Process for the preparation of cyclopropane compounds using diazo-compounds

Abstract

A process for the continuous production of a reaction product of a diazo-compound and a substrate in a multi-stage flow reactor is disclosed.

Claims

1. A process for the continuous production of a reaction product of a diazo compound and a substrate in a multi-stage continuous flow reactor comprising at least two connected reactors in fluid communication with each other, wherein a reaction product formed at each stage of the multistage sequence is continuously discharged from an upstream reactor in which it is formed, and is fed as a continuous fluid stream into a downstream reactor whereupon it can undergo further conversion, the process being characterized in that a precursor compound of the diazo compound that is continuously formed in an upstream reactor is continuously discharged into a downstream reactor under conditions conducive to convert it into the diazo-compound, and wherein said downstream reactor contains the substrate with which the diazo-compound reacts to form the reaction product of the diazo compound and the substrate.

2. The process according to claim 1, wherein the diazo-compound is selected from the group consisting of diazoalkanes, diazocarboxylates, and combinations thereof.

3. The process according to claim 1, wherein the substrate is selected from the group consisting of olefins, aldehydes, ketones, cyclic ketones, carboxylic acids, acid chlorides, and combinations thereof.

4. The process according to claim 3, wherein the substrate is an olefin selected from di- or polyolefins.

5. The process according to claim 3, wherein the olefin is selected from the group consisting of isoprene, myrcene, farnesene, β-springene, and combinations thereof.

6. The process according to claim 1, wherein the diazo compound precursor is an N-alkyl-N-nitroso compound selected from the group consisting of N-methyl-N-nitroso urea (MNU), N-Methyl-N-nitroso-p-toluenesulfonamide, methyl N-methyl-N-nitroso carbamate, ethyl N-methyl-N-nitroso carbamate, N-methyl-N-(2-methyl-4-oxopentan-2-yl)nitroso amide, and combinations thereof.

7. The process according to claim 1, wherein the diazo compound is diazomethane, the substrate is an olefin and their reaction product is a cyclopropanated compound.

8. The process according to claim 7, wherein the diazomethane precursor compound formed in the upstream reactor is mixed with the olefin substrate, a catalyst and an organic solvent to form an organic phase, and this organic phase is mixed with an aqueous phase containing base to form a biphasic reaction mixture, which reaction mixture reacts to form the cyclopropanated compound in the downstream reactor.

9. The process according to claim 8, wherein the diazomethane precursor compound formed in the upstream reactor is contained in the organic phase of a biphasic mixture, and wherein the organic phase containing the diazomethane precursor compound is separated from the aqueous phase of the biphasic mixture before being mixed with an olefin substrate, catalyst and organic solvent.

10. The process according to claim 7, comprising the steps of: providing a multi-stage continuous flow reactor comprising at least two connected reactors in fluid communication with each other; at each stage of the reaction, continuously feeding a fluid stream of reactants in a directed flow into a reactor under conditions conducive to form a reaction product; continuously discharging from that reactor a fluid stream of reaction product, and feeding it into a successive reactor there to undergo further conversion to form a successive reaction product; characterized in that the diazomethane precursor compound formed as a reaction product in a reactor is continuously fed into a successive reactor under conditions conducive to convert the precursor compound into diazomethane, and wherein said successive reactor contains a the olefin substrate with which the diazomethane reacts to form the cyclopropanated compound.

11. The process according to claim 10, wherein the diazomethane precursor compound, discharged from the reactor is mixed continuously with the catalyst, substrate and solvent to form an organic phase, and this organic phase is mixed and reacted continuously with an aqueous phase containing a base in the successive reactor thereby to form the cyclopropanated compound.

12. The process according to claim 1 carried out under flow conditions.

13. The process according to claim 12, wherein the reaction is carried out in a microreactor.

14. The process according to claim 2, wherein the diazoalkane comprises at least one of diazomethane or diazoethane.

15. The process according to claim 2, wherein the diazocarboxylate comprises diazoacetate.

16. The process according to claim 4, wherein the substrate is an olefin selected from di- or polyolefins having at least one terminal (mono-substituted) double bond.

17. The process according to claim 11, wherein the base comprises KOH.

18. The process according to claim 10, wherein the multi-stage continuous flow reactor comprises at least three connected reactors in fluid communication with each other.

19. The process according to claim 10, wherein the multi-stage continuous flow reactor comprises at least four connected reactors in fluid communication with each other.

Description

(1) The following figures serve to further illustrate the invention or specific aspects of the invention:

(2) FIG. 1 shows a schematic representation of a multistage continuous flow reactor suitable for a process for the continuous production of a “diazo reaction product”;

(3) FIG. 2 shows a schematic representation of a reaction chamber in which the diazo reaction product is formed, and a new addition mode of the reagents is applied in the flow process;

(4) FIG. 3 shows a schematic representation of an alternative reaction chamber in which the diazo reaction product is formed, and a batch-related addition mode of the reagents is applied in the flow process; and

(5) FIG. 4 shows a schematic representation of an alternative multistage continuous flow reactor, suitable for a process for the continuous production of a “diazo reaction product”, further comprising multiple Continuously Stirred Tank Reactors (CSTRs) and biphasic settlers.

(6) In the process of converting a precursor compound (e.g. Liquizald) into diazomethane and further into the diazo reaction product, the sequence of addition of reactants and reagents is of particular importance. When such a synthetic step is carried out as described in prior art batch processeses, (e.g. as in WO 2015059290), Liquizald in a suitable organic solvent, such as toluene, is added to a vigorously stirred biphasic mixture comprising a substrate containing an olefinic double bond, such as farnesene, a suitable catalyst, e.g. Pd(acac).sub.2 and a base, e.g. KOH. A representative example of such a batch process is set out in the examples (see Example 1). However, adopting this addition sequence in the method of the present invention using continuous flow is problematic. More particularly, it creates additional complexity in the apparatus set-up, which requires a greater number of pumps and microreactors owing to the need for pre-mixing and pumping of the biphasic mixture before the addition of Liquizald. Furthermore, the reaction results in a much lower conversion of substrate, which can possibly be attributed to the pumping of inhomogeneous mixtures and poor phase mixing in the reactor. This is further demonstrated in the Examples, and more particularly Examples 7 and 8 which are representing a set up according to FIG. 3, as well as a comparison of FIGS. 2 and 3.

(7) An addition sequence wherein an aqueous phase (e.g. KOH in water) and an organic phase (Pd(acac).sub.2, Liquizald and Farnesene in toluene) are continuously mixed, was found to be much more efficient and simple regarding optimizing conversions and technical set-up (FIG. 2). An example of this reaction in a single microreactor, which corresponds to a reactor or a reaction chamber, and residence unit is set-forth in Example 2. FIG. 2 illustrates the arrangement of the last reaction chamber in FIG. 1 or FIG. 4, in which the diazo reaction product is formed, but can also be regarded as an independent continuous flow reactor.

(8) Steps in the multi-stage process of forming a cyclopropanated compound described hereinabove may be carried out in a biphasic medium comprising an organic and aqueous phase. By way of example, after formation of 4-methyl-4-(methylamino)pentan-2-one, the successive reaction products formed during each stage of the process are soluble or dispersible in an organic phase. Therefore, if it is desired, aqueous-soluble or dispersible effluents or by-products can be continuously separated from the organic phase by means of phase separation techniques.

(9) Separation of the precursor compound—Liquizald—from the aqueous phase is of particular importance because Liquizald is formed under acidic conditions, whereas the formation of diazomethane from Liquizald is carried out in a biphasic mixture under basic conditions. Accordingly, before Liquizald is continuously fed into a reaction chamber to form diazomethane, it is desirable to separate the Liquizald-containing organic phase from the aqueous phase by suitable phase separation means, and it is furthermore beneficial before the subsequent synthetic step is undertaken, to remove traces of acid from the Liquizald-containing phase by washing it with portions of a suitable aqueous base.

(10) Phase separation and washing with aqueous base can be carried out under continuous flow conditions by passing the biphasic reaction mixture containing the precursor compound (e.g. Liquizald) through a continuous extraction apparatus, such as Continuously Stirred Tank Reactors (CSTRs). The use of counter current extraction in multistep flow sequences is described for example by K. P. Cole et al. in Science 356, 1144-1150 (2017). Ideally CSTRs are miniaturized to decrease the volumes of Liquizald during extraction and washing, which can be done for example with liquid-liquid separators (LLS) and agitated cell extractors (ACX), which decrease the stationary amount of Liquizald, and feeding the Liquizald-containing phase into a successive reactor for the cyclopropanation step. The principle of using liquid-liquid separators (LLS) is described for example in technical notes from companies such as Zaiput Flow Technologies or in multistep flow publications which describe the use of such LLS, see for example J. Britton, C. L. Raston, Chem. Soc. Rev. 46, 1250-1271 (2017). The principle of using agitated cell extractors for counter-current extraction is generally well known in the art, and described for example in technical notes from companies such as AM Technology or in reviews describing the principles of such techniques, see for example F. Visscher, J. van der Schaaf, T. A. Nijhuis, J. C. Schouten, Chemical Engineering Research and Design 91, 1923-1940 (2013).

(11) FIG. 1 depicts a schematic representation of a process and apparatus for the continuous production of a cyclopropanated compound in accordance with a particularly preferred embodiment of the invention, which Figure and description serve to further describe the invention. The description refers, for sake of simplicity, to the continuous synthesis of Liquizald and its use for the cyclopropanation of Farnesene, but any other precursor compound, such as any N-alkyl-N-(2-methyl-4-oxopentan-2-yl)nitrous amide precursor can be made and any other alkene substrate used for cyclopropanation. The following elements are arranged in FIG. 1: reservoirs 101-107, pumps 111-119, flow in tube represented by arrows, microreactor 121, residence coil with static mixture 131, quenching vessel 141, plug flow reactors 151-152, liquid-liquid separator LLS 161, back pressure regulator 171, agitated cell extractor (ACX) 181, T-connections 191-194. With reference to FIG. 1: in a first step mesityl oxide (from reservoir 101, through feed A, pump 111) and methyl amine (from reservoir 102, through feed B, pump 112) are mixed in a T-piece (191, T.sub.1) and a plug flow reactor (151, pfr.sub.1) in which these components react to produce 4-methyl-4-(methylamino)pentan-2-one (MOMA=Mesityl Oxide/Methyl Amine adduct); in a second and third step an acid in toluene (from reservoir 103, through feed C, pump 113) and sodium nitrite in water (from reservoir 104, through feed D, pump 114) are added in two subsequent T-pieces (192, T.sub.2 and 193, T.sub.3) whereupon these components react to provide N-methyl-N-(2-methyl-4-oxopentan-2-yl) nitrous amide (Liquizald) in the third T-piece (193, T.sub.3) and a subsequent plug flow reactor (152, pfr.sub.2). Alternatively sodium nitrite can be added first, then the acid; in a fourth and fifth step the resulting biphasic mixture from step 3 is separated by a liquid-liquid separator (161, LLS, step 4), the acidic aqueous waste is removed (passing a back pressure regulator 171). The organic phase moves forward (pump 115), and residual acid is removed by continuous washing (step 5) in an agitated cell extractor (181, ACX), to which aqueous base is added (from reservoir 105, through feed E, pump 116); in a sixth step the organic layer containing Liquizald is pumped (pump 117) into a T-piece (194, T.sub.4) where it is mixed with the polyene substrate (Farnesene) and catalyst in toluene (from reservoir 106, through feed F, pump 118); The resulting Liquizald-containing mixture is pumped into a microreactor (121) where it is mixed in a seventh step with aqueous base (from reservoir 107, through feed G, pump 119), whereupon the cyclopropananation product is formed in the microreactor 121 and a subsequent residence coil 131. The resulting cyclopropanated product mixture is directed from the residence coil into a vessel 141 containing a quench solution where the biphasic mixture is permitted to settle. The quench-solution is used for safety reasons. In case of fully converted Liquizald and diazomethane quenching is not necessary.

(12) After step seven the biphasic mixture comprises the cyclopropanated compound (the diazo reaction product) in an organic phase and an aqueous phase containing a waste mixture, which can be separated by normal batch-wise separation because reaction parameters can be adjusted to ensure that this mixture does not contain any unreacted Liquizald and/or diazomethane. Optionally it is of course possible to continue with the work-up and purification of the product under flow conditions.

(13) There now follows a series of examples that serve to further illustrate the invention. The following examples are given to illustrate preferred embodiments of the invention. It will be understood that these examples are illustrative and the invention is not to be considered as restricted thereto.

(14) General:

(15) Non-polar GC condition: 100° C., 1 min, 20° C./min to 240° C., 5 min, 240° C. GC Agilent 7890B Series GC system. Non-polar column: HP-5 from Agilent Technologies, 0.32 mm×0.25 mm×30 m. Carrier gas: hydrogen. Injector temperature: 230° C. Split 1:50. GC calibration with Liquizald 80% (as determined by NMR) and decane. rpa %=relative peak areas.

(16) Non-polar GCMS-conditions: 50° C., 2 min, 20° C./min to 240° C., 35° C./min to 270° C. GCMS Agilent 5975C MSD with HP 7890A Series GC system. Non-polar column: BPX5 from SGE, 5% phenyl 95% dimethylpolysiloxan 0.2 mm×0.25 mm×12 m. Carrier gas: helium. Injector temperature: 230° C. Split 1:50. Flow: 1.0 ml/min. Transfer line: 230° C. MS-quadrupol: 150° C. MS-source: 230° C. rpa %=relative peak areas.

Example 1 (Comparative)

(17) The cyclopropanation reaction under batch conditions is set forth hereinbelow:

(18) ##STR00001##

(19) Liquizald 46% in toluene (18.3 g, 53 mmol, 1.7 eq) is added at 45° C. and over 2 h to a stirred biphasic mixture of Farnesene 98% (6.4 g, 31 mol) and Pd(acac).sub.2 (6.3 mg, 20 μmol, 0.065%) in toluene (27 ml) and KOH (8 g, 0.12 mol) in water (34 ml). After 30 min at 45° C. and complete conversion as detected by GC the reaction mixture is poured upon acetic acid (25 ml). After phase separation the organic phase is washed with water, dried over MgSO.sub.4, filtered, and the solvents are evaporated. The brown orange oil (7 g) contains 92% Δ-Farnesene, 8% of Δ.sub.2-Farnesene according to GC which corresponds with a GC-calibrated purity (84%) of Δ-Farnesene to a corrected yield of 86% Δ-Farnesene based on Farnesene.

(20) ##STR00002##

Example 2

(21) Cyclopropanation reaction under flow conditions and new addition mode with 0.065% Pd(acac).sub.2, as set forth in FIG. 2.

(22) Equipment: Sigma-Aldrich Microreaction Explorer Kit No. 19979-1kt, © 2007, Version 1.02, Sigma Aldrich Production GmbH, Buchs, Switzerland. Microreaction plate: MR #437383, Little Things Factory, Volume 0.7 ml, PTFE tube, borosilicate glass, ceramics, max. inner pressure 6.5 bar. Two 100 ml round bottom flasks. 200 ml sulfonation-flask with acetic acid as quench. One gas washing bottle with acetic acid and one safety bottle. Two Ismatec pumps. Residence coil: 1.25 m stainless steel tube, 5 mm inner diameter, 18 ml volume, filled with 19 static polypropylene mixers. The self-made residence unit is described in the bachelor work of Fabian Ruethi (“TPO continuously for Silvanone”, page 72 and 124, Zürcher Hochschule für Angewandte Wissenschaften, 2012): 1.25 m stainless steel tube, 5 mm inner diameter, 18 ml volume, filled with 19 static mixers. Static mixers: 6.5 mm length, 4.8 mm diameter. Pumps, the microreactor and the residence coil are connected by teflon tubings with 1.5 mm inner diameter.

(23) Feed A: Farnesene of 98% purity (6.4 g, 31 mmol), Pd(acac).sub.2 (6.3 mg, 0.02 mmol. 0.065 mol %) in toluene (5 ml) [Note 1], and Liquizald 33% in toluene (25.7 g, 53.6 mmol, 1.7 mol-eq) pre-mixed in toluene (13 ml) pumped with 3.4 ml/min into microreactor and residence unit. Feed B: 3.5 M aqueous KOH (34 ml, 117 mmol) pumped with 2.2 ml/min into microreactor and residence unit.

(24) Feed A (from reservoir 201) and feed B (from reservoir 202) are pumped (pumps 211 and 212) through the set-up of FIG. 2 into microreactor 221 and residence coil 231 in an oil bath heated to 75° C. outer and at 70-60° C. inner temperature [Note 2]. The product flow (5.6 ml/min, residence time 3.3 min in microreactor and residence unit) is quenched (in vessel 241) sub-surfacely and under cooling (at 10-20° C. inner temperature to avoid solidification of HOAc) through acetic acid (15 g, 0.25 mol) [Note 3]. The biphasic quench mixture is transferred to a separation funnel, the phases are separated, and the organic phase is washed with water (100 ml), dried over MgSO.sub.4 (30 g), filtered and evaporated under reduced pressure. Crude Δ-Farnesene (6.9 g) is obtained as brown oil with a purity of 82.5%, as determined by GC- and NMR-calibration [Note 4]. GCMS of this material indicates a quantitative conversion to Δ-Farnesene (92% rpa) and Δ.sub.2-Farnesene (8% rpa) and the absence of Liquizald. Yield of Δ-Farnesene from Farnesene (corrected by purity): 84% [Note 4]. The analytical data of Δ-Farnesene and Δ.sub.2-Farnesene were identical with the ones described in WO 2015059290.

(25) Notes:

(26) [1] The catalyst Pd(acac).sub.2 is soluble in toluene.

(27) [2] Sensor in the microreactor not connected with the reaction stream.

(28) [3] Pure HOAc was used because complete cyclopropanation in the microreactor had to be proven and possible post-reactions in the quenching vessel had to be excluded. Excess HOAc (2×molar excess over KOH used) guaranteed non-basic quench conditions. Nitrogen is produced in the microreactor by controlled decomposition of the diazomethane and is led from the quench vessel through a washing bottle filled with pure HOAc. Only traces of MeOAc (˜0.02%) were detected in the washing bottle by NMR and only after several similar optimization experiments using the same washing bottle. The quench-solution can be applied for safety reasons. In case of fully converted Liquizald and diazomethane, quenching is not necessary.

(29) [4] The content of Δ-Farnesene was determined by NMR- and GC-calibration to 82.5%.

(30) A comparison of Example 1 (batch reaction) and Example 2 (flow reaction) shows that under otherwise identical conditions yields and purities are (within an error of 2%) similar under batch and flow conditions.

Example 3

(31) The flow reaction described in Example 2 was repeated under nearly identical conditions and in the same equipment but without residence unit:

(32) Feed A: Farnesene of 98% purity (6.76 g, 32 mmol), Pd(acac).sub.2 (6.5 mg, 21 μmol, 0.065 mol %) and Liquizald 38% in toluene (23.34 g, 55.3 mmol, 1.7 mol-eq) pre-mixed in toluene (18 ml) pumped with 1.7 ml/min into the microreactor without residence unit. Feed B: 3.5 M aqueous KOH (33 ml, 114 mmol) pumped with 1 ml/min into the microreactor without residence unit.

(33) Feed A and feed B are pumped (as described in example 2 subsequent but without residence coil) into the microreactor in an oil bath heated to 75° C. outer and at 70-60° C. inner temperature. The product flow (2.7 ml/min, residence time 15 sec in the microreactor) is quenched sub-surfacely and under cooling (at 10-20° C.) through acetic acid (15 g, 0.25 mol). A GC taken 30 min after completed quench shows Liquizald (26%), Farnesene (67%) and Δ-Farnesene (7%).

(34) This experiment shows that a 20-fold reduced residence time due to the absence of a residence unit drastically reduces the conversion of Farnesene and Liquizald to Δ-Farnesene.

Example 4

(35) The flow reaction described in Example 2 was run with 1.2 mol-eq Liquizald (instead of 1.7 mol-eq versus Farnesene):

(36) Feed A: Farnesene of 98% purity (8.96 g, 43 mmol), Pd(acac).sub.2 (8.8 mg, 0.03 mmol) and Liquizald 40% in toluene (20.8 g, 52.6 mmol) pre-mixed in toluene (22 ml) pumped with 3.4 ml/min into microreactor and residence unit. Feed B: 1.75 M aqueous KOH (23 ml, 40 mmol) pumped with 1.4 ml/min into microreactor and residence unit.

(37) Feed A and feed B are pumped through the set-up of FIG. 2 into microreactor and residence coil in an oil bath heated to 75° C. outer and at 70-60° C. inner temperature. The product flow (4.8 ml/min, residence time 4 min) is quenched sub-surfacely and under cooling (10-20° C.) through acetic acid (15 g, 0.25 mol) [Note 3]. The biphasic quench mixture is transferred to a separation funnel, the phases are separated and the organic phase is washed with water (100 ml), dried over MgSO.sub.4 (30 g), filtered and evaporated under reduced pressure. Crude Δ-Farnesene (11.6 g) is obtained as brown oil with a purity of 66%, as determined by GC- and NMR-calibration]. GCMS of this material indicates a 91% conversion to Δ-Farnesene (85% rpa) and Δ.sub.2-Farnesene (6% rpa) and the absence of Liquizald. Yield of Δ-Farnesene from Farnesene (corrected by purity): 82%.

(38) In comparison with example 2 this result shows that with less Liquizald (Liquizald/Farnesene ratio 1.2 instead of 1.7) the conversion of Farnesene is still extensive, and that the corr. yield of Δ-Farnesene (82%) is still comparable to the one obtained using more (1.7 eq) Liquizald (84% in example 2). It must be mentioned at this point that incomplete conversions are rather acceptable than over-cyclopropanation to Δ.sub.2-Farnesene because the bis-cyclopropanated byproduct cannot be recycled, whereas unconverted Farnesene can be, after distillative separation from Δ-Farnesene, recycled as substrate.

Example 5

(39) The flow reaction described in Example 2 was run with 0.02% catalyst (instead of 0.065% catalyst).

(40) Feed A: Farnesene of 98% purity (6.48 g, 31 mmol), Pd(acac).sub.2 (2 mg, 6.4 μmol) and Liquizald 43% in toluene (19.8 g, 54 mmol, 1.7 eq) pre-mixed in toluene (24 ml) pumped with 1.7 ml/min into microreactor and residence unit. Feed B: 3.55 M aqueous KOH (34 ml, 120 mmol) pumped with 1.1 ml/min into microreactor and residence unit.

(41) Feed A and feed B are pumped as described in example 2 and FIG. 2 into microreactor and residence coil. Total Flow: 2.8 ml/min. Residence time: 6.5 min. The product flow is quenched subsurfacely through acetic acid. A GC taken 30 min after complete quench indicates a 82% conversion to Δ-Farnesene (77% rpa) and Δ.sub.2-Farnesene (4% rpa) and the absence of Liquizald.

(42) This example shows that at much lower catalyst concentration (0.02 instead of 0.065 mol %) still acceptable conversions are obtained in flow.

Example 6

(43) The flow reaction described in Example 5 was run in batch mode.

(44) Liquizald 37% in toluene (23.1 g, 53 mmol, 1.7 eq) is added at 45° C. and over 2 h to a stirred biphasic mixture of Farnesene 98% (6.4 g, 31 mol) and Pd(acac).sub.2 (2 mg, 6.3 μmol, 0.02%) in toluene (27 ml) and KOH (8 g, 0.12 mol) in water (34 ml). A GC taken 30 min after complete quench indicates a 78% conversion to Δ-Farnesene (70%), Δ.sub.2-Farnesene (4%) and another monocyclopropanated byproduct (2%).

(45) A comparison of the results from batch (example 6) and flow mode (example 5) shows, that conversion and purity obtained under flow conditions (82% with 77% rpa Δ-Farnesene) are at least as good if not even better than the ones obtained under batch conditions (77% with 70% rpa Δ-Farnesene).

Example 7

(46) Cyclopropanation reaction under flow conditions at 0.065% catalyst/substrate ratio and with batch-(example 1)-related addition mode as set forth in FIG. 3:

(47) Equipment: as in example 2 but two microreactors (instead of one) and three pumps (instead of two).

(48) Feed A: Farnesene of 98% purity (6.3 g, 31 mmol), Pd(acac).sub.2 (6.2 mg, 0.02 mmol. 0.065 mol %) in toluene (13 ml) pumped with 1.5 ml/min. Feed B: 3.5 M aqueous KOH (34 ml, 118 mmol) pumped with 2.2 ml/min. Feed C: Liquizald 33.5% in toluene (24.9 g) and toluene (1 ml) pumped with 1.9 ml/min.

(49) Feed A (from reservoir 301) and feed B (from reservoir 302) are pumped (pumps 311 and 312) through the set-up of FIG. 3 into the first microreactor (321, mrx 1), where a biphasic mixture is preformed. From the first microreactor the flow (3.7 ml/min) is directed into the second microreactor (322, mrx 2) where it is combined with feed C (from reservoir 303) which is pumped (pump 313) with 1.9 ml/min into the second microreactor 322. From mrx 2 the flow is directed into the residence coil 331 from which the product flow (5.6 ml/min) is quenched (in vessel 341) sub-surfacely and under cooling (at 10-20° C. inner temperature through acetic acid (15 g, 0.25 mol). Work-up as described in example 2 gives a mixture (7 g) of unconverted Farnesene and crude Δ-Farnesene as a brown oil with a purity of 47%, as determined by GC- and NMR-calibration. GC of this material indicates Farnesene (44% rpa), Δ-Farnesene (56% rpa) and traces of Liquizald and Δ.sub.2-Farnesene. Yield of Δ-Farnesene from Farnesene (corrected by purity): 48%.

(50) This experiment shows that an addition mode as in batch example 1 which adds Liquizald to a preformed 2-phase mixture of Farnesene, Pd(acac).sub.2 in toluene and KOH in water, has two main disadvantages under flow conditions: a higher complexity (more pumps and microreactors) a significantly lower conversion (56% instead of 100%) of Farnesene which is probably due to pumping of an inhomogeneous mixture (feed A+feed B) into the second microreactor.

Example 8

(51) Cyclopropanation reaction under flow conditions at 0.02% catalyst/substrate ratio and with batch-(example 1)-related addition mode as set forth in FIG. 3:

(52) Feed A: Farnesene of 98% purity (6.4 g, 31 mmol) and Pd(acac).sub.2 (1.9 mg, 0.02 mmol, 0.02 mol %) in toluene (13 ml) pumped with 0.74 ml/min. Feed B: 3.5 M aqueous KOH (34 ml, 118 mmol) pumped with 1.1 ml/min. Feed C: Liquizald 39% in toluene (21.3 g) and toluene (5 ml) pumped with 0.96 ml/min.

(53) The flow reaction described in example 7 was carried out with the same equipment set-up but with slightly changed feeds and with 2 mg (6.4 μmol, 0.02 mol %) Pd(acac).sub.2 instead of 6.2 mg (0.02 mmol, 0.065 mol %). Work-up as described in example 2 gave a mixture (7 g) of unconverted Farnesene and crude Δ-Farnesene as a brown oil with a purity of 43%, as determined by GC- and NMR-calibration. GC of this material indicates Farnesene (47% rpa), Δ-Farnesene (50% rpa), Δ.sub.3-Farnesene (3% rpa) and traces of Δ.sub.2-Farnesene and Liquizald. Yield of Δ-Farnesene from Farnesene (corrected by purity): 44%.

(54) ##STR00003##

(55) This experiment shows similar effects as in example 7 at lower catalyst concentrations.

Example 9

(56) A multistep flow experiment as set forth in FIG. 4.

(57) Equipment: as in the flow reaction described in Example 2 (and represented by FIG. 1) but with some additional devices: five additional feeding pumps (total seven) 6 peristaltic pumps 3 additional T-connections (total 4) two plug flow reactors (pfr) five additional reservoirs for additional reagents, reactants and solvents (total seven) four CSTR's (continuously stirred tank reactors) with biphasic settlers 2 back pressure regulators

(58) Pumps, T-connections, microreactors, plug flow reactors and CSTR's are connected by teflon tubings. The T-connections are also made of teflon and are, for example, available from Sigma-Aldrich (Merck). The plug flow reactors are 1.25 m stainless steel tubes. The residence coil filled with static mixers is described in example 2.

(59) Feed A: mesityl oxide (700 g, 7 mol). Feed B: methyl amine 40% in water (580 ml, 7.5 mol). Feed C: acid (11.4 mol) in 700 ml toluene. Feed D: NaNO.sub.2 (543 g, 7.6 mol) in water (820 ml). Feed E: base 25% in water (1400 ml). Feed F: Farnesene (640 g, 3 mol) and Pd(acac).sub.2 (0.2 mg, 0.65 mmol. 0.02 mol %) in toluene (550 ml). Feed G: 3.5 M aqueous KOH (3400 ml, 11.7 mol). Flow rates, residence times and tube diameters are omitted for clarification.

(60) Feed A (from reservoir 401) and feed B (from reservoir 402) are pumped (pumps 411 and 412) through the set-up of the scheme above into T-connection T.sub.1 (491) and the first plug flow reactor (451, pfr.sub.1) where these reactants undergo Michael addition to the mesityloxide methylamine adduct (MOMA), which is directed into T-connection T.sub.2 (492), where acid in toluene (from reservoir 403, through feed C) and NaNO.sub.2 in water (rom reservoir 404, through feed D) are subsequently added through two pumps (413, 414). After acidic nitrosation in the T-connection T.sub.3 (493) and second plug flow reactor (452, pfr.sub.2) the resulting Liquizald in a biphasic mixture in toluene and water is fed into a first CSTR (461a, continuously stirred tank reactor) and a biphasic settler (461b), where the phases are separated into an acidic waste water stream (passing a back pressure regulator 471) and a Liquizald stream in toluene, which is fed into the next three CSTR's (462a, 463a, 464a) and the corresponding biphasic settlers (462b, 463b, 464b) by several peristaltic pumps. Washing is affected with aqueous base (from reservoir 405, feed E) which is pumped (pump 415) countercurrently from the fourth CSTR (464a) and its biphasic settler (464b) via third CSTR (463a) and its biphasic settler (463b) to the second CSTR (462a) from which the basic waste water stream is discarded passing the biphasic settler (462b) and a back pressure regulator 472. A stream of essentially acid-free Liquizald in toluene is directed (pump 416) from the biphasic settler 464b of CSTR 4 (464a) to T-connection T.sub.4 (494). Concomitantly, Farnesene and catalytic Pd(acac).sub.2 (from reservoir 406, feed F) are pumped (pump 417) into T-connectionT.sub.4 (494) giving a homogeneous organic stream which is directed into microreactor (421) and residence unit (431) rc.sub.3 filled with static mixers. Concomitantly, homogeneous aqueous base (from reservoir 407, feed G) is pumped (pump 418) into the microreactor (421) and residence unit (431, rc.sub.3). After cyclopropanation reaction at 60° C. in the microreactor and residence unit filled with static mixers the product is quenched sub-surfacely (vessel 441) and under cooling through acetic acid (1.5 kg, 22.5 mol) The biphasic quench mixture is separated, the organic phase washed with water (10 l) dried azeotropically and evaporated under reduced pressure. Crude Δ-Farnesene (600 g) is obtained as brown oil with a purity of 80% (GC rpa) containing Δ.sub.2-Farnesene (3% rpa) and unconverted Farnesene (16% rpa). Yield of Δ-Farnesene from Farnesene (corrected by purity): 78%. The crude product is further purified by distillation.