Selectivity of the First Two Glycerol Dehydrogenation Steps Determined Using Scaling Relationships

Glycerol is a byproduct
of biodiesel production and an abundant
feedstock that can be used for the synthesis of high-value chemicals.
There are many approaches for glycerol valorization, but, due to the
complicated reaction mechanism, controlling which products are produced
is challenging. Here, we describe glycerol’s chemical selectivity
for different metallic catalysts using descriptors for carbon (mainly
*C, *CH2OH) and oxygen (mainly *O, CH3O*). The
quality of these descriptors and the weighted combinations thereof
are validated based on their fit, via linear regression, to the binding
energies of all reaction intermediates generated in the first two
glycerol dehydrogenation steps on a number of close-packed Ru, Co,
Rh, Ir, Ni, Pd, Pt, Cu, Ag, and Au surfaces. We show that *CH2OH is a better descriptor than *C for the studied carbon-bound
intermediates, which is attributed to the observation that the adjacent
*OH group interacts with the surface. This leads to a negative oxygen
dependence, which can be generalized to similar alcohol-derived adsorbates.
Furthermore, we show that CH3O* is a better oxygen descriptor
than *O for the studied intermediates. This is mainly attributed to
the difference between the single and double bonds, as we show that
*OH is closer to the accuracy of CH3O*. Multilinear regression
with different combinations of *C, *O, and *OH is comparable in accuracy
to that of *CH2OH and CH3O*. Scaling relationships
are used to determine the selectivity map for glycerol dehydrogenation.
The results show that the first dehydrogenation is selective toward
two different intermediates (one bonded via the secondary carbon and
the other via the secondary oxygen) depending on the relative bond
strength of the carbon and oxygen descriptors. The second dehydrogenation
step results in five intermediates, again depending primarily on the
relative bond strength of carbon and oxygen to the surface. The selectivity
maps can be used together with kinetic considerations and experimental
data to find catalyst candidates for glycerol dehydrogenation.