Research on the migration and transformation law of sulfur in a molten salt heating tire pyrolysis reactor: Experiments and neural network field-driven molecular dynamics simulations

The molten salt heating tire pyrolysis reactor is characterized by high thermal efficiency, making it promising for application in the tire pyrolysis industry. However, there is a lack of prior research on the characteristics of its products, particularly regarding the migration and transformation patterns of sulfur within those products. This study focuses on the sulfur transformation characteristics in the products of the proposed molten salt heating tire pyrolysis reactor. Using sulfur-containing cross-linked natural rubber molecules as the model, a neural network potential was trained based on the results, and molecular dynamics simulations of the pyrolysis process were conducted using this potential to clarify the sulfur transformation patterns. The results indicate that approximately 60% of the sulfur remains in the pyrolytic char after the pyrolysis process, around 35% exists as sulfur-containing compounds in the pyrolysis oil, and less than 5% is present in gaseous form in the pyrolysis gas. The sulfur in the char from waste tire pyrolysis primarily exists in the form of aliphatic sulfur, and as the pyrolysis temperature increases, the proportion of aliphatic sulfur in the char gradually decreases. Both H₂S and SO₂ concentrations increase with rising pyrolysis temperatures, with the concentrations of H₂S and SO₂ rising from 836 µL/g and 74 µL/g at 450 °C to 987 µL/g and 101 µL/g at 525 °C, respectively; the concentration of H₂S is significantly higher than that of SO₂. The atomic forces calculated from the neural network potential align well with those obtained from DFT calculations, with a mean absolute error (MAE) of 0.381 eV/Å and a root mean square error (RMSE) of 0.607 eV/Å.