Deep reaction network exploration of glucose pyrolysis

Area of Science:

• Biomass Pyrolysis
• Computational Chemistry
• Chemical Reaction Networks

Background:

• Understanding biomass pyrolysis is crucial for producing valuable chemicals.
• The reaction network for simple sugars like glucose is highly complex and not fully resolved.
• Existing models struggle with the depth and number of products in glucose pyrolysis.

Purpose of the Study:

• To develop a comprehensive reaction network for [Formula: see text]-D-glucose pyrolysis.
• To identify reaction pathways leading to major experimental pyrolysis products.
• To aid in catalyst design for enhanced biomass conversion.

Main Methods:

• Employed transition-state-guided reaction exploration.
• Utilized graph-based rules and Dijkstra algorithm for identifying kinetically relevant pathways.
• Computed over 31,000 reactions and transition states (semiempirical QM) and ~7,000 (DFT).

Main Results:

• Generated the largest biomass pyrolysis reaction network to date.
• Successfully mapped pathways to most significant experimental pyrolysis products.
• Identified new low-barrier mechanisms resolving experimental discrepancies and explaining high yields of key products like HMF.

Conclusions:

• The developed computational approach provides a complete picture of glucose pyrolysis pathways.
• This method successfully explains yields of key products and resolves experimental anomalies.
• The approach is transferable to other complex biomass pyrolysis reaction network prediction problems.