UTFacultiesTNWEventsPhD Defence Thimo te Molder | Cellulosic Glycols - Identification and prevention of catalyst deactivation

PhD Defence Thimo te Molder | Cellulosic Glycols - Identification and prevention of catalyst deactivation

Cellulosic Glycols - Identification and prevention of catalyst deactivation

The PhD defence of Thimo te Molder will take place (partly) online and can be followed by a live stream.

Thimo te Molder is a PhD student in the research group Sustainable Process Technology (SPT). Supervisors are prof.dr. S.R.A. Kersten and prof.dr. J. Lange, co-supervisor is dr. M.P. Ruiz Ramiro all from the Faculty of Science & Technology (S&T).

This thesis has investigated the conversion of lignocellulosic biomass to ethylene glycol (EG) via tungstate catalysed hydrogenolysis.

The specific goals of this work were: 1) identification and preferably quantification of catalyst deactivation phenomena due to the presence of components (other than holocellulose) in lignocellulosic biomass, and in parallel, 2) the development of suitable pretreatment techniques that can effectively remove the identified catalyst poisons (prevention). Although this thesis specifically targeted the production of EG, the lessons learned are likely relevant to other catalytic routes that target the conversion of biomass to fuels or chemicals using hydrogenation catalysts.

1.1.    Hydrogenolysis protocol

Catalyst deactivation studies were performed by evaluating the glycol and sugar alcohol yield after hydrogenolysis of utnreated and treated biomass substrates and mixtures of relevant model components of the biomass. Microcrystalline cellulose, which is free of impurities, was run as a reference. We developed a hydrogenolysis protocol that, in contrast to the literature, allows decoupling of the acid hydrolysis, aldol-cleavage and hydrogenation functionalities. The protocol offers great flexibility and therefore allows to assign eventual yield deficit to the deactivation of one (or several) of the catalysts. Moreover, we have developed a HPLC method to measure the concentration of soluble sodium polytungstate in the reactor effluent, which is presumably the active species. This method thereby provides a standalone indication of the state of the tungstate catalyst, which in combination with the obtained product yields delivers valuable information on the state of the hydrogenation catalyst.

Industrial processes for the manufacturing of base chemicals typically require a catalyst consumption below 1 tonneProduct / kgCatalyst[1]. For the present hydrogenolysis process, however, the desired catalytic reactions compete with thermal side reactions. Therefore, a minimum amount of catalyst is required to obtain acceptable product yields. For single batch experiments we found that for the tungstate catalyst a maximum biomass-to-catalyst ratio of about 32 g/g was viable. For the hydrogenation catalyst this ratio was about 8 g/g. Unfortunately, these ratios are far off from the industrial window. Although batch experiments have their limitations, they are useful for initial screening.

We found that a high glycol yield was obtained when running in excess of both catalysts. In fact, it turned out that wood powder, millimetre sized particles and “single particles” (Slice size = 22 mm (ø) x 3 mm), gave the same combined glycol yield (~50 wt.%) as a cellulose reference test when an excess of catalyst was applied. This means that the biomass structure, e.g. cellulose accessibility, is not limiting the yield. A low glycol yield should therefore be attributed to catalyst deactivation. We then tuned the hydrogenolysis protocol for maximum sensitivity towards catalyst poisons, i.e. by operating at the threshold catalyst to biomass ratio that is needed for high glycol yields. We can thus consider four potential outcomes: 1) no deactivation, 2) only deactivation of the tungstate catalyst, 3) only deactivation of the hydrogenation catalyst, and 4) deactivation of both catalysts, see Table 1.

Table 1: Hydrogenolysis scenario’s when ran under maximum sensitivity towards catalyst poisons

N

W-catalyst

Hydrogenation catalyst

Dominant product

 

Active?

Active?

 

 

 

 

 

1

EG

2

Sugar alcohol (SA)

3

Humins

4

Humins

Poisoning of the hydrogenation catalyst can always be identified from the product slate. However, deactivation of the tungstate catalyst can only be observed from the product slate when the hydrogenation catalyst is active as both scenarios 3 and 4 lead to the formation of humins, see Table 1. Moreover, solely relying on the product slate can be misleading as the formation of one of the products could be selectively blocked. For example, we observed that lignin present in the feed hampers the hydrogenation of sugars to sugar alcohols (SA), but not that of glycolaldehyde to EG, see Table 2. As a consequence, scenario 6 from Table 2 could be mistakenly interpreted as scenario 3 or 4 from Table 1.

Table 2: Hydrogenolysis scenario's when ran under maximum sensitivity towards catalyst poisons with lignin present in the feed

N

W-catalyst

Hydrogenation catalyst

Dominant product

 

Active?

Active?

 

 

 

 

 

5

EG