燃料中の硫黄分を機械学習を用いたRGB画像解析により単段抽出・プラズモンセンシングする手法

Mercaptans are compounds of great concern because they are one of the main sources of sulphur in fuels and can reduce engine performance. Diesel-biodiesel blends have partially replaced diesel fuel as part of efforts to promote greener energy. In Brazil, the current legislation permits up to 10 ppm of sulphur in diesel. Considering the impact of sulphur compounds produced within combustion engines, careful monitoring of fuels is essential to ensure a greener future, and methodologies to easily detect sulphur are needed, especially if they can be adapted for in situ measurements and are suitable for upcoming regulatory changes.

In this study, gold nanoparticles embedded in glycerol were used to extract, preconcentrate, and detect sulphur in biodiesel samples, with a mercaptan as sulphur source. The color changes due to nanoparticle aggregation were measured and processed using a smartphone app and machine learning to quantify sulphur. We demonstrate that this approach is suitable for determining this species in biodiesel at levels as low as 1.39 ppm, with the AuNPs@GLY-to-fuel ratio being easily adjustable, achieving lower limits of detection. Also, through SERS spectroscopy, we have demonstrated that the working principle behind the detection is the strong interaction between the thiol moiety and gold. This methodology can be easily customized, offering an interesting advantage from a practical perspective, as sulphur limit values vary worldwide and are influenced by the energy transition, which is a priority for public policies.

The primary feature of our system is that it combines extraction, preconcentration, and sensing in a single step, eliminating the need for sample preparation or conditioning, and bringing the analysis to the field. The herein-developed sensor is low-cost, user-friendly, and can be customized and adapted to probe sulphur in commercial diesel, as verified here by SERS spectroscopy.