The catalytic upgrading of furfural (FAL) to furfuryl alcohol (FOL), furan (FH), and 2-methylfuran (2-MF) is crucial for biofuel production (where F = Furyl radical). FAL, derived from xylose dehydration, serves as a key intermediate, but its reactive carbonyl (-CHO) and furan ring enable multiple reaction pathways, creating selectivity challenges. Experimental studies on palladium-based catalysts show diverse product distributions, while copper favors FOL selectivity with minor 2-MF formation. Platinum primarily facilitates hydrogenation to FOL, with a competing decarbonylation route forming FH. However, the mechanistic basis for these trends remains unclear.
Using Density Functional Theory (DFT), we analyzed the full reaction network on Pd(111), Cu(111), and Pt(111), considering FAL hydrogenation to FOL, decarbonylation to FH (FCO → F + CO), and dehydrogenolysis to 2-MF. On Pd(111), all three products can form due to moderate activation barriers, with FH being the most selective. The rate-determining step (RDS) in FH formation follows the decarbonylation pathway, with an activation energy (Ea) of 107.67 kJ/mol. On Cu(111), FOL selectivity dominates via FCHOH and FCH2O pathways, with the RDS at FCH2O formation (Ea = 72.36 kJ/mol). FH formation is unlikely due to the high FCO barrier (Ea = 114 kJ/mol). On Pt(111), FOL is the major product, with FCHOH hydrogenation as the RDS (Ea = 73.72 kJ/mol). FH forms through FCO (Ea = 102.83 kJ/mol), while 2-MF is suppressed due to the high FCH2 barrier (Ea = 151 kJ/mol).
Our DFT study effectively explains experimental selectivity trends by linking catalytic behavior to reaction energetics. Additionally, we developed Brønsted-Evans-Polanyi (BEP) and transition-state scaling (TSS) relationships for different reaction classes on these surfaces.