Drift-free ferroelectric photodetection with fast temporal response via thermal diffusion engineering.

Self-powered ferroelectric photodetectors offer an attractive platform for low-power optoelectronics, exploiting intrinsic bulk photovoltaic effects to achieve bias-free operation. However, their practical deployment has been hindered by slow temporal response and photocurrent drift, mainly due to inefficient thermal dissipation that leads to temperature buildup within the device under continuous illumination. Here, we report a thermal-diffusion-engineering approach that reconfigures the thermal environment of ferroelectric device to suppress lateral heat dissipation and enable efficient vertical heat extraction. Compared to conventional architecture, the engineered drift-free device exhibits a photoresponse speed improvement of over three orders of magnitude, along with complete drift suppression, enabling high-fidelity imaging with minimal crosstalk. Infrared thermography and COMSOL simulations confirm distinct thermal environment in conventional and drift-free devices, revealing strong heat accumulation in conventional and efficient heat extraction in drift-free architectures. This work highlights thermal diffusion engineering as a key device-design parameter for enhancing ferroelectric optoelectronic performance, paving the way for scalable, bias-free energy-harvesting photodetectors and neuromorphic imaging systems.