Kraft, Johannes (2017)
Heterogeneously catalysed hydrogenolysis of glycerol to 1,3-propanediol.
Technische Universität Darmstadt
Ph.D. Thesis, Primary publication
|
Dissertation Johannes Kraft Endfassung -
Text
Dissertation Endfassung.pdf - Accepted Version Copyright Information: CC BY-SA 4.0 International - Creative Commons, Attribution ShareAlike. Download (7MB) | Preview |
Item Type: | Ph.D. Thesis | ||||
---|---|---|---|---|---|
Type of entry: | Primary publication | ||||
Title: | Heterogeneously catalysed hydrogenolysis of glycerol to 1,3-propanediol | ||||
Language: | English | ||||
Referees: | Claus, Prof. Dr. Peter ; Vogel, Prof. Dr. Herbert | ||||
Date: | 2017 | ||||
Place of Publication: | Darmstadt | ||||
Date of oral examination: | 16 October 2017 | ||||
Abstract: | The strong increase in the biodiesel production in recent years has immensely increased the availability of glycerol as a building block for the chemical industry. However, the selective conversion of glycerol to high-value products like 1,3-propanediol is still difficult and lacks efficiency, resulting in this process being a bottle-neck for the ready availability of valuable high tech products like poly trimethylene terephthalate (PTT), an important polymer for the textile industry. With derivative products like this, the percentage of materials based on renewable resources could be increased significantly, contributing to the diminution of the dependence on fossil resources and to the containment of the climate change. The conversion of bio-based glycerol is already performed on industrial scale as a biotechnological process, burdened with typical drawbacks of these kinds of processes like low space-time-yields, the need for co-reactants etc. Also, this one as well as actual heterogeneously catalysed industrial processes suffers from yields far below 100%. In principle, heterogeneous catalysts should be able to solve these problems, however, the selective removal of the secondary hydroxyl group has turned out to be far more complex than the formation of 1,2-propanediol and the performance of literature catalysts still leaves a lot to be desired. In order to construct better catalysts, the functionality of the catalyst and the reaction mechanism needs to be understood. So far, several reaction mechanisms have been proposed for the variety of catalysts tested on this reaction. In most cases, an acid-catalysed mechanism, either with or without hydrogen spillover, has been assumed whereas a direct hydrogenolysis mechanism was proposed for state-of-the-art bimetallic iridium rhenium catalysts. Based on publications right before the beginning of this study, this type of catalyst has been chosen as one of the starting points for the investigation of the mode of operation of the catalyst and the reaction mechanism. Besides, several attempts for innovative catalysts, e.g. using tungsten carbide, other tungsten components or different hydrogenating metals, have been made during the first phase of this work, but, unfortunately, none of them achieved promising results. Therefore, this work focussed then on the investigation of the reaction mechanism. As the results with Ir-Re catalysts published by the Tomishige group could not be reproduced and the reaction conditions, especially the very low amount of material compared to the size of the reactor (6 g of glycerol solution in a 190 mL reactor), seemed very impractical, new standard conditions were defined in a series of preliminary tests. Most of the reactions of this study have been performed for 8 h at 433 K and 5 MPa of constant hydrogen pressure, enabled by an external hydrogen tank that reduced the necessary gas volume inside the reactor. The most important parameters that were chosen to be examined were the active metals, the support materials and the pre-treatment conditions. Within the preliminary tests, the importance of an inserted protecting vessel was detected and hence used in all following reactions, probably protecting the catalyst against iron ions from the steel wall of the reactor. The support material was found to strongly influence the conversion of glycerol whereas the selectivity remained rather unchanged. The best results could be obtained with Ir-Re catalysts based on SiO2-containing materials like zeolites (H-ZSM-5, H-BEA and MCM-41) or silica (G-6 and Q-6), with the availability of free Si-OH groups on the surface being a very important factor for the catalyst performance, probably due to glycerol adsorption on these sites. Other support materials like alumina or carbon supports proved to be less active in the hydrogenolysis. Besides, no correlation between acidic sites, BET surface area or CO uptake and catalyst performance could be found. Based on these results, SiO2 (G-6) and H-ZSM-5 (80) were chosen as the main support materials for this study. Regarding the active metals, the good performance of bimetallic iridium-rhenium catalysts could be confirmed. The presence of rhenium, which had to be in direct contact with the noble metal, led to a massive increase in the reaction rate at which the selectivity did not undergo big changes. The impregnation order was also found to be crucial, iridium had to be impregnated first, followed by rhenium without an intermediate reduction step. However, this was only valid for iridium and rhenium, modifications of the impregnation order of iridium and an eventual tertiary metal did not show this effect. Rhenium on its own turned out to be almost inactive and did not form any 1,3-propanediol whereas Iridium on its own formed a similar product distribution as the bimetallic catalyst, but needed a lot more time for it. Replacing iridium by platinum in bimetallic catalysts with rhenium did not lead to great changes, mainly a slight decrease in glycerol conversion. In contrast to that, ruthenium and rhodium increased the conversion, but formed less 1,3-PDO, resulting in a much lower yield than in case of Ir-Re catalysts. The pre-treatment was thoroughly investigated, showing that the calcination in air at high temperatures of 773 K or more, which has been used in most works described in the literature, actually inhibited the performance of the catalyst, probably due to an agglomeration of the metal particles and maybe also due to remaining oxygenised species. A simple reduction at 503 K in flowing hydrogen after the impregnation and drying was found to promote the reaction. In addition, in a later stage of the work, an additional reduction in-situ in water at 473 K and 7 MPa hydrogen pressure increased the conversion and also the selectivity to 1,3-propanediol, allowing a reduction of the temperature to 393 K at a reaction time of 20 hours, reaching a yield of 21% at 43% conversion, alike literature results of a similar reaction system reported by a group independently from the Tomishige group.[70] It could be shown that the performance improvement was neither caused by a simple re-reduction after a possible oxidation of the metals during the transfer into the reaction vessel with air contact nor by the presence of dissolved rhenium originating from remaining unreduced precursor. The presence of water as well as the presence of hydrogen was found to be decisive for the success of the in-situ reduction pre-treatment. The change in selectivity indicated a change in the catalyst structure which might be caused by the high hydrogen pressure during the reduction, compared to the reduction ex-situ. In a small series of experiments with different solvents, organic solvents with hydroxyl groups like ethanol and 1,2-butanediol reduced the conversion of glycerol, probably due to adsorption of the solvent on active sites. 1,2-butanediol was slowly converted by the catalyst whereas ethanol, with only one hydroxyl group, was hardly converted, similar to 1- and 2-propanol. A reaction in pure glycerol led to amounts of products formed that were comparable to the reaction of a 20%wt. aqueous glycerol solution, indicating that the catalyst is saturated and the availability of glycerol is not a limiting factor. The importance of the presence of water could not be examined due to the water content of the solvents, precluding the possibility of working under water-free conditions even at the very beginning of the reaction. In accordance with literature results, 1,2-propanediol was converted more or less with the same rate as glycerol whereas 1,3-propanediol reacted much slower. 1,2-propanediol formed about three times more 1-propanol than 2-propanol, indicating that the primary hydroxyl group was more likely to adsorb and the neighbouring hydroxyl group underwent the hydrogenolysis. Experiments with deuterium instead of hydrogen gave an insight into the mechanism of the hydrogenolysis over an Ir-Re/SiO2 (G-6) catalyst. It could be shown that the catalyst actually exchanges hydrogen (or deuterium, in this case) atoms between the hydrogen gas phase and water. Another finding was that glycerol and hydrogenolysis products undergo a rapid dehydration and rehydration on the catalyst which led to highly deuterated species of reactant and products. The dehydration activity had been observed in other experiments, once in the absence of hydrogen and once measuring the adsorption of glycerol in DRIFTS, but the deuterium experiments led to a very clear result and also showed the dimension of the dehydration and rehydration reaction, leaving virtually no molecule untouched. It could therefore be concluded that the reaction occurs via an acid-catalysed mechanism. A recent work by Falcone et al., using mainly heavy loaded Pt-Re/SiO2 catalysts, that has been performed in parallel to this one came to the same conclusion about the mechanism, though based on different tests and arguments, hence complementing the present work.[91] This group was also not able to reproduce the results of the Tomishige group. Concluding, the main achievement of this work is to show that the Ir-Re catalyst, different from what was proposed so far, actually works as a bifunctional dehydration-hydrogenation catalyst. The results hence combine with the works published for Rh-Re and Pt-Re catalysts by the groups around Prof. Dumesic and Prof. Davis, based on DFT calculations and the aforementioned study, but contradict the proposals of the group around Prof. Tomishige, who has published most articles on Ir-Re catalysts for this reaction so far. A detailed mechanism has been proposed, according to the results of the present study. The influence and importance of several parameters, especially regarding pre-treatment, have been investigated and described in order to simplify the work of following researchers by pointing out several details like the protecting inserted vessel (glass or Teflon), in-situ reduction and abstinence of calcination that cause a big effect on the performance of the catalyst. In terms of reaction outcome, the results are very similar to the ones in a paper published by Deng and Scott,[70] however, conversion and selectivity reported by the Tomishige group could not be reproduced, at least not with calcined catalysts as described in the literature. Not calcined catalysts led to a similar selectivity, but at lower conversion. The best yield of 1,3-propanediol registered was 21% at 43% conversion at 393 K and 5 MPa H2 after 20 hours reaction time. Typical space-time yields were 2.5 mmol1,3-PDO/gcat·h at 393 K and 4.5 mmol1,3-PDO/gcat·h at 433 K, equalling around 31 and 56 mmol, respectively, of 1,3-propanediol formed per hour and per gram of metal (counting iridium and rhenium). These space time yields are in the range of what has been claimed to be the highest one for platinum-tungsten catalysts[78] and also similar to the ones calculated from publications of other works using Ir-Re catalysts. Comparing these results with the literature of the biotechnological process, in which space-time-yields are calculated in g h-1 L-1, the results of this study equal around 2 g h-1 L-1 at 393 K and 3.5 g h-1 L-1 at 433 K under the used reaction conditions and are therefore in the typical range of literature results for bio-tech processes (which stayed usually below 5 g h-1 L-1). However, in case of the heterogeneously catalysed process, the space-time-yield relating to the volume (g h-1 L-1) can be increased by simply adding more catalyst – in contrast to the biological process which is limited by relatively low maximum substrate concentrations resulting in maximum end concentrations of 1,3-PDO of less than 100 g/L. Furthermore, the biological process has several other disadvantages like very long reaction times and the more complex separation of the desired product from the reaction solution.[100] High space-time-yields also usually implicate a low yield of 1,3-PDO like 0.30 mol/mol in the example with the highest reported space-time-yield of 16.4 g h-1 L-1.[101] In comparison to the formation of 1,2-propanediol, the space-time-yield reached for the formation of 1,3-propanediol remain clearly smaller. The highest yield of 1,2-propanediol reached in our group and serving as a benchmark for upcoming studies was 22.1 g1,2-PDO / (gCu·h) at 493 K, which equals 290 mmol1,2-PDO / (gCu·h) or 70 mmol1,2-PDO / (gCat·h). Even though the higher temperature of the process to 1,2-PDO leads to higher values, it seems unlikely that comparable yields could be obtained for 1,3-PDO at higher temperatures with the known catalytic systems. It is also questionable whether catalysts aiming to 1,3-PDO and based on a dehydration-hydrogenation mechanism will ever be able to outperform the catalysts designed for the formation of 1,2-propanediol, due to the unfavourable stability of the intermediate 3-hydroxypropanal, compared to the far more stable acetol. In summary, with the elucidation of the mechanism and several important parameters, this work is another piece in the puzzle of the “perfect” catalyst for the hydrogenolysis of glycerol as part of the restructuration of the raw material base of modern industrial chemistry, away from fossil resources and to the point of an industry using renewable resources only. |
||||
Alternative Abstract: |
|
||||
URN: | urn:nbn:de:tuda-tuprints-69140 | ||||
Classification DDC: | 500 Science and mathematics > 540 Chemistry | ||||
Divisions: | 07 Department of Chemistry 07 Department of Chemistry > Ernst-Berl-Institut > Fachgebiet Technische Chemie 07 Department of Chemistry > Ernst-Berl-Institut > Fachgebiet Technische Chemie > Technische Chemie II |
||||
Date Deposited: | 26 Oct 2017 06:44 | ||||
Last Modified: | 09 Jul 2020 01:54 | ||||
URI: | https://tuprints.ulb.tu-darmstadt.de/id/eprint/6914 | ||||
PPN: | 419464638 | ||||
Export: |
View Item |