Crystal Facet Engineering-Photoexcitation-Machine Learning Synergy on Cu₂O/CuO Heterojunctions for High-Performance...

Triethylamine (TEA), a typical biogenic amine indicative of protein spoilage and a common toxic chemical pollutant, requires highly sensitive detection to ensure both food safety and occupational health. However, existing TEA sensors are often limited by low response and high operating temperatures. To address these limitations, we synthesized Cu₂O/CuO heterojunctions in situ via a semisacrificial template method. By modulating the molecular weight of polyvinylpyrrolidone (PVP), we precisely controlled the growth of Cu₂O {111} and {100} facets. This approach enabled the controllable synthesis of Cu₂O/CuO nanomaterials with distinct morphologies-cubes, truncated octahedrons, and octahedrons-resulting in significant differences in the exposed crystal facet. At 80 °C, the response of Cu₂O/CuO-1.5 to 100 ppm TEA increased by 1.63-fold under photoexcitation, from 1229 to 2007. At room temperature, photoexcitation enhanced the response of Cu₂O/CuO-1.5-10000 to 100 ppm TEA by 1.5 times (from 80 to 120) and shortened the response/recovery times from 22 s/58 s to 20 s/42 s. Nevertheless, the sensor remained susceptible to humidity interference. To mitigate this, we systematically evaluated multiple machine learning models and selected the extra trees algorithm, achieving high-precision humidity compensation (R² = 0.9966). This work provides an effective strategy for high-performance TEA detection in complex environments by integrating crystal facet engineering, photoexcitation, and machine learning.