Design of the reversible biphasic arrangement in the microfluidic chip reactor for asymmetric hydrogenation reactions

Highlights

• Temperature driven reversible biphasic solvent system was designed/optimised.

• It was utilised for steroselective hydrogenation of MAA to (R)-MHB.

• The reaction was catalysed by optically pure (R)-Ru-BINAP complex.

• The reaction was performed in a microfluidic chip reactor.

• The biphasic system utilised a mixture [N_14,222][Tf₂N]/MeOH/H₂O.

Abstract

Design of the reversible biphasic system for a practical use in asymmetric hydrogenation performed in a microfluidic chip reactor is reported. Methylacetoacetate (MAA) was transformed to (R)-methylhydroxybutyrate over optically pure (R)-Ru-BINAP as a model reaction. The study was an iteration towards the design, description and optimization of the temperature driven reversible biphasic system in the mixed [N_R,222][Tf₂N]/methanol/water phase by varying the parameter of the structure of the ionic molecule, starting from [NH₄][Tf₂N] up to [N_14,222][Tf₂N]. At temperatures, and other conditions providing high conversions and optical yields, the reaction mixture was monophasic. At lower temperatures, the mixture became biphasic for ionic liquids with long alkyl chains (namely [N_14,222][Tf₂N]) due to their strong non-polar character. The formed ionic liquid phase accommodated the chiral Ru complex, the water/methanol phase the reaction products. After the reaction, when the mixture was cooled, over 90% of the catalytic complex was kept in the ionic liquid phase. Viscosity and density data were also discussed. It seems the reversible biphasic system offers a way to facilitate the separation of the chiral Ru-complex from the reaction mixture. Due to high sensitivity of the complex to handling conditions, its reuse still requires further optimization.

Introduction

Many finechemicals are industrially requested as optically pure isomers. The corresponding technologies are usually operated as batch processes. Continuous tubular reactors are rather exceptional. The fine-chemical industry still mostly relies on the batch or semi-batch reactor infrastructure. The continuous regime, however, may bring many advantages (Gutmann et al., 2015; Kluson et al., 2016a, 2019). First, the scaling-up is usually more feasible for the continuous process. The flow routes developed and optimized in a laboratory can be usually scaled to production amounts with minimal re-optimization and without major changes in the synthetic path. In the case of high value added chemicals it is important to design the chemical engineering part and the process part together with the optimal performance of the chemical reaction. In this respect microfluidic chip reactors represent a fully functional practical platform for the design and optimisation of reaction conditions under continuous flow regime. Because of its high surface area to volume ratios and a small reactor volume it benefits of unique mass and energy transport capabilities. The excellent heat and mass transfer characteristics, together with the fact that the reaction is resolved along the length of the reaction channel, enables a precise control of the residence time of intermediates or products (Gutmann et al., 2015; Kluson et al., 2016a, 2019).

Asymmetric reactions (those yielding optically pure products), namely hydrogenations over optically pure noble metal complexes, have been treated extensively in the past (Knowles and Noyori, 2007; Govender et al., 2016; Blaser et al., 2001; Xie et al., 2018). For industrial purposes big effort had been spent to immobilize the catalytic complex on various types of solid matter to transform the “issue” to the heterogeneous arrangement with its simpler separation, recovery, and a possible reuse (Tito and Kluson, 2016; Kluson et al., 2016b). Despite the effort this could be (in most cases) achieved only when compromising with the optical purity of the principle product. It should be added that even in the homogeneous arrangement, when the optically pure catalyst is dissolved in the reaction mixture, this very delicate and sensitive property needs to be deliberately addressed.

In the past years some papers (Hallett and Welton, 2011; Hintermair et al., 2010; Olivier-Bourbigou et al., 2010; Picquet et al., 2004; Oechsner et al., 2009; Steinrueck and Wasserscheid, 2015; Giernoth, 2007; Floris et al., 2010) have brought evidence that room temperature ionic liquids of various molecular structures might be well suited for the accommodation of such catalytic complexes. Among others N-alkyl-triethylammonium bis(trifluoromethylsulfonyl)imides = [N_R,222][Tf₂N] were found effective (Fig. 1) (Floris et al., 2009). In the summary formula R denotes for the length of the alkyl chain, the “222” subscript abbreviates the three ethyl groups. Obviously, there must be a significant role of the length of their alkyl chain in the cationic part of the molecule (N-alkyl-triethylammonium)⁺. When elongating the non-polar part the overall alkane character of the ionic liquid molecule may finally prevail above the ionic character. It has a straight impact on many other properties such as viscosity, density, diffusivity, coordination ability, miscibility (phase behaviour), overall molecular energetics, etc. (Papovic et al., 2016; Cerna et al., 2011; Ghatee et al., 2013; Camargo et al., 2016; Rocha et al., 2016; Moosavi et al., 2016).

The complexity of the subject could be well documented when employing a thoroughly described reaction model, such as asymmetric hydrogenation of β-ketoesters (Kluson et al., 2019, 2016b; Oechsner et al., 2009; Floris et al., 2010, 2009; Kluson et al., 2017a; Floris et al., 2011), for example methylacetoacetate (MAA) to (R)-methylhydroxybutyrate ((R)-MHB) over (R)-Ru-BINAP in the mixed methanol/([N_R,222][Tf₂N] (Fig. 2). The process starts with the achiral compound (MAA), which is being transformed to the optically pure product ((R)-MHB). The chirality is provided by the homogeneous chiral bidentate phosphine (BINAP) complex of ruthenium ((R)-Ru-BINAP). It is noteworthy that this unique catalyst was winning a Nobel Prize for its inventor, Ryori Noyori, in 2001 (Knowles and Noyori, 2007).

In our initial works we treated the issue of the role of the molecular structure of various ionic liquids, including also [N_R,222][Tf₂N], on the stereoselective reaction course exclusively in the batch mode (Floris et al., 2010, 2009; Cerna et al., 2011; Floris et al., 2011; Bartek et al., 2005). In parallel some effort had been spent on the theoretical evaluation of the energetically most stable conformers of the series of the used [N_R,222][Tf₂N]. The conformers with extreme features were used for the interpretation of the experimentally observed selectivity-structure effects (Dytrych et al., 2014).

At that stage it was sought to operate the already optimised reaction (in the batch mode) in the continuous regime. We attempted to transfer the whole system to the microfluidic chip reactor. Advantages of this approach and experimental arrangement have been described elsewhere (Yao et al., 2018; Fukuyama et al., 2008; Chinnusamy et al., 2012; Munirathinam et al., 2015; Tsaoulidis et al., 2013; Kashid et al., 2011; Wilms et al., 2009). The study was first targeted on comparative measurements of the pressure drop in two distinct geometries, the microcapillary and the microfluidic reactor. In both cases the prediction was calculated via the Hagen-Poiseuille law, which relates (under laminar flow) the pressure drop with channel radius, length, liquid flowrate and viscosity. The viscosities of corresponding liquids were also determined experimentally. It revealed that the predicted and experimental pressure drop values matched very well for the microcapillary, while they did not match for the microfluidic reactor (Kluson et al., 2017b). In the microfluidic reactor, the measured pressure drop data were higher than the predicted one. The channel cross section was approached by a capillary radius for the microcapillary, and by a hydraulic radius for the microfluidic reactor. The discrepancy can be explained that the real channel cross-section geometry differs considerably from the expected elliptical shape (Kluson et al., 2016a, 2017b). The found deflections and trends were interpreted with intention to validate the microfluidic chip for practical flow conditions of the used fluids involving the series of ([N_R,222][Tf₂N]. It also included the description of their flow properties in the confined space in ternary mixtures with methanol and water.

After that three microfluidic chip reactors (different in length and volume), incorporated in the microreactor platform, were operated isothermally under steady state conditions (at temperature levels from 363 K to 433 K) in the model reaction using the [N_8,222][Tf₂N]/methanol/water phase (R = 8, the alkyl chain contains 8 carbon atoms) (Kluson et al., 2017a). In the reactions, up to the temperature level of 413 K, the conversions were still increasing (up to 98%). The achieved enantiomeric excess (ee) was over 99% towards the (R)-(+)-methyl-3-hydroxybutanoate isomer. At the same temperature the highest achieved conversion was also observed for the comparative experiment only in pure methanol. However, in this case the ee parameter reached only 94%. With help of thermodynamic calculations it was shown the solvents actively participated in the re-coordination of the catalytic complex. Besides the complex coordination (impact on ee) it also revealed that [N_8,222][Tf₂N] imposed an evident stabilisation effect on the otherwise very sensitive chiral catalyst due to energetic reasons caused by molecular dynamics of the relatively long alkyl chain (Kluson et al., 2017a; Dytrych et al., 2014). At this part it was declared that the microreactor platform was successfully adapted, and the reaction system validated for the confined and specific microchip space.

In the current communication we report on the design of the temperature driven reversible biphasic system in the microfluidic chip reactor intended for a practical use in the model asymmetric hydrogenation. The major message is the targeted use of master solvents (specific mixtures of ionic liquids) for the chemical synthesis in order to recycle an optically pure catalyst and separate it from the product and other reactants. It is a study on a molecular design of [N_R,222][Tf₂N] ionic liquids which provide the desired phase behavior: a single phase at reaction and two phases for separation (schematically shown in Fig. 3), while ensuring that the catalyst recycling is possible. Methylacetoacetate (MAA) transformation to (R)-methylhydroxybutyrate ((R)-MHB) over optically pure (R)-Ru-BINAP is again employed. It is “shaped” as the iteration towards the temperature driven reversible biphasic arrangement in the mixed [N_R,222][Tf₂N]/methanol/water phase by varying the parameter of the molecular structure of ionic molecules, starting from [NH₄][Tf₂N] up to [N_14,222][Tf₂N] (number of carbons of the alkyl chain R = 2, 4, 6, 8, 10, 12, 14). Before approaching the reaction the viscosity and density data are carefully discussed. Then trends of the phase behavior for selected ternary mixtures [N_R,222][Tf₂N]/methanol/water are introduced and discussed in the form of triangular diagrams. The reversible biphasic system is believed to facilitate the separation of the chiral Ru-complex from the reaction mixture that can then be potentially reused in the reaction.