Stereoselective synthesis of optical isomers of ethyl 4-chloro-3-hydroxybutyrate in a microfluidic chip reactor

  • Petr KlusonEmail author
  • Petr Stavarek
  • Vera Penkavova
  • Hana Vychodilova
  • Stanislav Hejda
  • Natalie Jaklova
  • Petra Curinova
Full Paper


Ethyl (R)-4-chloro-3-hydroxybutyrate ((R)-CHBE) is a versatile fine chemistry intermediate. It is used as a precursor in the synthesis of several pharmacologically valuable products, including L-carnitine. It is usually produced by means of stereoselective biotechnology methods in enzymatic reactions. An alternative preparation strategy towards ethyl (R)-4-chloro-3-hydroxybutyrate is based on the asymmetric hydrogenation of ethyl 4-chloro-acetoacetate (ECAA) to the optically pure product ((R)-CHBE) over ((S)-Ru-BINAP) catalytic complex. The reaction conditions were optimised first using (R)-Ru-BINAP yielding the (S)-CHBE isomer. All reactions were performed under continuous regime in a microfluidic chip reactor. Three different solvent phases were employed. The methanol/water phase, the ethanol/water phase, and the [N8,222][Tf2N]/methanol/water phase. The attained conversions were total in all cases already at 408 K. The parameter of enantioselectivity ee was 99.4% towards the (S)-CHBE for the system in which (R)-Ru-BINAP was accommodated in [N8,222][Tf2N]/methanol/water phase. In the case of methanol/water experiment the ee parameter reached 92.5%. For ethanol/water ee was 91.8%. The (R)-CHBE isomer over (S)-Ru-BINAP was obtained with ee = 99.3% in the [N8,222][Tf2N]/methanol/water phase at 408 K. For the reactions leading to (S)-CHBE apparent activation energies were evaluated. They were similar for MeOH/water and EtOH/water (110.5 and 110.7 kJ.mol−1). The apparent activation energies corresponding with the [N8,222][Tf2N]/MeOH/water system were much higher (of about 90 kJ.mol−1) reaching the level of 200 kJ.mol−1. The impact of the molecular structure of the main reactant was negligible as appeared from the comparison with hydrogenation of methylacetoacetate (MAA, ~ 200 kJ.mol−1). The effect of the presence of the [N8,222][Tf2N] ionic liquid on the sum of the activation energy dominated. The effectiveness of the enantioselective synthesis was additionally assessed by nuclear magnetic resonance employing the method of enantioselective complexation of the chiral compound with a chiral solvating agent.


L-carnitine Optical purity Stereoselective hydrogenation Ru complex Microfluidic chip reactor 


Carnitine (3-hydroxy-4-N,N,N-trimethylaminobutyrate) is an indispensable substance for energy metabolism, especially in tissues dependent on β-oxidation of fatty acids [1]. The main physiologic function of carnitine is the transport of activated long-chain fatty acids across the inner mitochondrial membrane and the reversed transport of any toxic acyl-groups originated in mitochondria. It also improves metabolism of glucose via stimulation of pyruvate dehydrogenase complex [1, 2]. It exists in two optical isomers, D- and L-. However, in higher living species it appears only in the L- form [1]. Carnitine intake is provided mainly by nutrition, the largest portion is available in meat. The organism is also able to supply its needs by an endogenous synthesis in liver, kidneys and brain from amino acids lysine and methionine [1, 2, 3]. Some patients with medical and/or genetic disorders are unable to receive it through metabolic pathways. In such cases it must be artificially added as a nutrient or as part of specially administered drugs. The genetically induced (primary) carnitine deficiency is a disorder with symptoms of cardiomyopathy, skeletal-muscle weakness, and hypoglycaemia. The secondary carnitine deficiencies may result of certain maladies such as chronic renal failure, or metabolic syndrome, and after long-term antibiotics therapies, due to significant malnutrition, etc. [4, 5, 6, 7, 8, 9, 10, 11, 12].

Ethyl (R)-4-chloro-3-hydroxybutyrate ((R)-CHBE) is referred to as a versatile precursor (intermediate) for several pharmacologically valuable products, including also L-carnitine [1, 13, 14, 15]. Despite several synthetic strategies have been developed, the enzymatic asymmetric synthesis is still the most common way. In this respect stereoselective carbonyl reductases (SCR, the nicotinamide cofactor-dependent enzymes) are tremendously important. They are universally known to catalyse reversible redox reactions between alcohols and aldehydes/ketones [16, 17, 18, 19, 20, 21]. During the past decade they have been considerably applied to the synthesis of chiral pharmaceutical intermediates for anti-cholesterol drugs, β-lactams antibiotics, anticancer drugs, and many others. On the other hand scale-up of the SCR-catalysed reactions might be complicated by their narrow substrate specificity, expensive cofactor dependency, substrate insolubility, low stability, etc. [16, 17, 18, 19, 20, 21].

Most of the L-carnitine producers are concentrated in China. The main market players are Lonza Group, Northeast Pharmaceutical, Hengtai Chemical and Chengda Pharmaceutical. The global production of L-carnitine increased from 6200 metric tons in 2011 to 9400 metric tons in 2015. It is available in three grades on the market: human feed grade, animal feed grade, and pharmaceutical grade. The market segment by application can be divided into animal food, health care products, functional drinks, medicine and others (

Here we report on an alternative strategy towards ethyl (R)-4-chloro-3-hydroxybutyrate and ethyl (S)-4-chloro-3-hydroxybutyrate starting with a common chemical, ethyl 4-chloro-acetoacetate (ECAA). This molecule is subjected to stereoselective hydrogenation (Fig. 1) catalysed by homogeneous chiral bidentate phosphine (BINAP) complex of ruthenium ((R)-Ru-BINAP, (S)-Ru-BINAP) [22, 23, 24, 25]. The process starts with the achiral compound (ECAA), which is being transformed to the optically pure product ((R)-CHBE) or ((S)-CHBE). The chirality is provided either by ((R)-Ru-BINAP) (Fig. 2a) or by ((S)-Ru-BINAP) [22, 23, 24].
Fig. 1

A simplified reaction scheme showing the hydrogenation of ECAA to two optical isomers of CHBE over chiral Ru-BINAP complex

Fig. 2

a(R)-[Ru(BINAP)Cl(p-cymene)]Cl = (R)-Ru-(BINAP), b [N8,222][Tf2N].

Several strategies have been described to preserve the optical purity and activity of the chiral organometallic complex in the stereoselective reaction [22, 23, 24, 25, 26]. One of them involves the use of room temperature ionic liquids as the “accommodation” medium for the sensitive noble metal complex [27, 28, 29, 30, 31, 32]. For the specific use of Ru-BINAP the ionic liquid N-octyl-triethylammonium bis(trifluoromethylsulfonyl)imide = [N8,222][Tf2N] (Fig. 2b) was identified recently as quite effective [27, 28, 33, 34, 35].

When considering a production of (R)-CHBE instead of its mere laboratory synthesis, the continuous regime should be preferred despite the fine-chemical industry dominantly relies on the existing batch or semibatch reactor infrastructure [36]. Scaling-up is usually more feasible for the continuous process. The flow routes developed and optimized in a laboratory can be scaled to production amounts with minimal re-optimization and without major changes in the synthetic path [36]. For such purposes some types of microreactors are quite useful [33, 34, 36, 37]. Namely, the microfluidic chip reactor represents 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 small reactor volume it benefits of unique mass and energy transport capabilities [36, 37]. The excellent heat and mass transfer characteristics, together with the fact that the reaction is resolved along the length of the reaction channel, enables precise control of the residence time of intermediates or products [36]. Safety aspects are also important reasons for carrying out a chemical reaction in a microreactor [38].

In this communication we report on the stereoselective hydrogenation of ECAA to (R)-CHBE or (S)-CHBE over (S)-Ru-BINAP or (R)-Ru-BINAP accommodated in [N8,222][Tf2N]/methanol/water (alternatively methanol/water or ethanol/water) phase, and performed in the microfluidic chip reactor (Fig. 3). The main intention of the study was to verify the feasibility of the reaction and process system for the possible technology scaling-up in the production of ethyl (R)-4-chloro-3-hydroxybutyrate pharmaceutical intermediate. To the best of our knowledge a similar treatment has not been published yet.
Fig. 3

The Chemtrix microfluidic chip reactor of the 3223 type (10 μl). Liquid inlet (1), gas inlet (2), quench inlet (3) (unemployed in this study), outlet (4), static mixers (5), meandering channel – reaction space (6), channel width W 300 μm, height H 120 μm

Materials and methods

Methanol, ethanol and water were of HPLC grade and they were supplied by Merck. [N8,222][Tf2N] was of IoLiTec origin with the declared >98.5% purity. ECAA, (R)-Ru-BINAP, (S)-Ru-BINAP, and both CHBE isomers were purchased from Aldrich. Optical purity of the Ru complexes was 99.5%.

Three microfluidic chip reactors (3222 = 5 μl, 3223 = 10 μl, 3227 = 19.5 μl; different in length and volume), were operated isothermally under hydrogen pressure of 20 bar and steady state conditions at temperature levels from 363 K to 433 K in the stereoselective reaction using the [N8,222][Tf2N]/methanol/water, methanol/water, or ethanol/water mixtures as the solvent phase.

The experimental setup was based on the Labtrix (Chemtrix) microreactor platform [33, 34] accommodating the glass microfluidic chip reactor (Fig. 3). All employed chips were of the same following parameters: channel width W 300 μm, height H 120 μm, channel cross-section A 32910 μm2, channel circumference P 737 μm, hydraulic equivalent radius of the channel RH = 2A/P = 89.3 μm. By varying the length of the reactor’s channel and hence the overall volume, the parameter of the residence time t [s] of the liquid mixture in the reactor can be evaluated for a wide range of flow rates as a ratio of the reactor volume V [μl] and the volumetric flowrate of gas and liquid Q(L + G) [μl.s−1] t = V/Q(L + G).

The liquid was supplied by a linear pump (Fusion 200, Chemyx Inc.), and it flowed through temperature (Pt100) and pressure (M 11, WIKA) sensors to the first inlet of the microfluidic chip. The second inlet of the microfluidic chip was connected to the hydrogen line (99.9%, Linde); the hydrogen flowrate was controlled by a mass flow controller (Bronkhorst). The molar composition of the reaction mixtures involved [N8,222][Tf2N] in amount of 9%, water in 19%, methanol in 60%, (S)- or (R)-Ru-BINAP in 0.05%, and ECAA in 12%. As alternatives, methanol/water and ethanol/water solvent phases were also used (see Table 1). For all reaction mixtures the ECAA concentration (cECAA = 1.53 [mol l−1]) and molar ratio of (S)- or (R)-Ru-BINAP/ECAA were kept constant (0.004 [−]). The catalyst concentration was cRucat = 5.6 g.l−1. Experiments with (S)-Ru-BINAP were performed only in the [N8,222][Tf2N]/methanol/water phase.
Table 1

Compositions of the tested mixtures

Mixture component







M, [g/mol]







Mixture composition, [mol. %]


















The hydrogen flowrates were in about 2x stoichiometric molar excess with respect to ECAA. It corresponded to the hydrogen flowrates 600 μl/min at the liquid flowrate 10 μl/min. The hydrogenation experiments were conducted under continuous dosing of the liquid and the gaseous phases, with a single pass of the mixture through the microchip reactor. The reaction was quenched by cooling the mixture that was leaving the reaction space to 278 K.

The reaction samples were analysed on a gas chromatograph Trace GC 1310 (Termo Scientific GC) equipped with an autosampler TRIPLUS 100 and the FID detector. The used column was RTX 50 (Restek) containing 50% of phenylmethylsiloxane (60 m × 0.25 mm × 0,25 μm). The data were verified with the enantioselective column Rt-βDEXsm (30 m × 0.32 mm × 0.25 mm) by Restek. Diethylether was the analysis solvent, with the addition of 2 wt.% of 4-methylanisole as the internal standard.

Enantiomeric ratio in the obtained product was verified by 1H NMR spectroscopy using a chiral solvating agent (CSA), the (Sa)-2,2’-bis[N’-3,5-bis(trifluoromethyl)phenyl]ureido-1,1’-binaphthalene. The proton NMR signals of individual enantiomers cannot be distinguished by NMR, but the signals of diastereomers can be differentiated. The CSA is capable of formation of diastereomeric complexes with enantiomers of the product. The complexation leads to separation of signals belonging to complexes with individual enantiomers and their ratio is obtained then by integration of these signals. The NMR spectra were recorded on the Agilent 400MR DDR2 spectrometer (1H 400 MHz, 13C 100 MHz). The chemical shifts are expressed in parts per million and are referenced to the residual signal of the solvent used. More details to the method are given elsewhere [39].

Results and discussion

Course of the stereoselective hydrogenation of ECAA to (S)-CHBE over (R)-Ru-BINAP carried out in the continuous regime in the microfludic chip reactor was assessed first. Three different solvent phases were employed. The first contained methanol/water. For the second solvent phase, methanol was substituted with ethanol to inspect a possible re-esterification of ethyl 4-chloro-acetoacetate with methanol under reaction conditions in the previous case. The third solvent phase comprised, besides water, [N8,222][Tf2N] ionic liquid and methanol. The protective role of the ionic liquid provided to the chiral Ru complex was sought. As already shown [40], presence of water may restrict the formation of acetals arising in the side reaction. The acid-catalysed nucleophilic addition of alcohol on >C=O is a reversible process and water may shift the equilibrium towards the carbonyl compound. Water molecules also contribute to the polar character of the solvent phase, which is important mainly for the phase with [N8,222][Tf2N] bearing a long C8 alkane chain. Viscosities, densities and phase behaviour of various ternary mixtures of [N8,222][Tf2N]/methanol/water were reported recently [34].

Conversions of ECAA (XR) vs. reaction temperatures are plotted in Fig. 4. Each of the isothermal levels represents a separate experiment. Under studied conditions the overall reaction rate was apparently insufficient to indicate any reaction progress below 363 K. At low conversions certain indication of a pseudo induction period appears. (The induction period is usually referred to for plots in which time is an independent variable.) It is more pronounced for the [N8,222][Tf2N]/methanol/water system (same for the appearance of the inflection point). It is followed by a steep increase nearly up to the temperature level of the total conversion. The ascendant part sharply changes the trend at ~ 408 K. Then a flat region is established. No contribution of the reversible reaction (dehydrogenation) was observed for all the three reaction systems at temperatures over 408 K.
Fig. 4

ECAA conversion as a function of temperature for the three used solvent systems and (R)-Ru-BINAP; 3223 microchip, QL = 10 μl.min1, QG = 600 μl.min1, 20 bar

Certain differences in the conversion-temperature plots could be ascribed to different partial mechanisms contributing to the overall reaction. The involvement of the (R)-Ru-BINAP complex brings a series of consecutive/parallel steps, each requiring a certain amount of activation energy. As already shown by Wolfson et al. [41] they are closely linked with the coordination/re-coordination of a reactant to the active noble metal core of the organometallic complex. The situation is not by far trivial when [N8,222][Tf2N] is involved. As demonstrated by Dytrych et al. [35] the stabilisation effect of the ionic liquid molecule onto the catalytic complex is thermodynamically driven.

The experimental data (Fig. 4) were utilised for evaluating the apparent activation energies (EAap) for each of the reaction systems. It was based on the assumption of the first order kinetics. One of the corresponding Arrhenius plots is depicted in Fig. 5 (here for [N8,222][Tf2N]/MeOH/H2O). Values of EAap are listed in Table 2 together with the EAap for stereoselective hydrogenation of methylacetoacetate (MAA) obtained recently [33] under same experimental conditions, also in [N8,222][Tf2N]/MeOH/H2O.
Fig. 5

The Arrhenius plot for the reaction system with [N8,222][Tf2N]/MeOH/H2O for ECAA hydrogenation over (R)-Ru-BINAP

Table 2

Apparent activation energies


Solvent phase

EAap [kJ/mol]













The table reveals quite clear trends. The apparent activation energies are (within the experimental error) virtually identical for the systems MeOH/water and EtOH/water (110.5 and 110.7 kJ.mol−1). At this point it must be added that the participation of the re-esterification reaction with methanol in the case of MeOH/water system was insignificant. Also the formation of acetals upon the reaction was completely marginal. Its suppression with water was very effective. The apparent activation energies corresponding with the [N8,222][Tf2N]/MeOH/water system are much higher (of about 90 kJ.mol−1) reaching the level of 200 kJ.mol−1. It is noteworthy that the impact of the molecular structure of the main reactant is completely negligible as we may see from values for ECAA and MAA. Obviously the effect of the presence of the ionic liquid on the sum of the activation energy dominates. The coordination of [N8,222][Tf2N] with Ru is energetically “expensive”.

The plot in Fig. 6 shows the conversions of ECAA and MAA when varying the hydrodynamic residence time. This was achieved either by changing the internal volume of the microchip or by changing the liquid and gas flowrates. The data approximation obeyed the first order kinetics XR = 100(1 − ekt). In this equation XR [−] denotes for the ECAA (MAA) conversion, k [s−1] is the apparent kinetic constant, and t [s] is the reaction time. The solution of a mass balance for a batch reactor can be adopted here since the reaction time can be expressed by the hydraulic residence time calculated as a ratio of a microchip internal volume V and sum of actual gas and liquid volumetric flowrates (t = V/Q(L + G)). For the actual gas volumetric flowrate, pressure and temperature correction was necessary. The inlet gas flowrate indicated by a mass flow controller in units Nml.min−1 (i.e. at normal conditions 273 K, 101.3 Pa) was corrected according the equation of state to actual reaction pressure and temperature. Further, the amount of hydrogen consumed by the reaction was evaluated on the basis of achieved conversion and this amount was subtracted from the inlet hydrogen flowrate providing an approximation to actual gas volumetric flowrate through the microfluidic chip.
Fig. 6

The ECAA (closed points) and MAA (open points) conversions vs. V/Q(L + G) plot; ([N8,222][Tf2N]/MeOH/H2O), T = 408 K, microchip 3223 (●,○), and microchip 3222 (▲,∆). Correlation coefficients of the exponential fit are 0.9819 (ECAA) and 0.9695 (MAA)

The evaluated kinetic constant kECAA = 0.079 [s−1] has been obtained at temperature T = 408 K and for [N8,222][Tf2N]/MeOH/H2O (correlation coefficient 0.9819). In comparison with MAA, the hydrogenation of ECAA proceeded faster under the same operating and solvent conditions (kMAA = 0.055 [s−1], correlation coefficient 0.9695). It accounts for a higher reactivity of ECAA. The deviation of experimental data from the predicted curve in Fig. 6 is more pronounced for shorter residence times. It is due to that at short residence times in range of few seconds the inaccuracy of pump and gas mass flow controller could be amplified.

As repeatedly mentioned the attained conversions were total for all studied reaction systems (at 408 K). On the other hand the achieved optical purities of the product already differed. The parameter of enantioselectivity (ee) was 99.4% towards the (S)-CHBE isomer for the system in which (R)-Ru-BINAP was accommodated in [N8,222][Tf2N]/methanol/water phase. When reminding that the optical purity of the Ru complex was 99.5% this result could be seen as very acceptable. In the case of methanol/water experiment the ee parameter reached only 92.5%, and it was 91.8% for ethanol/water. Besides gas chromatography with the stereospecific column the enantiomeric ratio in the obtained products was also determined by 1H NMR spectroscopy using a chiral solvating agent [39]. The two corresponding NMR spectra are seen in Fig. 7a). By integration of the signals of enantiomer (R), which are not hindered by signals of the major enantiomer (S), the ratio of enantiomers was evaluated.
Fig. 7

A part of 1H NMR spectrum of products obtained by ECAA hydrogenation using (a) reaction with (R)-Ru-BINAP; major peaks belong to the complex of the chiral solvating agent with (S)-CHBE; (b) reaction with (S)-Ru-BINAP; major peaks belong to the complex of the chiral solvating agent (R)-CHBE; (c) superimposed spectra clarifying the assignment of residual peaks of the corresponding opposite enantiomer in both reaction mixtures

The cost of [N8,222][Tf2N] in comparison with methanol or ethanol is incomparably higher of course. Is it then worth to utilise it because of 7% of ee as it is the case here? The answer is unambiguously “yes”. It is not an exaggeration to say that on the practical level there is not already a big difference to obtain the product of the asymmetric reaction with the optical purity of 72% or 92%. The subsequent purification and separation steps are necessary in either case. Regardless the chosen method (chiral membranes, chiral preparative chromatography, re-crystallisation, etc.) they are all complicated, quite sensitive to external conditions, usually quite unstable in time regarding ee, and always expensive. On the other hand the stereoselective synthesis, namely the one relying on the optically active homogeneous catalytic complexes, focuses on the preparation of the product with the highest achievable optical purity in one reaction step. Ideally the ee should be located on the level of the “chirality provider” (here (R)-Ru-BINAP or (S)-Ru-BINAP), with no need of any further processing for improving the optical yields. Majority of the current products in the pharmaceutical industrial segment, which are produced as optically active compounds, are required as pure isomers.

Another concern when employing the ionic liquid phase in such reactions covers its potential utilisation for the catalyst recovery and reuse. In the specific case of tetra-alkyl ammonium bistriflimides there is an attractive option to perform the reversible biphasic reaction system (arrangement). At the reaction temperatures, and other conditions providing high conversions and optical yields, the reaction mixture is monophasic. At lower temperatures, for ionic liquids with long alkyl chains (e.g. [N14,222][Tf2N]), the system reveals the biphasic behaviour due to strong non-polar effects. The ionic liquid phase accommodates the chiral Ru complex, the water/methanol phase then the reaction products. After the reaction, when the reaction mixture is cooled to the room temperature, the catalytic complex is selectively kept in the ionic liquid phase, and it could be straightforwardly reused in the next reaction. The corresponding paper dedicated to the design, optimisation and practical performance of the reversible biphasic system is now being finalised.

In the introduction paragraphs it was repeatedly emphasised that for L-carnitin, the ethyl (R)-4-chloro-3-hydroxybutyrate intermediate is required. As shown above when (R)-Ru-BINAP was employed the reaction yielded ethyl (S)-4-chloro-3-hydroxybutyrate with a very high enantiomeric excess (ee = 99.4%, at 100% conversion of ECAA). Besides the gas chromatography this fact was fully proven by the NMR analysis. On the other hand stereoselective hydrogenation of methylacetoacetate (MAA) over (R)-Ru-BINAP yielded the (R) isomer [33]. The clarification is necessary at this point. The difference is rather formal and it must be interpreted with help of the IUPAC nomenclature rules.

When determining the absolute configuration ((R) or (S)) the substituents must be numbered according to the IUPAC priorities – a heavier atom bears the priority regarding the distance from the centre of chirality. Figure 8 shows ethyl (R)-3-hydroxypentanoate and ethyl (S)-4-chloro-3-hydroxybutyrate. It is noteworthy that the (S) configuration in spite of the orientation of substituents on the chirality centre is exactly the same as in the case of ethyl (R)-3-hydroxypentanoate. To verify that the developed process of stereoselective hydrogenation of ECAA is fully valid also for (R)-4-chloro-3-hydroxybutyrate, the reaction was repeatedly performed at previously optimised conditions but with (S)-Ru-BINAP. The origin and the optical purity of this complex were the same as for (R)-Ru-BINAP. The reaction was carried out only in the [N8,222][Tf2N]/methanol/water phase. At 408 K the achieved enantioselectivity towards ethyl (R)-4-chloro-3-hydroxybutyrate varied from 99.2 to 99.3%. It could be seen as nearly identical as for ethyl (S)-4-chloro-3-hydroxybutyrate (ee = 99.4%) (within the experimental error). The corresponding NMR spectra of ethyl (R)-4-chloro-3-hydroxybutyrate using the chiral solvating agent is presented in the Fig. 7b). The Fig. 7c) shows the superimposed spectra clarifying the assignment of residual peaks of the corresponding opposite enantiomer in reaction mixtures obtained with (R)- and (S)-Ru-BINAP. The assignment of individual signals was confirmed by analysis of complexes with analytical standards of both enantiomers of CHBE.
Fig. 8

Ethyl (R)-3-hydroxypentanoate (left), and ethyl (S)-4-chloro-3-hydroxybutyrate (right)

The character of the gas-liquid flow was observed with the 3223 chip (5 μl) and the MeOH/water phase. The in-situ images (Fig. 9) taken during the reactions clearly show that gas and liquid flow in separated slugs (Taylor flow [42]), however, their properties change considerably. From the pictures taken at three different temperature levels it is evident that the higher the temperature the higher the gas slugs, while the liquid slugs became shorter and occurs fewer. It is therefore obvious that the solvent (MeOH) was partially evaporated at such conditions due to relatively high vapour pressure. Note that the equilibrium vapour pressure of MeOH at 428 K is about 15 bars, which is located fairly below the working pressure (20 bar). Nevertheless, the partial evaporation of the solvent did not affect the ECAA conversion that was nearly 100% for the temperatures above 408 K.
Fig. 9

The gas-liquid flow through the reactor microchannel during the hydrogenation of ECAA at three different temperature levels, microchip 3223


Practical feasibility of the stereoselective hydrogenation of ECAA to (R)-CHBE, an intermediate for L-carnitine, over (S)-Ru-BINAP performed in the microfluidic chip reactor was verified as an alternative approach to standard biotechnology methods. The reaction conditions were first optimised using the (R)-Ru-BINAP complex yielding the (S)-CHBE isomer. Three different solvent phases were employed. The methanol/water phase, the ethanol/water phase, and the [N8,222][Tf2N]/methanol/water phase. The attained conversions were total in all cases at 408 K and higher. On the other hand the achieved optical purities of the principal product, (S)-CHBE, differed. The parameter of enantioselectivity ee was 99.4% towards the (S)-CHBE for the system in which (R)-Ru-BINAP was accommodated in [N8,222][Tf2N]/methanol/water phase. In the case of methanol/water experiment the ee parameter reached only 92.5%, for ethanol/water 91.8%). (R)-CHBE over (S)-Ru-BINAP was obtained with ee = 99.3% in the [N8,222][Tf2N]/methanol/water phase at 408 K. For the reactions in which (S)-CHBE was received apparent activation energies were evaluated. They were very similar for the systems MeOH/water and EtOH/water (110.5 and 110.7 kJ.mol−1). The apparent activation energies corresponding with the [N8,222][Tf2N]/MeOH/water system were much higher (of about 90 kJ.mol−1) reaching the level of 200 kJ.mol−1. It is noteworthy that the impact of the molecular structure of the main reactant was completely negligible as appeared from the comparisons of activation energies for ECAA hydrogenation and previously used methylacetoacetate (MAA, also ~ 200 kJ.mol−1). The effect of the presence of the [N8,222][Tf2N] ionic liquid on the sum of the activation energy dominated. Finally it must be added that the participation of the re-esterification reaction with methanol in the case of MeOH/water system was insignificant. Also the formation of acetals upon the reaction was marginal. Its suppression with water was very effective.



The authors acknowledge the Czech Science Foundation - Grant No. 15-04790S for funding this research.

Compliance with ethical standards

Conflict of interest

On behalf of all authors, the corresponding author states that there is no conflict of interest.


  1. 1.
    Cibulka R (2005) Metabolic effects of carnitine and its role in medicine. Clin Biochem Metab 13:24–28Google Scholar
  2. 2.
    Vaz FM, Wanders R (2002) Carnitine homeostasis and human disease: Advances in metabolite analysis. LabMedica 19:14–17Google Scholar
  3. 3.
    Karmaniolas K, Ionnidis P, Liatis S, Dalamanga M, Papalambros T, Migdalis I (2002) Primary carnitine deficiency in a male adult. J Med 14:105–110Google Scholar
  4. 4.
    Bjorklund G, Dadar M, Pen JJ (2019) Chronic fatigue syndrome – suggenstions for a nutritional treatment in the therapeutic approach. Biomed Pharmacother 109:1000–1007CrossRefGoogle Scholar
  5. 5.
    Guerreiro G, Faverzani J, Diaz JCE (2018) Oxidative damage in glutaric aciduria type I patiens and the protective effects of L-carnitine treatment. J Cell Biochem 119:10021–10032CrossRefGoogle Scholar
  6. 6.
    Belsky JB, Wira CR, Jacob V (2018) A review of micronutrients in sepsis of thiamine, L-carnitine, vitamin C, selenium and vitamin D. Nutr Res Rev 31:281–290CrossRefGoogle Scholar
  7. 7.
    Kalhori Z, Mehranjani MS, Azadbakht M (2019) L-carnitine improves function by reducing inflammation, oxidative stress and apoptosis in mice following induction of polycystic ovary syndrome. Reprod Fertil Develop 31:282–293CrossRefGoogle Scholar
  8. 8.
    Duca Y, Calogero AE, Cannarella R (2019) Current and emerging therapeutic agents for idiopathic male infertility. Expert Opin Pharmacother 20:55–67CrossRefGoogle Scholar
  9. 9.
    Wang R, wang L, Zhang C (2019) L-Carnitine ameliorates peripheral neuropathy in diabetic mice with a corresponding increase in insulin-like growth factor-1 level. Mol Med Rep 19:743–751Google Scholar
  10. 10.
    Endo S, Takahashi T, Sato M (2018) Effects of L-carnitine supplementation for chronic stroke patient. J Stroke Cerebrovas Dis 27:3342–3244CrossRefGoogle Scholar
  11. 11.
    Song JJ, Kyung J, Cho YK (2018) Effectiveness of high-dose carnitine complex treatment in patients with nonalcoholic fatty liver disease. J Gastroenterol Hepatol 33:423–424Google Scholar
  12. 12.
    Koeth RA, Lam-Galvez BR, Kirsop J (2019) L-carnitine in omnivorous diets induces an atherogenic gut microbial pathway in humans. J Clin Investig 129:373–387CrossRefGoogle Scholar
  13. 13.
    Patel RN (2011) Biocatalysis: synthesis of key intermediates for development of pharmaceuticals. ACS Catal 1(9):1056–1074CrossRefGoogle Scholar
  14. 14.
    Marzi M, Minetti P, Moretti G, Tinti MO, De Angelis F (2000) Efficient enantioselective synthesis of (R)-carnitine from glycerol. J Org Chem 65:6766–6769CrossRefGoogle Scholar
  15. 15.
    Wang GJ, Hollingsworth RI (1999) Synthetic routes to L-carnitine and L-gammaamino-beta-hydroxybutyric acid from (S)-3-hydroxybutyrolactone by functional group priority switching. Tetrahedron Asymmetry 10:1895–1901CrossRefGoogle Scholar
  16. 16.
    Matsuyama A, Yamamoto H, Kobayashi Y (2002) Practical application of recombinant whole-cell biocatalysts for the manufacturing of pharmaceutical intermediates such as chiral alcohols. Org Process Res Dev 6:558–561CrossRefGoogle Scholar
  17. 17.
    You ZY, Liu ZQ, Zheng YG (2014) Characterization of a newly synthesized carbonyl reductase and construction of a biocatalytic process for the synthesis of ethyl 4-chloro-3-hydroxybutanoate with high space-time yield. Appl Microbiol Biotechnol 98:1671–1680CrossRefGoogle Scholar
  18. 18.
    Applegate GA, Cheloha RW, Nelson DL, Berkowitz DB (2011) A new dehydrogenase from Clostridium acetobutylicum for asymmetric synthesis. Chem Commun 47:2420–2422CrossRefGoogle Scholar
  19. 19.
    Chen X, Liu ZQ, Lin CP (2016) Efficient biosynthesis of ethyl (R)-4-chloro3-hydroxybutyrate using a stereoselective carbonyl reductase from Burkholderia gladioli. BMC Biotechnol 16:70–75CrossRefGoogle Scholar
  20. 20.
    He YC, Tao ZC, Zhang X, Yang ZX, Xu JH (2014) Highly efficient synthesis of ethyl 4-chloro-3-hydroxybutanoate and its derivatives by a robust NADH-dependent reductase from E-coli CCZU-K14. Bioresour Technol 161:461–464CrossRefGoogle Scholar
  21. 21.
    Liu ZQ, Ye JJ, Hong HB, Yan JB, Lin Y, Chen ZX, Zheng YG, Shen YC (2015) Upscale production of ethyl 4-chloro-3-hydroxybutanoate by using carbonyl reductase coupled with glucose dehydrogenase in aqueousorganic solvent system. Appl Microbiol Biotechnol 99:2119–2129CrossRefGoogle Scholar
  22. 22.
    Blaser HU, Malan C, Pugin B, Spindler F, Steiner H, Studer M (2003) Selective hydrogenation for fine chemicals: recent trends and new developments. Adv Synth Catal 345:103–151CrossRefGoogle Scholar
  23. 23.
    Knowles WS, Noyori R (2007) Pioneering perspectives on asymmetric hydrogenation. Acc Chem Res 40:1238–1239CrossRefGoogle Scholar
  24. 24.
    Blaser HU, Spindler F, Studer A (2001) Enantioselective catalysis in fine chemicals production. Appl Catal A 221:119–143CrossRefGoogle Scholar
  25. 25.
    Floris T, Kluson P, Slater M (2011) Stereoselective hydrogenation of methyl acetoacetate over structurally different chiral ruthenium complexes. React Kinet Mech Catal 102:67–74CrossRefGoogle Scholar
  26. 26.
    Preetz A, Drexler HJ, Schulz S, Heller D (2010) BINAP: rhodium-diolefin complexes in asymmetric hydrogenation. Tetrahedron Asymmetry 21:1226–1231CrossRefGoogle Scholar
  27. 27.
    Floris T, Kluson P, Bartek L, Pelantova H (2009) Quaternary ammonium salts ionic liquids for immobilization of chiral Ru-BINAP complexes in asymmetric hydrogenation of β-ketoesters. Appl Catal A 366:160–165CrossRefGoogle Scholar
  28. 28.
    Cerna I, Kluson P, Bendova M, Floris T, Pelantova H, Pekarek T (2011) Intensification of the use of ionic liquids as efficient reaction co-solvents in asymmetric hydrogenations. Chem Eng Process 50:264CrossRefGoogle Scholar
  29. 29.
    Hallett JP, Welton T (2011) Room-temperature ionic liquids: solvents for synthesis and catalysis. 2. Chem Rev 111:3508–3576CrossRefGoogle Scholar
  30. 30.
    Olivier-Bourbigou H, Magna L, Morvan D (2010) Ionic liquids and catalysis: recent progress from knowledge to applications. Appl Catal A 373:1–56CrossRefGoogle Scholar
  31. 31.
    Picquet M, Poinsot D, Stutzmann S, Tkatchenko I, Tommasi I, Wasserscheid P, Zimmermann J (2004) Ionic liquids: media for better molecular catalysis. Top Catal 2:139–143CrossRefGoogle Scholar
  32. 32.
    Oechsner E, Schneiders K, Junge K, Beller M, Wasserscheid P (2009) Highly enantioselective Ru-catalyzed asymmetric hydrogenation of β-keto ester in ionic liquid/methanol mixtures. Appl Catal A 364:8–14CrossRefGoogle Scholar
  33. 33.
    Kluson P, Stavarek P, Penkavova V, Vychodilova H, Hejda S, Bendova M (2017) Microfluidic chip reactor and the stereoselective hydrogenation of MAA over Ru-BINAP in the [N8,222][Tf2N] – methanol – water mixed phase. Chem Eng Process 115:39–44CrossRefGoogle Scholar
  34. 34.
    Kluson P, Stavarek P, Hejda S, Penkavova V, Bendova M, Vychodilova H, Vlcek D (2017) Molecular structure effects of [NR,222][Tf2N] ionic liquids on their flow properties in the microfluidic chip reactor – a complete validation study. Chem Eng Process 111:57–64CrossRefGoogle Scholar
  35. 35.
    Dytrych P, Kluson P, Slater M, Solcova O (2014) Theoretical interpretation of the role of the ionic liquid phase in the (R)-Ru-BINAP catalyzed hydrogenation of methylacetoacetate. React Kinet Mech Catal 111:475–479CrossRefGoogle Scholar
  36. 36.
    Gutmann B, Cantillo D, Kappe CO (2015) Continuous-flow technology - a tool for the safe manufacturing of active pharmaceutical ingredients. Angew Chem Int Ed 54:6688–6728CrossRefGoogle Scholar
  37. 37.
    Kluson P, Stavarek P, Hejda S, Penkavova V, Bendova M, Vychodilova H (2016) Microreactors and microfluidic reactors for the synthesis of fine chemicals. Chem List 110:892–899Google Scholar
  38. 38.
    Kockmann N, Thenée P, Fleischer-Trebes C, Laudadioc G, Noël T (2017) Safety assessment in development and operation of modular continuous-flow processes. React Chem Eng 2:258–280CrossRefGoogle Scholar
  39. 39.
    Curinova P, Dracinsky M, Jakubec M, Tlusty M, Janku K, Izak P, Holakovsky R (2018) Enantioselective complexation of 1-phenylethanol with chiral compounds bearing urea moiety. Chirality 30:798–806CrossRefGoogle Scholar
  40. 40.
    Bartek L, Drobek M, Kuzma M, Kluson P (2005) Notes on acetal formation in stereoselective hydrogenation of methyl-3-oxobutyrate on Ru-BINAP chiral complex. Catal Commun 6:61–65CrossRefGoogle Scholar
  41. 41.
    Wolfson A, et al (2003) The role of the solvent in the asymmetric hydrogenation of β-keto esters with Ru-BINAP. J Mol Catal A Chem 198(1):39–45.
  42. 42.
    Angeli P, Gavriilidis A (2008) Hydrodynamics of Taylor flow in small channels: a review. J Mech Eng Sci 222:737–751CrossRefGoogle Scholar

Copyright information

© Akadémiai Kiadó 2019

Authors and Affiliations

  • Petr Kluson
    • 1
    • 2
    Email author
  • Petr Stavarek
    • 1
  • Vera Penkavova
    • 1
  • Hana Vychodilova
    • 1
  • Stanislav Hejda
    • 1
  • Natalie Jaklova
    • 3
  • Petra Curinova
    • 1
  1. 1.Institute of Chemical Process Fundamentals v. v. iCzech Academy of SciencesPragueCzech Republic
  2. 2.Institute of Environmental Studies, Faculty of Nature ScienceCharles University in PraguePragueCzech Republic
  3. 3.Chemistry section, Faculty of Nature ScienceCharles University in PraguePragueCzech Republic

Personalised recommendations