Full-dimensional neural network potential energy surface for the multi-channel photodissociation of HNSiO via its S₁ band

Area of Science:

• Chemical Physics
• Theoretical Chemistry
• Photochemistry

Background:

• The HNSiO molecule's photodissociation dynamics are crucial for understanding reactive chemical processes.
• Investigating the first excited state (S₁) provides insights into energy transfer and fragmentation pathways.

Purpose of the Study:

• To construct a full-dimensional potential energy surface (PES) for the S₁ state of HNSiO.
• To investigate the photodissociation dynamics and product distributions of HNSiO in the S₁ state using quasi-classical trajectory calculations.

Main Methods:

• Development of a neural network-based PES using over 91,000 <i>ab initio</i> points.
• Quasi-classical trajectory (QCT) calculations to simulate photodissociation dynamics.
• Analysis of 17 stationary points (5 minima, 12 transition states) on the S₁ PES.

Main Results:

• Identified six potential product channels for HNSiO photodissociation.
• Dominant products (>90%) are HN + SiO and H + NSiO across various energy ranges.
• Product distribution is sensitive to total energy, with H + NSiO favored at higher energies (>7.1 eV).
• Significant isomerization observed due to low energy barriers between isomers.

Conclusions:

• The photodissociation of HNSiO in the S₁ state is complex, with HN + SiO and H + NSiO as major outcomes.
• Isomerization plays a key role, influencing the final product yields.
• The dynamics share similarities with the photodissociation of HNCO in its S₁ band.