(421ab) Growth of Humins During Acid-Catalyzed Carbohydrate Conversion

The acid catalyzed conversion of cellulose produces glucose, fructose and HMF among other intermediates leading to levulinic and formic acids. Levulinic acid is a platform chemical that can be converted into a variety of chemicals, additives or alternative fuels. In addition, the acid catalyzed conversion process generates a dark-colored solid, commonly referred to as humins. The formation of humins is undesirable since it reduces the selectivity for levulinic acid and the humins have little value. The mechanism or mechanisms by which humins form are not understood, and the effect of the composition of the reacting mixture upon their structure has not been studied.

We have first examined the formation of humins during the acid catalyzed conversion of HMF, using SEM, TEM and FTIR to study the solids that are produced. The results are consistent with a mechanism whereby 2,5-dioxo-6-hydroxy-hexanal, formed from HMF, undergoes aldol addition/condensation with HMF, itself, and the growing humin molecule. Consistent with this mechanism, the FTIR spectrum of humins formed from HMF reveal the presence of the furan ring, but the absence of the carbonyl group of HMF. If benzaldehyde is added to the reacting mixture, benzyl groups are incorporated into the humin structure.

The hypothesis that humins grow via aldol addition/condensation of 2,5-dioxo-6-hydroxy-hexanal has kinetic implications. Specifically, 2,5-dioxo-6-hydroxy-hexanal has three carbonyl groups, each of which can participate in aldol addition/condensation. If these three carbonyl groups were equally reactive, one would expect this intermediate to quickly react with three HMF molecules, after which it would cease to grow. Clearly this cannot explain the formation of high molecular weight humins. If humins grow via aldol addition/condensation of 2,5-dioxo-6-hydroxy-hexanal with a growing humin molecule, then the carbonyl groups of that growing humin molecule must be less reactive than the carbonyl groups of 2,5-dioxo-6-hydroxy-hexanal.

Unfortunately, this kinetic consequence of the proposed mechanism can not be tested through kinetics experiments for reasons that will be discussed. However, the first step in the acid-catalyzed aldol addition/condensation reactions is the formation of an enol from an aldehyde or ketone with an available alpha hydrogen. We have used DFT to estimate the free energy of enol formation involving each of the carbonyls. The free energy of formation of enols from 2,5-dioxo-6-hydroxy-hexanal and from the condensation product of 2,5-dioxo-6-hydroxy-hexanal and HMF that involve C2 and C3 are considerably smaller than any of the other possible enols. This is consistent with the proposed mechanism for humin growth.

We have additionally examined the humins formed from glucose and from fructose. The infrared spectra are quite similar in all three cases. This suggests that 2,5-dioxo-6-hydroxy-hexanal, formed from HMF, is the key intermediate in all three cases. Indeed, the spectra of the humins formed from the hexoses indicate the presence of furan rings. This is not particularly surprising, as one might not expect direct conversion of the hexoses to humins, at least not via aldol addition/condensation. The hexoses can assume three primary conformations, two are ring structures that do not possess carbonyl groups (and consequently can not participate in aldol reactions). The third form is a linear aldehyde or ketone that can participate in aldol reactions. However, over 99% of the hexoses exists in the ring form, so it is expected of the reagents that could react with 2,5-dioxo-6-hydroxy-hexanal, HMF will be the most abundant. This is consistent with the observation of a furan functionality in the humins formed from glucose and fructose.

There are also differences in the morphology of the humins, depending upon the reaction environment. In most cases we have studied, the humins take the form of spherical particles. However, at high carbohydrate to acid ratios, spheres were not observed. Addition of other components to the reacting mixture (e. g. other aldehydes) can also affect the morphology of the humins.

Ideally, one would prefer to completely suppress the formation of humins, thereby maximizing the yield of levulinic acid. If this can not be accomplished, though, an alternative approach is to modify the humins that do form so that they have a higher value. In this regard, two observations from this study are particularly significant: the morphology of humins can be modified by variation of reaction conditions and it is possible to incorporate additional chemical functionalities in the humins that form. (The latter was demonstrated by the incorporation of benzyl groups in humins formed from HMF.) Eventually these findings might be developed to allow the production of humins with a specific morphology and a specific chemical functionality.