additionally, this more recent work also unveils a crucial benefit of operating NMRs in a temperature fluctuation noise dominated regime. In particular, we experimentally demonstrate that a hierarchically structured, high-Q-factor NMR can operate in the temperature fluctuation noise dominated regime over an unprecedented measurement bandwidth, thereby maintaining peak thermal sensing performance at faster sensing speeds (i.e., up to 30 times the thermal response time of the NMR).

Lastly, we functionalize a relatively large 3.2 × 3.2 mm square membrane NMR with a localized titanium metasurface absorber, targeting wavelengths ranging 0.5 to 3 terahertz. By striking a balance between frequency stability and thermal responsivity, we experimentally achieve a specific detectivity of 𝐷* ≈ 3.4 × 10⁹ cm · √Hz / W and a noise equivalent power NEP ≈ 36 pW / √Hz , which significantly outperforms the best commercial terahertz pyroelectric detectors. Despite non-idealities in our displacement readout interferometer, this experimentally achieved 𝐷* is well within an order of magnitude of the fundamental thermal detectivity limit imposed by background photon fluctuations (i.e., 𝐷* ≈ 1.8 × 10¹⁰ cm · √Hz / W ). This work, therefore, creates a pathway to achieving the previously unattained fundamental thermal sensing limit at room temperature.