Localized Thermal Gradients On-Chip by Radiative Cooling of Silicon Nitride Nanomechanical Resonators

TL;DR

This study demonstrates on-chip radiative cooling of silicon nitride nanomechanical resonators, creating localized thermal gradients that could power microscale renewable heat engines, with potential for significant temperature drops.

Contribution

We experimentally show radiative cooling of a silicon nitride nanomechanical resonator on-chip, enabling localized thermal gradients for energy harvesting applications.

Findings

Achieved temperature drops of up to 9.3 K during the day and 7.1 K at night.

Successful radiative cooling observed across multiple experiments and conditions.

Model predicts potential cooling of up to 67 K with improvements.

Abstract

Small scale renewable energy harvesting is an attractive solution to the growing need for power in remote technological applications. For this purpose, localized thermal gradients on-chip--created via radiative cooling--could be exploited to create microscale renewable heat engines running on environmental heat. This could allow self-powering in small scale portable applications, thus reducing the need for non-renewable sources of electricity and hazardous batteries. In this work, we demonstrate the creation of a local thermal gradient on-chip by radiative cooling of a 90 nm thick freestanding silicon nitride nanomechanical resonator integrated on a silicon substrate that remains at ambient temperature. The reduction in temperature of the thin film is inferred by tracking its mechanical resonance frequency, under high vacuum, using an optical fiber interferometer. Experiments were conducted on 15 different days during fall and summer months, resulting in successful radiative cooling of the membrane in each case. Maximum temperature drops of 9.3 K and 7.1 K are demonstrated during the day and night, respectively, in close correspondence with our heat transfer model. Future improvements to the experimental setup could improve the temperature reduction to 48 K for the same membrane, while emissivity engineering potentially yields a maximum theoretical cooling of 67 K with an ideal emitter.