The discovery eliminates the need for bulky external sensors and could make future photonic chips—used in data centers, telecom, and quantum systems—smaller, cheaper, and far more stable.

A tiny tweak in chip design could usher in a big leap for light-based data processing. Researchers at Columbia University have discovered that a thin-film metallic resistor—already used for tuning photonic circuits—can double as a precise, on-chip thermometer.
The world of integrated photonics—where light, rather than electricity, carries and processes information—has hit a thermal snag. Slight temperature shifts can throw off the delicate resonance frequencies of photonic devices, disrupting performance. Until now, the fix has been bulky external sensors. But by re-imagining a common thin‐film platinum resistor that already exists in many photonic chips, the team has shown that the resistor’s temperature-sensitive electrical behavior can be harnessed for real-time thermal monitoring and control.
In the experiment, the researchers placed a platinum thin-film resistor directly above a high-Q photonic microcavity. They observed that as the device heated or cooled, the resistor’s electrical resistance shifted in a measurable way. That essentially turned the resistor into a built-in thermometer—no extra components needed. They then locked a commercial distributed-feedback laser to the cavity and kept the output wavelength stable to within one picometre over two full days.
The innovation is foundry-compatible and platform-agnostic: it can be adopted in different chip configurations, including silicon ring modulators and quantum photonic systems. The built-in thermometer promises to shrink the size and cost of photonic systems by removing external sensors, and could accelerate the deployment of photonic devices for data centres, telecom, and quantum computing.
By leveraging a component that’s already in the chip, the researchers solved a persistent thermal issue in photonics with elegance and efficiency. It’s a smart tweak with big implications—smoother integration of photonic and electronic circuits, lower system overhead, and faster progress toward scalable, real-world photonic applications.









