Search for a command to run...
Vertical-cavity surface-emitting lasers (VCSELs) are the dominant light sources in short-reach optical interconnects, where cost, efficiency, and scalability are critical. However, their modulation bandwidth, output power, and signal integrity degrade markedly as ambient temperature rises and self-heating increases, making accurate device-level temperature awareness indispensable. Existing approaches rely on embedded sensors or forward-voltage monitoring, which require calibration, additional hardware, or pilot overhead, and are therefore not well suited for in-service operation. This work introduces a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">pilot-free</i> and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">sensor-free</i> method for inferring VCSEL operating temperature directly from payload signals. We establish, through an electro–thermal rate-equation model, that temperature rise manifests as a systematic reduction in the entropy of the optical waveform. Leveraging this property, we develop a regression-based estimator that achieves sub-<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$5^{\circ }$</tex-math></inline-formula>C accuracy in simulation. The results demonstrate that entropy-based payload analysis provides a principled and low-cost proxy for junction temperature, with potential for integration into high-speed link management.