Similar to the simulations carried out in vacuum (14), ab initio MD simulations of xy­lose degradation in water started when a protonated (3-D-xylose was positioned in a unit cell surrounded by 32 water molecules (43). During the course of our entire simulation (~5 ps), the initiation proton attached to the -OH groups on the sugar ring was ob­served to be transferred back to the surrounding water molecules. This transfer is rapid and occurs in less than 100 fs for all of the hydroxyl groups on the xylose ring. Once the proton was transferred to the neighboring water molecule, it is quickly transferred to other water molecules and away from the sugar molecule. This result shows that proto­nation is probably the rate-limiting step in sugar degradation under acidic media because our earlier simulations in vacuum demonstrate that protonated (3-D-xylose molecules de­compose rapidly. Figure 9.4 shows the proton transfer from xylose C2-OH to the sol­vent water molecules from our simulations. The simulations started with a protonated xylose molecule surrounded by 32 water molecules. After 34 fs, a neighboring water molecule forms a bond with the protonated hydroxyl group (C2-OH+-OH2) as shown in the figure. After 62 fs, the proton from the C2-OH transfers to a water molecule, forming an H3O+. After the proton was transferred from the xylose hydroxyl group to the water molecule, it was quickly transferred to other water molecules and away from the xylose molecule due to the strong hydrogen bonding interactions between the water molecules and the high proton mobility. It appears that protonation is a slow rate-limiting step.

In order to test the notion that protonation is the rate-limiting step, reaction barriers have been estimated using the hybrid density functional B3LYP (18). A single water molecule has a lower proton affinity (44), PA = 165 kcal mol-1, than xylose (Table 9.1 PA = 186.7-191.3 kcal mol-1), but the MD simulations appear to indicate that a proton will be transferred from xylose to a neighboring water cluster. CBS-QB3 calculations (18) show that for water clusters, the proton affinity increases with cluster size due to the increased stability of the hydronium ion. These calculations show that for a four-molecule cluster, the proton affinity has increased to 220.2 kcal mol-1. This and the MD simulations suggest that bulk water will have a higher proton affinity than xylose. The high proton affinity of bulk water relative to xylose must then add energy to the intrinsic energy barrier for the dehydration of xylose. This could explain the discrepancy between the energy barriers calculated in vacuum shown inTable9.1 (~16kcalmol-1 for the formation of furfural) and the experimentally observed barriers in solution, about 32 kcal mol-1. This was investigated further with static electronic structure calculations using B3LYP. Energy barriers for the initial step of xylose conversion to furfural were calculated in the presence of water clusters. With water clusters present, the barrier for xylose dehydration increased (18) to around 30 kcal mol-1, which is consistent with experimental values. This suggests that the experimentally measured barrier for xylose dehydration contains an energy contribution due to transferring the proton from the solvent to the substrate.