Fraunhofer-Gesellschaft

Publica

Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

New Routes for Efficient and Sustainable Oxymethylene Ethers Synthesis

 
: Salem, Mohamed Kamal Ouda
: Klein, Harald; Kraume, Matthias; Schaadt, Achim

:
Volltext urn:nbn:de:bvb:91-diss-20200108-1507846-1-1 (11 MByte PDF)
MD5 Fingerprint: 2e6334ba78fc493ed69de83253f5497d
Erstellt am: 14.3.2020


München, 2020, XII, 169 S.
München, TU, Diss., 2020
Englisch
Dissertation, Elektronische Publikation
Fraunhofer ISE ()

Abstract
The C1-oxygenates oxymethylene dimethyl ethers (OME) are attracting interest as potential diesel substitutes as they offer significant emission reductions upon combustion (e.g. particulate matter, NOx etc.), as well as being potentially attractive low vapour pressure “green” solvents (e.g. in CO2 capture applications). They are synthesized based on methanol (MeOH), which can be source from fossil (e.g. via steam methane reforming) or renewable (e.g. CO2 hydrogenation) feed stocks. In this thesis a potentially scalable and feasible OME synthesis process (denoted heron as the “direct OME synthesis”) is described based on the endothermic catalytic dehydrogenation of MeOH to yield anhydrous Formaldehyde (FA) and H2, followed by a low temperature acid catalyzed OME synthesis step. Anhydrous FA is directly introduced to the OME synthesis reactor which leads to very high single path target product OME3-5 yield of14% [g g-1product]. The process flow sheet was implemented via a hybrid simulation platform combining the merits of using “in-house” developed reactor models together with commercially available software and algorithms. The combined chemical and phase equilibrium of the OME synthesis reaction was described by solving the equilibrium relations using Newton-Raphson approach. Experiments in batch autoclave using MeOH and para-formaldehyde (p-FA) feed were carried out using commercial DW50X2® and Amberlyst 36® catalysts at different temperatures and feed ratios. The results were used to validate the OME reactor model implemented in Matlab®. Parametric study results showed that very high OME3-5 yield could be achieved at FA/MeOH molar ratio 1.8-2, temperature of 60 °C and the least water content in the feed. Additionally, the results emphasized that the temperature is not significantly influencing the equilibrium reaction. Based on the previous results, the performance of the key process step of MeOH endothermic dehydrogenation to FA and H2 was defined. A continuous test setup was constructed for testing the dehydrogenation reaction. An annular counter current reactor was developed for this kinetic controlled reaction system. Tests were carried out using Na2CO3 at T = 650-700 °C, with MeOH feed concentration <10 vol.% and GHSV between 7-35 ×103 h-1. The best performance achieved was 40% MeOH conversion at >90% FA selectivity. This performance was in the vicinity of the desired results; however the catalyst selectivity towards FA was reducing after several hours on stream while the MeOH conversion was increasing. Catalyst developments for this reaction system and scaling up the innovated reactor concepts are to be further investigated. Nevertheless, the results obtained from steady state tests of this reaction were used for extraction of experimental reaction rate constants at 690 °C and implement a simple global kinetic model of this reaction system. OME product yield is limited by the chemical equilibrium and recycle of non-reactants and OME fractions other than the desired OME product is essential for achieving satisfying process yield. Modelling this complex reaction equilibrium system involving 32components participating in 29 simultaneous reactions while considering recycling is cumbersome. The OME equilibrium composition considering recycle was then described by adopting non-stoichiometric Gibbs unconstrained minimization technique using stochastic global optimizer. This model was validated against batch experimental dataset using trioxane (TRI) and methylal (OME1) as the feed over Amberlyst 36® catalyst. The model results enabled defining the optimum FA/MeOH feed molar ratio of 1.8 which results in the desired final OME3-5 product distribution at a very high yield of 80.3 wt.%(MeOH to OME3-5) and minimum recycle ratio of ca. 2.8 [molrecycle mol-1feed]. The OME reaction equilibrium model results together with the FA global kinetic reactor model results were both implemented using Matlab® software and integrated in a CHEMCAD® platform were all the process flow sheet components are included. The integration was done by including a user-defined Excel® unit in CHEMCAD® which interfacing the two software via a VBA node. A physical property model was implemented in CHEMCAD® and the rigorous separation columns algorithms were used for separating the OME product mixture. The process heat integration was done using PinCH 2.0 software. A VBA code for stream data extraction, process streams segmentation according to phase change and heat capacities evaluation was developed for data preparation for PinCH software. Finally, a simple production cost model was implemented and the process key performance indicators were defined. After the process heat integration, the steam consumption was reduced by 16.1% and cooling water consumption was reduced by 30.4%.At annual production capacity of 35 kt OME3-5, the overall process yield MeOH toOME3-5 is 80.3%, the process energy efficiency is 71.7%, the specific steam consumption is 2.31 MWh/t OME3-5 and the production cost is 951.5 US$ per t OME3-5(0.16 €ct./kWh). The MeOH cost is the major production cost factor with 47% of the production cost share followed by the energy cost of 22.13% and then the production capacity. At large annual production capacity of 1000 kt and MeOH feed cost of300 US$ per t, the assessed production cost using the process described in this work is US$ 598.7 per t OME3-5 (0.10 €ct./kWh). This represents the lowest production cost in comparison with the available literature process and demonstrates the potential of the presented OME production process. The hybrid model developed in the frame of this thesis allows reliable and robust process evaluation and can be applied for different process concepts once the process components and parameters are correctly defined.

: http://publica.fraunhofer.de/dokumente/N-581283.html