(711g) Green Solvents for the Hydroformylation of Long Chain Alkenes

A major trend in the chemical industry is to develop more sustainable products and processes. This is a two-fold process: on one side, the sustainability of existing processes is being improved and on the other, new processes utilizing renewable feedstocks are being developed. For the latter, homogeneous metal transition catalysis (HTMC) is a promising avenue for converting renewable sources into valuable intermediates and end-products. Homogeneous catalysis has many favorable properties including lower reaction temperatures and pressures, which is more suitable for use with renewable feedstocks that require milder reaction conditions than do conventional sources. One disadvantage of using HTMC is the challenging separation and recycle of the dissolved catalyst after the reaction. The catalyst typically consists of a transition metal and an organic ligand, which are frequently highly valuable components, and must be recycled for a sustainable and economic feasible process [1]. As is often the case, large amounts of solvents are used in HTMC for both facilitating the reaction and performing the post-reaction catalyst recovery. In the frame of sustainability, these solvents should also be environmentally benign, preferably having acceptable environmental, health and safety (EHS) criteria.

The reaction and separation can be performed using a novel type of solvent system, a thermomorphic multicomponent system (TMS). In this solvent mixture, the reaction is conducted at a higher reaction temperature in a homogeneous liquid phase, while the separation of catalyst and product takes place using liquid extraction at a lower, separation temperature. This temperature-dependent phase behavior is achieved by using two solvents with a miscibility gap whose critical point lies between the reaction and the separation temperature [2]. The TMS concept addresses the limitation of the widely used Ruhrchemie/Rhône-Poulenc Process for the hydroformylation of short-chain alkenes [1]. Here, the reactant alkenes are dissolved in an aqueous solution and the product aldehydes form a second liquid phase that can be siphoned off. The water-soluble catalyst is almost insoluble in the aldehyde phase, minimizing the catalyst loss. The concept, which is very efficient for short alkenes, cannot be applied for long chain alkenes since the solubility of the reactant in the aqueous catalyst phase is poor, resulting in an inefficient, diffusion-limited reaction. In a TMS, however, the reaction can take place in single liquid phase regime without mass transfer limitations.

The TMS principle has already been successfully applied to the hydroformylation of 1-decene or 1-dodecene using a solvent system comprised of dimethylformamide (DMF) and decane [3, 4]. This system, which is highly efficient from a thermodynamic point of view, suffers from the use of the developmentally toxic solvent DMF. Because of its EHS characteristics, DMF is included on the list of substances of very high concern according to the REACH legislation (Regulation, Evaluation and Authorization of Chemicals) [5]. In order to find a replacement solvent for DMF, a computer-aided TMS selection strategy has recently been developed [6]. Promising TMS components are selected by predicting thermodynamic properties using COSMO-RS and EHS properties using the QSAR software VEGA [7, 8]. Some characteristics considered in this selection procedure are the reactant, product and gas solubilities in the solvent candidates, the boiling points of the solvents, the LLE behavior and 14 different EHS properties (predicted by 33 QSAR models) such as toxicity, persistence, flash point and others.

In this contribution, the screening methodology is briefly introduced, and focus is placed on the experimental validation of the theoretical results. The evaluation of the green solvent candidates is divided into the following steps. Firstly, the boiling point should be checked at ambient conditions. Although, accurate boiling point predictions are available, checking the boiling point first is reasonable because the experiment dependents on one substance and has a low expenditure. Additionally, this step also allows one to evaluate the thermal stability of the solvent, a key element for the process performance. Secondly, the temperature dependency of the liquid-liquid equilibrium (LLE) of the binary solvent mixture is investigated by checking whether the critical point lies within the desired temperature range. This can be done quickly by cloud point experiments and should be the next step in the experimental investigation because of its simplicity and the large reduction of uncertainty in actual LLE behavior, since theoretical predictions thereof are still prone to errors. However, in order to ensure TMS functionality, the liquid phase behavior is also evaluated using a mixture with both the reactant and the product. Therefore, the phase behavior of a mixture of the reactant and the solvents is checked at the reaction temperature, and an analogous procedure is conducted for a mixture comprised of the product and the solvents at the separation temperature. These compositions represent the beginning of the reaction in a batch process, or respectively, for the subsequent separation under the assumption of full conversion. Afterwards, the LLE of the product and the solvents under separation is measured in detail under separation conditions. With these measurements one can then check the feasibility of the product separation. Next, the performance of the reaction within the new solvents is evaluated in terms of solvent stability and adequate reaction rates. Semi-batch experiments are conducted through which samples are taken out of the reactor to evaluate the reaction progress along the reaction time and to check the general functionality. Finally, and perhaps most importantly, the catalyst partitioning between the two liquid phases needs to be investigated, which is the key for an efficient catalyst recycling. This is chosen as the last step because for these measurements an ICP-MS is used and performing measurements in organic solvents using an ICP-MS is cumbersome. However, the use of an ICP-MS is necessary because the leaching of the catalyst into the product phase is extremely low (parts per billion down to parts per trillion), which is not possible with most analytical devices. Here, the best TMS are those with the lowest amounts of catalyst leaching.

Altogether, this experimental approach allows for the comparison of different solvents regarding their applicability as a TMS used in a hydroformylation process and shows the potential for green solvents to replace DMF. The collected data also makes it possible to perform a more rigorous process optimization in subsequent investigations, where decisions about solvent composition can be included. In this manner, a more ecologically beneficial process may be realized, facilitating the development of more sustainable processes for the chemical industry.

Acknowledgment

This work is funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) - TRR 63 „Integrated Chemical Processes in Liquid Multiphase Systems“ (Gefördert durch die Deutsche Forschungsgemeinschaft (DFG) - TRR 63 "Integrierte chemische Prozesse in flüssigen Mehrphasensystemen" (Teilprojekt B9) - 56091768.). Steffen Linke is also affiliated with the “International Max Planck Research School for Advanced Methods in Process and Systems Engineering - IMPRS ProEng” at the Max Planck Institute for Dynamics of Complex Technical Systems Magdeburg.

References

[1] Arno Behr and Peter Neubert. Applied homogeneous catalysis. Wiley-VCH-Verl., Weinheim, 2012.

[2] Arno Behr, Guido Henze, Dietmar Obst, and Barbara Turkowski. Selection process of new solvents in temperature-dependent multi-component solvent systems and its application in isomerising hydroformylation. Green Chemistry, 7(9):645, 2005.

[3] Kevin McBride, Nicolas Maximilian Kaiser, and Kai Sundmacher. Integrated reaction-extraction process for the hydroformylation of long-chain alkenes with a homogeneous catalyst. Computers & Chemical Engineering, 105:212-223, 2017.

[4] Benjamin Hentschel, Gregor Kiedorf, Martin Gerlach, Christof Hamel, Andreas Seidel-Morgenstern, Hannsjörg Freund, and Kai Sundmacher. Model-Based Identification and Experimental Validation of the Optimal Reaction Route for the Hydroformylation of 1-Dodecene. Industrial & Engineering Chemistry Research, 54(6):1755-1765, 2015.

[5] European Parliament. Registration, Evaluation, Authorisation and Restriction of Chemicals: REACH. https://echa.europa.eu/documents/10162/3f575bea-5a96-4f18- a110-d294d37422c7, 2006.

[6] Steffen Linke, Kevin McBride, and Kai Sundmacher. Systematic green solvent selection for the hydroformylation of long chained alkenes. ACS Sustainable Chemistry & Engineering, UNDER REVIEW, 2020.

[7] Andreas Klamt. Conductor-like Screening Model for Real Solvents: A New Approach to the Quantitative Calculation of Solvation Phenomena. The Journal of Physical Chemistry, 99(7):2224-2235, 1995.

[8] Emilio Benfenati, Alberto Manganaro, and Giuseppina Gini. VEGA-QSAR: AI inside a platform for predictive toxicology. Italy Published on CEUR Workshop Proceedings, (Vol-1107), 2013.