preparation of enantiomers using high- pressure technologies · preparation of enantiomers using...

UNIVERSITY OF MARIBOR

FACULTY OF CHEMISTRY AND CHEMICAL ENGINEERING

Doctoral thesis

Preparation of enantiomers using high-pressure technologies

Author: Paul Thorey

Mentor: prof. dr. Maja Habulin

Co-mentor: prof. dr. Béla Simándi

Maribor, 2010

Preparation of enantiomers using high-pressure technologies

Abstract

The study of two different methods of obtaining chiral alcohols is proposed herein.

The requirement of the relatively new paradigm of green chemistry associated with clean

technologies such as biocatalysis or non-conventional solvents, dense gases, was

focused at. Indeed, the two methods of production of chiral alcohols were:

the conversion of acetophenone into (R)-1-phenylethanol in dense gases

catalysed by Lactobacillus brevis alcohol dehydrogenase and its

coenzyme, NADP/H;

the resolution of (±)-trans-1,2-cyclohexanediol by cocrystal formation with

tartaric acid followed by supercritical extraction.

In both cases high enantiopurities were achieved (ee>99%).

Key words:

High-pressure technologies, enantiomers, green chemistry, R-1-phenylethanol,

Lactobacillus brevis, alcohol dehydrogenase, NADP, liquid propane, enzyme deactivation,

resolution, trans-1,2-cyclohexanediol, tartaric acid, cocrystal, supercritical carbon dioxide,

extraction, X-ray diffraction, differential scanning calorimetry.

UDK: 66 – 987 : 544 . 122 . 3 (043 . 3)

2

CONTENT FIGURES .............................................................................................................. 8 TABLES .............................................................................................................. 12 ABBREVATIONS................................................................................................ 14 ABBREVATIONS................................................................................................ 14 SYMBOLS .......................................................................................................... 15 THANKS ............................................................................................................. 17 DEDICATION ...................................................................................................... 18 1. INTRODUCTION: in search of asymmetry................................................... 19 2. Bibliographical review.................................................................................. 24

2.1. Production of enantiomers ................................................................... 24 2.1.1. Enantiomers and stereoselective synthesis-importance of the

catalyst. 24 2.1.2. Biocatalysis ..................................................................................... 27

2.1.2.1. Generalities on enzymes.......................................................... 27 2.1.2.2. Biocatalysis in industry ............................................................. 29 2.1.2.3. Membranes .............................................................................. 30 2.1.2.4. Biphasic systems...................................................................... 31 2.1.2.5. Improving the stability of enzymes: immobilisation techniques . 32 2.1.2.6. Immobilisation of ADHs. ........................................................... 34 2.1.2.7. Enzyme-catalysed reactions in non aqueous solvents.............. 35

2.1.3. Resolution of racemic mixture. ........................................................ 37 2.1.3.1. Chiral chromatographic separation........................................... 37 2.1.3.2. Resolution by selective crystallisation: conglomerates and

racemates. 39 2.1.3.3. Resolution by formation of diastereoisomers............................ 40 2.1.3.4. Resolution of enantiomer by formation of diastereoisomeric salt.

42 2.1.3.5. Other methods of resolution of alcohols ................................... 48

3

2.1.3.6. Resolution of enantiomers by formation of a diastereoisomeric

cocrystal with tartaric acid instead of a salt.......................................................... 49 2.2. Conversion of acetophenone to R-1-phenylethanol using alcohol

dehydrogenase from Lactobacillus brevis (LBADH) .................................................... 52 2.2.1. Alcohol dehydrogenases require a coenzyme NADH and NADPH that

must be regenerated. .............................................................................................. 52 2.2.1.1. Generalities about the coenzymes ........................................... 52 2.2.1.2. Regeneration by a second enzyme .......................................... 55 2.2.1.3. Regeneration by the same enzyme: sacrificial substrate method.

56 2.2.2. Different studies with alcohol dehydrogenase from Lactobacillus

brevis 57 2.2.3. Alcohol dehydrogenase in non aqueous solvent.............................. 59 2.2.4. Goal of this work concerning LBADH in dense gases...................... 61

2.3. Resolution via the formation of diastereomeric complexes with (+)-

tartaric acid followed by extraction with supercritical carbon dioxide applied (±)-trans-

1,2-cyclohexanediol .................................................................................................... 61 2.3.1. Different methods of resolution of (±)-trans-1,2-cyclohexanediol

based on the formation of a covalent bond.............................................................. 62 2.3.2. Resolution of (±)-trans-1,2-cyclohexanediol by selective formation of

a cocrystal 62 2.3.3. Physical properties of CHD. ............................................................ 64 2.3.4. Method of resolution of CHD by formation of a cocrystal CHD-TA and

optimisation of the parameters of extraction: temperature and pressure.................. 66 2.3.5. Issues concerning the resolution of CHD by cocrystallisation and SFE

to be addressed in the present work........................................................................ 68 3. Conversion of acetophenone to R-1-phenylethanol ..................................... 70

3.1. Materials and methods......................................................................... 70 3.1.1. Materials ......................................................................................... 70

3.1.1.1. Reagent ................................................................................... 70 3.1.1.2. Biocatalyst................................................................................ 70

3.1.2. Methods .......................................................................................... 70

4

3.1.2.1. Preparation of the biocatalyst ................................................... 70 3.1.2.2. High-pressure view cell ............................................................ 71 3.1.2.3. Reaction with LBADH............................................................... 72 3.1.2.4. ADH activity test....................................................................... 73 3.1.2.5. Autoclave for incubation of biocatalyst...................................... 74

3.1.3. Analytical methods.......................................................................... 74 3.2. Results and discussion ........................................................................ 75

3.2.1. Reaction in water ............................................................................ 75 3.2.2. Preliminary test in heptane.............................................................. 75 3.2.3. Reaction at high-pressure ............................................................... 76

3.2.3.1. Reaction in propane with co-immobilised catalyst..................... 77 3.2.3.2. Reaction in biphasic system propane-water ............................. 78

3.2.4. Deactivation of LBADH.................................................................... 80 3.2.4.1. Deactivation of “untreated” LBADH .......................................... 80 3.2.4.2. Deactivation of LBADH in propane ........................................... 81 3.2.4.3. Reaction in biphasic systems propane-water............................ 81

3.3. Conclusion and future work.................................................................. 82 4. Resolution of (±)-trans-1,2-cyclohexanediol via the formation of

diastereomeric complexes with (+)-tartaric acid followed by extraction with supercritical

carbon dioxide. ............................................................................................................... 83 4.1. Materials and methods......................................................................... 83

4.1.1. Materials. ........................................................................................ 83 4.1.2. Determination of the structure of the co-crystal. .............................. 83 4.1.3. Supercritical fluid extractor. ............................................................. 83 4.1.4. Resolution of CHD with TA and SFE............................................... 85

4.1.4.1. Sample preparation. ................................................................. 85 4.1.4.2. Supercritical fluid extraction...................................................... 85 4.1.4.3. Raffination by alkaline treatment .............................................. 86

4.1.5. Analytical methods.......................................................................... 86 4.2. Results and discussion ........................................................................ 87

4.2.1. Characterisation of the cocrystal ..................................................... 87 4.2.1.1. The structure of the co-crystal. ................................................. 87

5

4.2.1.2. Characterisation of the co-crystal TA-RRCHD.......................... 91 4.2.2. Decomposition of the CoC in situ .................................................... 93 4.2.3. Description of the extraction............................................................ 96

4.2.3.1. Monitoring the evolution of the content of the extractor by XRD

and fractionning 96 4.2.3.2. Improving the enantiomeric excesses by leaving off an

intermediate fraction...........................................................................................102 4.2.4. Sample preparation........................................................................102

4.2.4.1. Investigation of the binaries RRCHD-TA and SSCHD-TA.

Coroboration by XRD. ........................................................................................102 4.2.4.2. Investigation of the cocrystallisation by XRD, ternary phase

diagram. 106 4.2.4.3. Two issues raised by the XRD studies: sodium hydrogen tartrate

(NaTA) and metastable compound.....................................................................109 4.2.4.4. Conditions of crystallisation/sample preparation......................110

4.2.5. Toward enantiopure products, molar ratio, double extraction .........113 4.2.5.1. Extraction with molar ratios varying .........................................113 4.2.5.2. Resolution repeated twice .......................................................115

4.3. Conclusion on the resolution of CHD by cocrystallisation and SFE and

further plan 118 5. CONCLUSIONS .........................................................................................120 6. LITTERATURE...........................................................................................121 7. APPENDIX .................................................................................................132

7.1. Challenge of green chemistry..............................................................132 7.1.1. Context of the development of green chemistry..............................132 7.1.2. 12 principle of green chemistry. Derived conceps. .........................134 7.1.3. Alternative solvent: supercritical fluids and SCCO2.........................136

7.2. Reaction run with coimmobilised NADP ad LBADH in non-aqueous

solvent, propane and heptane....................................................................................139 7.3. Miscibility of ACP, ISP, 1-phenylethanol, acetone and propane ..........140 7.4. Method of determination of the structure of the cocrystal CoC ............140 7.5. Theoretical ternary diagram with a liquid solution................................141

6

7.5.1. If no liquid solution between the CHD enantiomers exists. .............142 7.5.2. A liquid solution exists between the CHD enantiomers...................143

7.5.2.1. Gibbs free enthalpy of a binary mixture RacCHD and RRCHD or

SSCHD presenting a partial miscibility ...............................................................143 7.5.2.2. Ternary diagram......................................................................143

7.6. Results of the experiment for the determination of the phase diagram.

146 7.7. Diffractogram of samples prepared according to different methods.....152

7

"Perhaps looking-glass milk isn't good to drink" Said Alice to the cat.

(Lewis Caroll, Though the looking-glass)

8

FIGURES

Figure 1: Structure of RPE, SSCHD, and RRCHD, the asymmetric compound

targeted in this work........................................................................................................ 21 Figure 2: The first homogeneous asymmetric catalysis by a chiral metal complex.

....................................................................................................................................... 24 Figure 3: Enantioselective synthesis of (-)-mentol................................................ 25 Figure 4: Bifunctional Ru-BINAP(diamine) complexes for enantioselective

hydrogenation of simple ketones (acetophenone into R-1-phenylethanol). ..................... 26 Figure 5: Illustration of how fast the enzyme drives a reaction............................. 28 Figure 6: Emil Fischer’s substrates...................................................................... 29 Figure 7: Classification of membrane filtration processes .................................... 30 Figure 8: Example of a continous conversion with a membrane .......................... 31 Figure 9 : A two-phased system involving a coenzyme-dependent enzyme ........ 32 Figure 10: L-amino acid production catalyzed by aminoacylase. ......................... 33 Figure 11: Three examples of carrier-coupling using the amino group of an enzyme

(Enz in this figure)........................................................................................................... 34 Figure 12: Binary mixture melting point diagram for a conglomerate-forming pair of

enantiomers (a) and a racemate-forming (b)................................................................... 39 Figure 13: Chemical and physical equilibria in the racemization for N-(2-

methylbenzylidene)phenylglycine amide......................................................................... 40 Figure 14: The resolution of rac-CHD by Chatterjee. ........................................... 41 Figure 15: Separation of enantiomers (+)-A and (-)-A combined with racemisation.

....................................................................................................................................... 42 Figure 16: Common resolution of a base B by the acid (L)-HA. ........................... 44 Figure 17 : a) Binary melting point phase diagram of salt p and salp n and b) their

solubility diagram. ........................................................................................................... 45 Figure 18 : Resolution with a molar ratio of 0.5.................................................... 46 Figure 19: Resolution of amphetamine by distillation ........................................... 47 Figure 20: Formation of phtalate derivatives, useful intermediates for resolution of

alcohol by diastereomeric salt formation ......................................................................... 48

9

Figure 21: Resolution of chiral alcohol by enzyme catalysed acylatation ............. 49 Figure 22: General activity of an alcohol dehydrogenase (ADH).......................... 53 Figure 23 : NADP+: Nicotinamide adenine dinucleotide ....................................... 53 Figure 24 : NAD is a hydrid acceptor and NADH a hydrid donnor........................ 53 Figure 25 : Coenzyme regeneration in the case of a reduction with NAD(P)H. .... 55 Figure 26 : Reactions catalysed by LBADH, conversion of acetophenone and

regeneration of NADPH by addition of isopropanol. ........................................................ 57 Figure 27: Comparison of 2 solid-liquid melting phase diagrams a) a usual for a

couple of enantiomers and b) CHD’s case with a solid solution....................................... 65 Figure 28: Principle of the resolution of (±)-CHD by co-crystal formation followed

by an extraction in SCCO2 (mr=0.5)................................................................................ 66 Figure 29: Decomposition of the residuum........................................................... 67 Figure 30 : Scheme and picture of the high-pressure reactor. ............................. 72 Figure 31: Example of a test of enzyme activity................................................... 74 Figure 32: Autoclave for the measurement of ADH deactivation in propane ........ 74 Figure 33: Conversion of ACP to RPE in water.................................................... 75 Figure 34: Reaction in heptane with co-immobilised catalyst (reaction G) ........... 76 Figure 35: Bioconversion in propane with immobilised catalyst............................ 78 Figure 36: Bioconversion in the biphasic system water/dense propane at 100 bar

....................................................................................................................................... 78 Figure 37: Bioconversion in the biphasic system water/dense propane at 30 bar 79 Figure 38: Bioconversion in the biphasic system water/dense propane at 200 bar

....................................................................................................................................... 79 Figure 39: Deactivation of an aqueous solution of LBADH at atmospheric pressure

and 36°C......................................................................................................................... 80 Figure 40: Deactivation of LBADH in powder form at atmospheric conditions...... 80 Figure 41: Deactivation of the preparations of LBADH in propane at 30 bar. ....... 81 Figure 42: Deactivation of an aqueous solution of LBADH in a biphasic system with

dense propane at 30 bar................................................................................................. 82 Figure 43: Supercritical fluid extractor.................................................................. 84 Figure 44: ORTEP diagram (Spek 2003) of the CHD-TA co-crystal (1)................ 87

10

Figure 45: The two dimensional infinite hydrogen bonded plane of the CHD-TA co-

crystal (1)........................................................................................................................ 90 Figure 46 : The inner TA layer of the sheet (Macrae et al. 2006) presenting its

hydrogen bonding system of co-crystal 1. ....................................................................... 91 Figure 47: Theoretical (in black) and experimental (in red) diffraction pattern of

CoC ................................................................................................................................ 91 Figure 48: FTIR: spectrum of the CHD-TA co-crystal Coc. .................................. 92 Figure 49: DSC melting peak of the pure CHD-TA co-crystal CoC in sealed Al-pan

at 10°C/min (mass 2.59 mg). .......................................................................................... 92 Figure 50: Simultaneous TG/DTA curves of the pure CHD-TA co-crystal 1 ......... 93 Figure 51: Extraction curves at different temperature and pressure, study of the

decomposition of CoC in situ. ......................................................................................... 94 Figure 52: Experimental powder pattern of the compound involved in the resolution

system. ........................................................................................................................... 97 Figure 53 : Loss of weight of the extractor according to the weight of CO2 and

sampling. The spline line is only indicative...................................................................... 97 Figure 54: Theoretical loss of weight of the extractor if no sample had been taken.

This figure does not show more information than the previous but has the advantage to

show which aspect the extraction curve would have if no sample had been taken.......... 98 Figure 55: Diffractograms of the different samples from the material inside the

extractor over the extraction............................................................................................ 99 Figure 56: Different fractions during extraction, their enantiomeric excesses......101 Figure 57: DSC curve of binary mixture corresponding to 1:1 molar ratio of a)

(R,R)-CHD and co-crystal (1) and b) (S,S)-CHD and (R,R)-TA, both exhibiting eutectic

melting behavior. ...........................................................................................................104 Figure 58: Melting binary phase diagram SSCHD TA. ........................................104 Figure 59: The problematic melting point phase diagram of RRCHD-TA. ...........105 Figure 60: Two ternary phase diagrams TA-SSCHD-RRCHD ............................107 Figure 61: DSC analysis of a sample of composition RRCHD:TA 16:84 featuring

the metastable compound “X” ........................................................................................110 Figure 62: Diffractograms of sample of mr=0.5 evaporated at different temperature

......................................................................................................................................112

11

Figure 63: Yield and F parameter with varying molar ratio..................................114 Figure 64: Enantiomeic excess with varying molar ratio......................................114 Figure 65: Second resolution of mixture 1 presenting an ee of SSCHD..............116 Figure 66: Second resolution of mixture 2 presenting an enantiomeric excess of

RRCHD..........................................................................................................................116 Figure 67: Sustainable development as a confluence of three domains: social,

economy, and environment............................................................................................132 Figure 68: Two determinant catalytic steps in the “green” synthesis of ibuprofen.

......................................................................................................................................136 Figure 69: phase diagramm (P,T) of a fluid.........................................................137 Figure 70: “No solid solution CoC>racem” phase diagram of RRCHD, SSCHD and

TA supposing that a racemic compound is formed and not a solid solution....................142 Figure 71 : “No solid solution CoC<racem” alternative phase diagram with RRCHD,

SSCHD, CoC, RacCHD, and TA....................................................................................142 Figure 72 : Gibbs molar enthalpy of a binary RRCHD-SSCHD ...........................143 Figure 73: “Solid solution and CoC>solCHD” ternary phase diagram .................144 Figure 74 : Variation of free energy of the system CHD + dn TA when CoC forms.

......................................................................................................................................145 Figure 75 : “Solid solution and CoC=solCHDlim” ternary phase diagram with

solCHD 2 .......................................................................................................................145 Figure 76: Example of ternary phase diagram where CoC can intake a small

fraction of SSCHD..........................................................................................................146

12

TABLES Table 1: The opposite enantiomers have different biological activities. ................ 22 Table 2: Examples of pharmaceuticals resolved by diastereomeric crystallisation in

the process ..................................................................................................................... 43 Table 3 : Coenzyme and their associated group.................................................. 52 Table 4 : Price of some nicotine adenine dinucleotides (from Jülich Chiral Solution

GmbH’s product portfolio (february 2007)) ...................................................................... 54 Table 5 : Some properties of Formate Dehydrogenase (FDH) (Product Portfolio

2007) .............................................................................................................................. 56 Table 6: Melting point and structural data for crystalline phases of trans-1,2-

cyclohexanediol and references...................................................................................... 64 Table 7: Comparasion of different results obtained for the resolution of CHD...... 68 Table 8: Result of the three conversions run in biphasic systems. ....................... 80 Table 9: Summary of crystallographic data, data collections, structure

determination and refinement for CHD-TA co-crystal (1)................................................. 88 Table 10: Intermolecular interactions in the crystal structure of CHD-TA co-crystal

(1). .................................................................................................................................. 89 Table 11: Melting point and enthalpy of fusion of the applied chemicals .............103 Table 12: DSC and XRD data of binary mixtures in the ternary system ..............103 Table 13: Result of the different experiments of further enantioenrichment of

mixture 1 and mixture 2 .................................................................................................117 Table 14: Waste of the different segment of chemical industry ...........................135 Table 15: Critical points of fluid presenting an industrial interest.........................137 Table 16: Result of the bioconversion of ACP into RPE in heptane and propane.

......................................................................................................................................139 Table 17: Miscibility in propane...........................................................................140 Table 18: The different sample prepared for investigation of the ternary system,

their composition and the phase observed by XRD........................................................146 Table 19: Composition of the different sample if the ternary does not present solid

solution. .........................................................................................................................147

13

Table 20: Which deviation do we observe from the model "no solid solution

CoC>racem"? ................................................................................................................147 Table 21: Composition of the different sample if the ternary presents a solid

solution and the deviations observed. ............................................................................147 Table 22: Sample for sample preparation prepared in different condition ...........152

14

ABBREVATIONS ACP Acetophenone

ADH Alcohol dehydrogenase

CHD Trans-1,2-cyclohexanediol

CoC cocrystal CHD-TA 1:1

DBTA O,O′-dibenzoyl-(2R,3R)-tartaric acid

DSC differential scanning calorimetry

ee enantiomeric excess

FDA Food and drug administration

FDH Formate Dehydrogenase

GC Gas chromatography

HPLC High performance (or pressure) liquid chromatography

HTA Sodium hydrogen tartrate

ISP isopropanol

LBADH Alcohol dehydrogenase from Lactobacillus brevis

NAD Nicotine adenine dinucleotide

NADH Nicotine adenine dinucleotide hydrogene

NADP Nicotine adenine dinucleotide phosphate

NADPH Nicotine adenine dinucleotide phosphate hydrogene

NaTA Hydrogen tartrato sodium

racCHD Racemic (±)-trans-1,2-cyclohexanediol

RPE R-1-phenylethanol

RRCHD (R,R)-trans-1,2-cyclohexanediol

RT Room temperature

SCCO2 Supercritical carbon dioxide

SF Supercritical fluid

SFE Supercritical fluid extraction

SolCHD Solid solution of racCHD with SSCHD or RRCHD

SSCHD (S,S)-trans-1,2-cyclohexanediol

SSTA (S,S)-tartaric acid

TA (R,R)-tartaric acid

15

SYMBOLS a activity no unit

c concentration (mol/L)

CO2rel relative weight of CO2 (mCO2/mCHDini) (g/g)

ee enantiomeric excess (%)

eeext or ee1 enantiomeric excess of the extract (%)

(first extraction)

eeext2 or ee2 enantiomeric excess of the extract (%)

(second extraction)

eeraf enantiomeric excess of the raffinate (%)

Ea activation energy (kJ/mol)

F F parameter no unit

g Gibbs free molar enthalpy (kJ/mol)

G Gibbs free enthalpy (kJ)

H enthalpy (kJ/mol)

Keq equilibrium constant no unit

µ chemical potential (kJ/mol)

µ0 chemical potential of a pure compound (kJ/mol)

P pressure (MPa, bar)

Pc critical pressure (MPa)

t time (h, min)

T temperature (°C, K)

Teu Eutectic temperature (°C, K)

Tc critical temperature (°C, K)

v reaction rate (mmol/min,

mmol/(min.U),…)

vi initial reaction rate (mmol/min,

mmol/(min.U),…)

R ideal gas constant (kJ/(mol K))

S entropy (kJ/(mol K))

V volume (L)

16

x fraction (%)

X fraction in a binary (%)

Y yield (%)

Yext or Y1 yield of the first extraction (%)

Y2 yield of the second extraction (%)

Yraf yield of raffination (%)

17

THANKS

I would like to warmly thank the people who made this work possible.

Friends and family. Colleagues and administrations.

As a complement:

“Hvala lepa”,

“Köszönöm szepa”,

Merci bcp,

“danke schön”,

“muchas gracias”!

18

DEDICATION

This work is dedicated to the memory of RENE HUBERT (1923-2010)

Introduction

19

1. INTRODUCTION: in search of asymmetry

Enantiomer comes from the Greek ἐνάντιος, opposite, and μέρος, part. Hands, feet are

enantiomeric pairs, they are mirror image to each other. A hand, a foot, snails are enantiomers for

an enantiomer is a molecule that does not superimpose on its mirror-image. Examples of synthetic

enantiomer and their biological activities are given in Table 1, where each line’s compounds are

mirror images. Whereas industrial production of enantiomerically pure products is a thorny

problem which has to be (partially) addressed herein, nature is remarkably able to perfectly

perform such a task with the helps of enzymes, the catalysts of life. The origin of the asymmetric

chirality in life is much discussed but we can more surely state that the need and the origin of

synthetic asymmetric molecules is to find in biological processes!

Biological organisms are built out of asymmetric compounds and the physiological

phenomena arise from highly precise molecular interactions in which chiral host molecules

recognize two enantiomeric guest molecules in different ways. Consequently many compounds

which are “active” are also chiral, and its stereoisomer often present dramatically different

activities toward a living organism as illustrated in Table 1 which includes the examples of the

present introduction. Industry found the source of chirality in natural products and still intensively

uses the building blocks produced in vivo. So many natural substances had applications in the fields

of chirotechnologies (Sheldon 1993): amino acid and sugar, which are eminently asymmetric, also

the biocatalyst and the secondary metabolites (alkaloids, essential oils…). Semisynthesis was

applied to create novel drug from traditional cures. A good illustration is the drug family derived

from opium poppy’s alkaloids. The catalytic potency of enzymes is taken advantage of by the

industry in processes based on biocatalysts or biotransformation as for antibiotics. Natural products

are the most prominent separation agents (for instance sugars for chromatographic columns) or

resolving agents.

The sectors of chemistry which needs asymmetric molecule are: pharmaceuticals, animals

heath products, agrochemical, electronic chemical, pheromones, flavours and fragrances. The scale

of this production spans from chiral synthons or highly active material produced at kilogram-scale

to amino acid which production reaches 105 tonnes per year. Agrochemical and pharmaceutical

commonly reaches tens of thousand of tonnes a year (Collins et al. 1992). The demand for

enantiomerically pure drugs has increased, so did the demand for stereospecific processes. Between

1983 and 1985, 95% of the drugs with an asymmetric carbon were sold under the racemic form,

while in 1992, only 25% of the drugs were sold as racemate, a equimolar mixture of both

Introduction

20

enantiomers (Collins et al. 1992) and the sale of enantiomer passed over the 120 billion marks in

the year 2000 (Kennedy et al. 2002). In 1998, 48% of the small molecule drugs approved by the

FDA were single enantiomers, and in less than half of those cases the chiral motif was made by

synthetic chemical methods. By 2007, the proportion of single enantiomers in small molecule drug

approvals had risen to 71%, and of these products, 70% had the chirality introduced by synthetic

chemical methodology. (Thayer 2008; Lennon et al. 2009)

Pairs of enantiomer generally have pharmaceutical activities (or efficacy) and only one

might be needed as for paclobutrazol (see Table 1: The opposite enantiomers have different

biological activities.) A similar issue is raised by essential oils as carvone or menthol or the

sweetener which taste or smell is rather different from one enantiomer to another: their

enantiopurification is compulsory. In some cases one of the enantiomers is not active and presents

no side effect and consequently the drug is often marked as racemic. This results in a poor yield:

half of the product is lost, as for ibuprofen, and consequently, when the production of a racemic is

shifted to a pure enantiomer the capacity of the process doubles, and the cost efficiency is

improved. Moreover the pharmacological study and validation of a drug is less complex for an

enantiomerically pure and the probability to have complex interaction between the active molecules

is reduced. Using the single active enantiomer can lower side effect of a drug, the same way it can

limit the environmental impact of agrochemicals.

Another reason for the development of chirotechnologies concerns the ownership of the

drugs. The development of a new drugs is extremely expensive : it costs about US$ 400 million

(2000 dollars) (Dimasi et al. 2003). When the patent of a drug sold as a racemate expires it is

possible to patent the sole active enantiomer to prolong the ownership on a molecule. The cost of

such an operation is obviously much lower than the discovery of a new molecule (Kennedy et al.

2002). For instance, the recently patented Xyzal (levocetirizine) is the active R-enantiomer of

cetirizine (Zyrtec, the racemate). Fluoxetine (prozac)) was subject of a dispute between two

companies about the commercialisation of fluoxetine as racemic or pure enantiomer (Kennedy et

al. 2002).

Thalidomide was a treatment for morning sickness. It possesses two enantiomers one

causes the desired sedative effect while the other was teratogenic and caused many foetal

disformation1 (Kennedy et al. 2002). Since this scandal the FDA and European Committee for

Proprietary Medicinal Products imposes that the activity of both enantiomers are known separately.

This decision increased a lot the cost of the development of a drug, therefore, pharmaceutical

1 Thalidomide racemises in physiological conditions. However they are used against leprosy and

might be again used against Aids disorders and tuberculosis.

Introduction

21

companies might decide to develop (and patent) an enantiomerically pure product. Hence, the other

enantiomer got the status of an undesired impurity.

On one hand, the authorities and the industrials’ interest for chirotechnologies has

motivated academic research in chemistry to study this relatively new topic. On the other hand, the

improvement of the legislation and products on the market took place because analytical

procedures allowed the determination of configurations or enantiomeric excesses, natural product

chemistry and biochemistry improved and stereoselective preparative methods were developed.

In this context much favourable to the development of chirotechnologies, the study of two

different methods of obtaining chiral alcohols is proposed herein. The requirement of the relatively

new paradigm of green chemistry associated with clean technologies such as biocatalysis or dense

gases as non-conventional solvents will be focused at. Indeed, the two proposed method of

production of chiral alcohol (see their structure on Figure 1) is the conversion of acetophenone into

(R)-1-phenylethanol in dense gases catalysed by Lactobacillus brevis alcohol dehydrogenase and

the resolution of (±)-trans-1,2-cyclohexanediol by cocrystal formation followed by supercritical

extraction.

OH

RPE

(R)-1-phenylethanol

OH

OH

SSCHD (S,S)- trans-1,2-cyclohexanediol

OH

OH RRCHD (R,R)- trans-1,2-cyclohexanediol

Figure 1: Structure of RPE, SSCHD, and RRCHD, the asymmetric compound targeted in

this work.

Introduction

22

Table 1: The opposite enantiomers have different biological activities.

H2NCOCH2

NH2 H

COOH

(S)-asparagine has a bitter taste.

H2NCOCH2

NH2 H

COOH

(R)-asparagine has a sweet taste.

O

(S)-carvone: caraway flavour

O

(R)-carvone: spearmint flavour

OH

O2N

OH

NHCOCHCl2

(R,R)-chloramphenicol is antibacterial.

OH

O2N

OH

NHCOCHCl2

(S,S)-chloramphenicol is inactive.

N NO

OH

O

Cl (R)-Levocetirizine is an antihistaminic.

N NO

OH

O

Cl S-Levocetirizine is not active.

ONH H

CF3

(R)-fluoxetine is antidepressant with

cardiac side-effect.

ONH H

CF3

(S)-fluoxetine is antidepressant with lower

side-effect.

NHO

ON

O

O (S)-thalidomide is teratogenic.

NHO

ON

O

O (R)-thalidomide has a sedative activity.

Used against leprosy

Introduction

23

O

OH

(S)-ibuprofen is an anti-inflammatory drug.

O

OH

(R)-ibuprofen is inactive.

NH

NH

OH

Et

H

OH

H Et (S,S)-ethambutol is tuberculostatic.

NH

NH

OH

Et

H

OH

H Et (R,R)-ethambutol causes blindness.

NN

N

tBu

Cl

OH

(R,R)-paclobutrazol is a fungicide.

NN

N

tBu

Cl

OH

(S,S)-paclobutrazol is a plant growth

regulator.

NO OF3C

COOBu

MeH

(S)-fluazifop butyl is inactive.

NO OF3C

COOBu

MeH

(S)-fluazifop butyl is an herbicide.

Bibliographical review

24

2. Bibliographical review

2.1. Production of enantiomers

2.1.1. Enantiomers and stereoselective synthesis-importance

of the catalyst.

Enantiomeric excess (ee) is a common way to describe the composition in enantiomers of a

mixture. Its formula is given in Equation 1. A common industrial standard for ee is above 70-80 %

(Collins et al. 1992). But higher grade can be demanded for certain purposes.

SRSR

ee

Equation 1 : Enantiomeric excess

The development of stereoselective chiral catalyst was the object of much effort and recent

development of not only biocatalysts (Sheldon 1993) but also abiological catalysts that present the

advantage to avoid the substrate limitation imposed by enzymes2. The catalytic asymmetric

synthesis is mainly based on chiral transition metal complexes or chiral acid and base. An example

is given in the Figure 2. The catalyst was improved by screening of Schiff base, ligand for the

metallo complex, leading to enantioselectivity up to 94%.

Figure 2: The first

homogeneous

asymmetric catalysis

by a chiral metal

complex.

Reproduced from Chirotechnology: the Industrial Synthesis of Optically Active

Compounds (Sheldon 1993)

An industrially-relevant synthesis based on stereoselective catalyst is the synthesis of (-)-

mentol that was developed by Noyori who was then rewarded with Nobel Prize in chemistry. The

2 For the importance of the catalyst in green chemistry and an overall introduction to this topic, see

annex 7.1.


25

key step was the asymmetric isomerisation of geranyldiethylamine catalyzed by an (S)-BINAP–Rh

complex in THF forming (R)-citronellal enamine, which upon hydrolysis gives (R)-citronellal in

96–99% ee (see Figure 3). The asymmetric reaction is performed on a nine-ton scale (Noyori

2001).

Figure 3: Enantioselective synthesis of (-)-mentol

Reproduced from (Noyori 2001)

Concerning the production of chiral alcohols green chemistry has presented different tools.

There are several techniques based on the stereoselective reduction, hydrogenation of ketones.

Several homogenous catalysts can be used for this purpose. The most effective homogeneous

hydrogenation catalysts are complexes consisting of a central metal ion, one or more (chiral)

ligands and anions which are able to activate molecular hydrogen and to add the two H atoms to an

acceptor substrate (Sheldon et al. 2007). Ru, Rh and Ir complexes stabilized by tertiary (chiral)

phosphorus ligands are the most active and the most versatile catalysts. The halogen-containing

BINAP–Ru(II) complexes is a precious catalyst for this reaction and allow ee superior to 90%

(Noyori 2001). Noyori rendered possible the enantioselective hydrogenation of aromatic ketones by

introducing a new class of Ru-BINAP (diamine) complexes (Noyori et al. 2001). In this case

hydrogen transfer is facilitated by ligand assistance. The company Takasago used this catalyst for


26

the production of (R)-1-phenylethanol in 99% ee using only 4 bar of hydrogen (see Figure 4), a

very moderate pressure compared to the 100 bar commonly used. There is consensus that the

transfer of the two H atoms occurs in a concerted manner as depicted in Figure 4 (Sheldon et al.

2007). This hypothesis explains the need for an N–H moiety in the ligand.

Figure 4: Bifunctional

Ru-BINAP(diamine)

complexes for

enantioselective

hydrogenation of

simple ketones

(acetophenone into

R-1-phenylethanol).

Reproduced from

Green Chemistry and

Catalysis (Sheldon et

al. 2007)

More rarely, heterogenous catalysts can catalyse the stereoselective reduction of ketone.

They are metal with chiral natural modifier such as Raney nickel system for β-functionalized

ketones (with tartaric acid (TA) as a modifier) or Pt catalysts modified with cinchona alkaloids for

α-functionalized ketones. In this last case, acetic acid or toluene as solvent, close to ambient

temperature and medium to high-pressure (10–70 bar) are sufficient to ensure high

enantioselectivities of 95 to 97.5% (Sheldon et al. 2007).

Some chiral catalysts were given as examples. Before concentration our attention to

biocatalyst and more specifically alcohol dehydrogenase it is necessary to give some definitions

necessary for the evaluation of the catalytic capacity. The activity of a catalyst is its ability to

perform a typical reaction in known condition: the activity is defined according to a certain

protocol that is widely accepted, shared by the researchers keen on comparing their results. The

activity is the initial rate of this reaction expressed in and µmol.min-1 and µmol.min-1/mgCatalyst is

often abbreviated as U and U/mgCatalyst, respectively. The activity is a very important parameter that

accounts for the deactivation (or hyper activation) of a catalyst. The half-life of catalyst is the time

after which half of the initial activity remains. The turnover is the number of catalytic cycles the


27

catalyst undertaken by unit of time. The total turnover is the number of catalytic cycle for a

conversion, this to say the number of synthesized molecules per number of catalyst molecules used.

2.1.2. Biocatalysis

2.1.2.1.Generalities on enzymes

The enzymes ARE the catalysts of life. These proteins are long chains from 100 to several

hundreds α-amino acids whose sequences are encoded in the DNA. This polypeptide chain forms

the primary structure of enzymes. Intramolecular bounds maintain the three-dimensional

structure of the protein. Secondary structure is the term given to local regions (10–20 amino

acids) of stable, ordered three-dimensional structures held together by hydrogen bonding, that is

non-covalent bonding between acidic hydrogens (O-H, N-H) and lone pairs. The three-dimensional

structure of protein sub-units, known as the tertiary structure, arises from packing together

elements of secondary structure to form a stable global conformation, which in the case of enzymes

is catalytically active. The packing of secondary structural units usually involves burying

hydrophobic amino acid side chains on the inside of the protein and positioning hydrophilic amino

acid side chains on the surface. The quaternary structure of an enzyme is the final structure of the

enzyme that can involve several proteins and cofactor or coenzymes (see 2.2.1.1). For more detail

please refer to Introduction to Enzyme and Coenzyme Chemistry (Bugg 2004).

Because of each enzyme’s millions year evolution, they are extremely good catalysts when

compared with man-made catalysts. They present three qualities that are speed, selectivity and

specificity. The speed of the enzyme in catalysing biochemical reaction can be illustrated by the

example found in Introduction to Enzyme and Coenzyme Chemistry (Bugg 2004): The rate of acid-

catalysed glycoside catalysis is accelerated 103-fold by intramolecular acid catalysis, but enzyme-

catalysed glycoside hydrolysis is 104-fold faster still – some 107 faster than the uncatalysed

reaction carried out at pH 1 (see Figure 5).


28

Figure 5: Illustration of how fast the enzyme drives a reaction

Reproduced from Introduction to Enzyme and Coenzyme Chemistry (Bugg 2004)

The selectivity and specificity of the enzyme is one of their best-known properties – and

also the most looked for. Actually, most enzymes have a limited range of accepted substrates and

the reaction they catalyse leads often to a unique product. Maybe the most striking of the extreme

stereoselectivity of the enzyme is the fact that in living organisms only one orientation is given to

the amino acid, the nucleic acid and so on. The discovery of this property of the enzyme was

underlined by Emil Fischer in 1894. He observed that the enzyme known as emulsin catalyzes the

hydrolysis of β-methyl-D-glucoside, while the enzyme known as maltase is active towards the α-

methyl-D-glucoside substrate (see Figure 6). This led Fischer to suggest his famous “lock-and-key”

theory of enzyme specificity, which he described in his own words as follows: “To use a picture, I

would say that enzyme and the glucoside must fit into each other like a lock and key, in order to

effect a chemical reaction on each other” (Vasic-Racki 2006)


29

Figure 6: Emil Fischer’s substrates

Reproduced from (Vasic-Racki 2006)

2.1.2.2.Biocatalysis in industry

They are important catalyst for industrial purposes, especially food industry. The

advantages of biocatalysis are the following: Enzymes (or micro-organisms) are renewable,

biodegradable and non toxic catalysts; they present good kinetic (high turnover) and high

selectivity (important enantiomeric excess and conversion and no protection step needed).

Moreover, they require only mild reaction conditions (low temperature, moderate pH and so on).

The History of Biostranformation – Dreams and Reality is told by Durda Vasic-Racki in a

thrilling way in Industrial Biotransformations (Liese et al. 2006) and we cannot report all

inventions and developments, and rather use some historical hallmarks as illustration and invite the

reader to this excellent historical introduction (Vasic-Racki 2006). The origin of biotransformation

can not be dated because the first microbial biotransformations, the production of alcoholic

brewage, cheese, vinegar were known before writing. As we need the restrain the scope of this

presentation we will refer only to enzyme-catalysed reaction, this is to say we exclude microbial

biotransformation and all those based on whole cell.

5 classes of enzyme are used for bioconversion: oxidoreductases (EC 1), transferases (EC

2), hydrolases (EC 3), Lyases (EC 4), isomerases (EC 5). Many oxidoreductases include the

NAD(P) dependent alcohol dehydrogenase, which will be considered latter. The enzyme class of

oxidoreductases possesses also enzyme which are not NAD(P)-dependent and whose use is

consequently easier - especially in non-aqueous solvent. Peroxidases and Polyphenol oxidase

consume hydrogen peroxide (H2O2) and O2, respectively. Desaturase are able to unsaturate an alkyl

chain (releasing H2) (Klibanov 2003; Liese et al. 2006). An example of liase that is industrially

relevant is the Penicillin amidase for manufacture of semi-synthetic β-lactam antibiotics. The


30

worldwide capacity is more than 20,000 t.a–1. Lipases are a relevant example of this class and will

be treated later for the reason of their prominence in non aqueous media. We would like to mention

some points on the different bioreactor setups which are relevant to our studies

2.1.2.3.Membranes

Filtration is the operation of separating solid from liquid. Membranes are filters which pore

size is so small that is can retain a fraction of the constituent of a mixture. They main application in

biocatalysis is to maintain an enzyme in solution but it exists also membranes that can keep the

coenzyme as well, as shown on Figure 7. It can also be functionalised to present a selectivitity

more specific than size as charge, polarity and so on.

Figure 7: Classification of membrane filtration processes

Reproduced from Industrial Biotransformations (Liese et al. 2006)

An example of continuously operated stirred tank reactor that uses enzyme membrane

reactor for the separation of the product from the enzyme 2-oxo-4-phenyl-butyric acid 2-hydroxy-

4-phenyl-butyric acid with D-lactate dehydrogenase and regeneration of coenzyme performed with

formate dehydrogenase is showed on Figure 8.


31

Figure 8: Example of a continous conversion with a membrane

Reproduced from Industrial Biotransformations (Liese et al. 2006)

2.1.2.4.Biphasic systems

If low solubility of substrate imposes a large reaction volume or if a substrate or product is

instable in water so that it is important to reduce the time it spends in the aqueous phase, a biphasic

system can be chosen instead of a single aqueous phase. Such two phase systems are depicted in

Figure 9. An example of a system that requires a coenzyme was chosen and includes apolar

products and substrates. The two phases are:

Aqueous phase. It contains the catalysts (the enzyme and the coenzyme if needed) and

hydrophilic substrates and products. The reaction takes place in this medium.

Organic water-immiscible phase. It “stores” the apolar/hydrophobic substrates and

products and it exchanges them with the other phase. Ideally, it provides the substrates to

the aqueous phase for the reaction to occur and extract the products.

Enzyme should be carefully chosen for this kind of system because the interface

water/organic solvent can deactivate them (Groger et al. 2003). A way to prevent the deactivation

of enzymes at this interface is to prevent them from reaching it with a membrane. If a membrane

had to be added on the reaction setup depicted in Figure 9 it would be in the aqueous phase close to

the organic phase so that the enzyme remains below the interface between the phases. Important

disadvantages of this technique are that it slows down the mass transfer and it increases pressure

drop.


32

Figure 9 : A two-phased system involving a coenzyme-dependent enzyme

The stereoselective hydrogenations of a ketone into an alcohol are often performed in such

media because the solubility of the product and substrate in water is usually low. (This is the case

of ACP and RPE.) It is also a simple way to separate the product from the catalysts. However

running reaction batch-wise and extract the product at the end is also common (Liese et al. 2006).

2.1.2.5.Improving the stability of enzymes: immobilisation

techniques

An important inconvenient of enzyme and biocatalyst in general is their instability that

represent an important cost. The catalyst is expensive and must be renewed often. Immobilisation is

a technique that improve the half-life of an enzyme. Another advantages is that the catalyst is easily

separable from the products. For more detail on immobilisation of enzyme please refer to Carrier-

bound Immobilized Enzymes (Cao 2005) where much information on this topic is found.

Many methods of enzyme immobilisation have been developed. The first is the simplest:

the enzyme is linked to the carrier by physical bounds. The advantages of this method are the

simplicity, the reversibility that allows the recyclig of the carrier and the fact that the enzyme does

not undergo much modification may lead to an unaltered activity. The main disadvantages come

from the weakness of the bound that often results in an important release of the enzymes into the

reaction media especially in presence of high desorption forces such as high ionic strength, pH and

so on. In organic solvent, where the enzyme is no soluble, physical absorption is a convenient


33

technique. The quantity of adsorbed enzyme is critical and a minimum monolayer coverage is often

required. Indeed protein molecules tends to maximize contact with the carrier surface by deforming

or unfolding, thus resulting in loss of activity, because of conformation changes, when coverage of

the carrier surface by the protein is below the monolayer. Additive as an inactive protein can fill the

surface on the carrier unoccupied by a too little quantity of enzyme. The absorbent should be

chosen with care because the interaction between the enzyme and the carrier will determine the

properties of the enzyme. The hydrophobicity/aquaphilicity balance of the support is decisive.

Hence, hydrophobic carriers is suitable for lipase immobilisation while some enzymes were more

favourably immobilised on more polar surface carrier able to form H-bond with enzyme or ionic

interaction. The first full scale industrial use of an immobilized enzyme was based on aminoacylase

immobilized on DEAE–Sephadex (weak ion exchanger) in a packed bed reactor. In 1969, they

started the industrial production of L-methionine as shown on Figure 10. This reaction is also an

example of resolution coupled with a racemisation. More examples will be given in 2.1.3

Figure 10: L-amino acid production catalyzed by aminoacylase.

Reproduced from History of Industrial Biotransformations (Vasic-Racki 2006)

Another technique of immobilisation is known as “entrapment”: a matrix is formed around

the enzyme with pores so fine that the enzyme can not leave. The matrix establishes multiple

physical bounds around the enzyme. Alginate gel formation is the more common technique of

entrapment. The water soluble alginate is mixed with the biocatalyst solution and dropped into a

calcium chloride solution in which water-insoluble alginate beads are formed. Carrageenan and

polyacrylamide gels are also widely used. Enzyme entrapment in sol-gel was applied to lipases

from Candida rugosa and porcine pancreas for reaction in SCCO2 and near critical propane. The

stability of the enzyme was improve and the reaction rate was much enhanced compared to reaction

with non-immobilised enzyme (Novak et al. 2003). Yeast ADH was immobilised by entrapment in

poly(AAmco-HEMA) gel (Soni et al. 2001)

The last method of immobilisation is the formation of a chemical bound between the carrier

and the enzyme. Several methods are possible and three are given as example in Figure 11. The

two first methods correspond to the coupling of the amino group of the enzyme, whereas in the last


34

case a carboxylic acid of the enzymes is activated with a diimide to be bounded to the carrier. It

should be emphasised that many other techniques are possible and many use a spacer, a longer

chain molecule that links the enzyme to the carrier. The enzyme is bounded to the carrier by

multiple bounds to the carrier and physical interaction takes place as for adsorption. The

advantages are numerous: the leakage of enzyme is slight, the conformation of the enzyme is

“frozen” on the carrier and this technique alters or improves the enzymes properties the more

radically.

OO

OH

NH2

NH2 Enz

NH2 Enz

EnzHOOC N NR

R

Enz O NR

O NHR

NH

Enz

O RHN NHR

O

O NH

OH

Enz

O NH

NH

EnzSupports SupportsCNBr

+

+

Figure 11: Three examples of carrier-coupling using the amino group of an enzyme (Enz

in this figure).

Another way to form a biocatalyst insoluble in water is the crosslinking technique: covalent

bound are formed between enzyme. The most common reagent is glutardialdehyde . This technique

was combined with precipitation: CLEA, cross-linked enzyme aggregates, are formed (Sheldon

2008).

2.1.2.6.Immobilisation of ADHs.

ADHs were seldom immobilised (Soni et al. 2001; Bolivar et al. 2006) for many processes

using them are run in a single phase (that is followed by an extraction), in biphasic systems or rely

on membranes (see 2.1.2). LBADH was immobilised by adsorption on glass bead for gas phase

reaction as it will be detailed in the next chapter. This method gave very good results for gas phase

reaction, allowing to lengthen its half-life considerably. LBADH (see 2.2 for more detail about this

enzyme.) was also immobilised for plug-flow reactor (Hildebrand et al. 2006) where the

regeneration is performed by isopropanol. This example is good example of how the different

technique can be combined: when a covalent bound is formed the enzyme establishes physical


35

bound to the carrier and the stability of the enzyme can be futher improved by crosslinking the

different enzyme’s residue left unreacted after immobilisation. In this example LBADH was

immobilised on amino-epoxy carrier supports prior to covalent binding to the epoxy groups of the

support, the protein physically adsorbed to the surface. Through additional amino groups on the

support, this adsorption process is facilitated and proceeds at low buffer concentrations. After the

adsorption process, bonds are formed between the epoxy groups and the nucleophilic groups of the

enzyme, resulting in a covalent multi-point attachment in which the enzyme conformation is more

rigid and therefore more stable against inactivation. At the end of this process of immobilization

the half-life of the enzyme was still the same as free in buffer, about 20 hours at 30°C. The half-life

of the preparation was increased over 500 h while 20 % of the initial enzyme activity remains by

proper immobilisation technique which combines the blocking of the remaining epoxy group of the

support with mercatoethanol and the cross-linking of enzyme with glutardialdehyde. A process run

with this catalyst achieved a conversion of 60% for enantiomerically pure RPE with TON of

2,500,000 and the catalyst was used over 10 weeks (Hildebrand et al. 2006).

2.1.2.7.Enzyme-catalysed reactions in non aqueous

solvents

The “natural” solvent for biocatalysis is water. However, it was discovered in the mid-

1980’s that enzymes are surprisingly active in organic solvents (Zaks et al. 1984; Zaks et al. 1985)

for only a small quantity of water is necessary to the maintaining of the ternary structure (as low as

0.02%, few water molecules for an enzyme molecule). Hydratation gives flexibility to the enzyme.

This flexibility is not only responsible for the enzyme to catalyse a reaction, the shape of the

enzyme adapting to the substrate, but also at the origin of the deactivation of the enzyme.

Deactivation of an enzyme corresponds to the unfolding of its chain. The deactivation can be of

two kinds: denaturation which is reversible and inactivation that is reversible (Fágáin 1995). The

water content of the non-aqueous media, or water activity, has a dramatic effect on the enzyme

stability and activity, water being exchanged between the surface of the solvent and the solvent.,

and must be optimised: too low water activity lead to too rigid enzymes laking activity and high

water activity can lead to fast deactivation. Too polar solvents are not suitable for enzyme-

catalysed reaction in non-aqueous solvents as they strip away water molecules from the surface of

the enzyme.

The advantages for the use of organic solvents instead of water in biocatalysis can be listed

(Vulfson et al. 2001; Klibanov 2003): the selectivity or activity can be modified and so can be more

beneficial to a desired reaction (Zaks et al. 1986), the slight solubility in water of certain

compounds, and/or stability in water limit the performance of the reaction in aqueous solvent.


36

Moreover, an extraction of the medium by an organic solvent, possibly contaminant and toxic

solvents like hexane, is necessary for recovering the products. The solubility of reactants is

increased, the non-enzymatic (spontaneous) and side reactions (products not stable in water)

eliminated. Enzyme stability can be enhanced, as well as the stability of coenzymes. But the

following problems may arise: many enzymes are quickly deactivated in organic solvents, the

reaction rate is low sometimes, solvents strongly inhibit certain enzymes, and aggregation of the

enzyme is a limit to mass transfer. Immobilisation generally considered as necessary in organic

solvent unless enzyme aggregate. A proper immobilisation improves the stability and/or activity of

biocatalysts.

Most reactions developed in non-aqueous solvents involve ester formation,

transesterification, hydrolysis. There are several reasons for this development. The enzyme

involved in those processes, as lipases, are found in vivo at the interface between phase, oil or fat,

and an aqueous phase, which explain their good stability and activity in non-aqueous solvents.

Indeed they are considered as “hard” enzymes: their structure is less flexible and, nonetheless,

active. This fact also explains why their immobilisation is particularly easy and efficient (Cao

2005). Lipases are well-known enzymes and have found many industrial applications such as large

scale used in detergent. Synthesis with lipase often involved apolar molecule which are not soluble

in water and so in this solvent the mass transfer of would slow down the reaction too much. In

absence of water the hydrolysis does not take place: in water many reactions are simply impossible

thermodynamically for the ester is hydrolysed. The case of ADH-catalysed reaction will be treated

in more details in 2.2.3.

Despite all those efforts to develop stereoselective catalysts, production of enantiomerically

pure molecule is often not feasible. Indeed, the result of most synthesis remains a racemic

compound of which only half is required. The separation of one enantiomer from the other is called

resolution. The main methods of resolution of enantiomers are presented in the next part. To the

point of view of the atom economy developed in a resolution3 is inferior to a neat stereoselective

synthesis for twice too many molecules are consumed and the atom efficiency is automatically

below 0.5 and decreases exponentially with the increase in the number of the asymmetric centres in

a molecule.

3 or, better, a resolution where one of the enantiomer is not desired or where the unwanted

enantiomer is recycled by racemisation. The case of the racemisation is important to the green chemisty’s

paradigm and would be consider in the coming part.


37

2.1.3. Resolution of racemic mixture.

This part will focus at the methods that enable the resolution of enantiomeric mixtures:

chromatographic method 2.1.3.1, separation of conglomerate 2.1.3.2, the resolution by formation of

diastereomers 2.1.3.3 (especially salts 2.1.3.4 and co-crystal 2.1.3.6). An example of a method

based on a biocatalyst is given in 2.1.3.5.

2.1.3.1.Chiral chromatographic separation

Chromatography is defined as a physical method of separation in which the components to

separate are distributed between two phases, one of which is stationary (stationary phase) while the

other (the mobile phase) moves in a definite direction. Separation of solutes injected into the

system arises from differential retention of the solutes by the stationary phase. The term liquid

chromatography is used when the stationary phase is a solid and the mobile phase a liquid. In the

case of gas chromatography the mobile phase is a gas while the stationary solid. Chromatographic

techniques, HPLC and GC, are routinely used as quantitative analytical method and now a wide

range of chiral columns allows the separation of the enantiomers. After a column with a high

loading capacity is successfully developped an analytical method can scaled-up.

Liquid chromatography and especially HPLC is the most important preparative method

used for the separation of enantiomers. The classical open-column liquid chromatography was

improved by the use of very small particles for the solid adsorbent stationary phase. Because of this

bed of packing material had much lower permeability, it became necessary to use a pump to

generate sufficient pressure to produce a flow rate high enough (Fekete 2008). The pressures used

for an analysis by HPLC are generally about 200 bar while the column can stand about 400 bar. In

the case of Ultra High Performance Liquid Chromatography the pressure are generally about 800

bar and the maximum pressure is about 1000 bar. Those techniques are usually gentle and

appropriate for unstable molecules. Although they have the reputation to be expensive and

ineffective for large scale separation, their quality has improved and chromatography is sometime a

first choice for resolution (Subramaniam 2001) and, according to this author, the separation are

rather easy to develop and appropriate for the small scale. The first HPLC column used for chiral

resolutions are based on polysaccharide, cellulose and starch, and the next developments were

based on substituted polysaccharides. Then, column came based on emulsion polymerisation of

acrylamides from amino acids. Fundamentally, three kinds of chiral stationary phases are

distinguished: chiral polymers, achiral matrices (mainly silica gel) modified with chiral moieties

(amino acid derivatives, crown ethers, cinchona alkaloids, carbohydrates, amines, tartaric acid

derivatives, cyclodextrins and binaphthol), and imprinted materials (chiral cavities) (Francotte

2005).


38

A first problem for enantiomeric resolution based on HPLC is the fact that separation is

performed batch-wise. Some methods allow a more intensive use of the equipment and the saving

of solvent: multiple close injection (the injection are performed repeatedly so that the support is

always involved in separation), recycling and peak shaving (the faction of solvent that contained

the overlapping of the two enantiomers is reinjected i.e. recycled). Another improvement is the

techniques of simulated moving beds which allows a continuous separation rather than proceeding

by batches (Rodrigues et al. 2001). The method of true moving bed is intuitive but hard to set:

liquid (equivalent to the mobile phase in HPLC) and a solid (absorbent equivalent to the stationary

phase) flow in opposite direction. The more retained molecule exits the separator with the solid

while the less goes out with the liquid. The process based on simulated moving beds are based on

this concept but there is no movement of the solid phase but the position of the inlet and outlet

steams move periodically simulating the movement of the solid absorbent (Rodrigues et al. 2001).

This technique was used by the pharmaceutical industry for the resolution of chiral drugs at large

scale production (>100g) including chiral buiding blocks as 1-phenylethanol (Negawa et al. 1992).

Despite its versatility, this technique represents a large investment prohibitive for small companies.

The inconvenient of chromatographic method is the high cost, notably in solvent that came

out very diluted out of the column and might contained additive such as a buffer, the matter related

to the scale-up, as the quality of the resolution might degrade at a larger scale. (Chiral) preparative

GC is rare (Subramaniam 2001). The column used was γ-cyclodextrine, cyclodextrine being the

most common type of (analytical) column for chiral GC. Chiral GC coupled with simulated moving

beds was applied to the resolution of (±)-enflurane, a volatile anesthetic. The use of membrane for

chiral resolution is at the moment limited, as well.

Chromatographic methods using supercritical fluids

Supercritical fluids got more and more important in the domain of chromatography

(Villeneuve et al. 2005) despite the investment that represents the equipment and the complexity

(or the lack of formation) regarding those techniques. The unique properties of supercritical fluids4

improves resolutions previously performed with HPLC (Phinney 2001): the higher diffusivity for

solute and the lower solubility compared to liquids leads to a higher efficiency and a shorter

analysis time. The pressure drop along the column which is a limit to scaling up liquid

chromatography processes is much lower using SF. The recovery of the product is easier and the

solvent consumption lowered. Considering all those advantages, the operation cost for preparative

SFC will be half of that of preparative HPLC (Villeneuve et al. 2005).It is to notice that the used of

4 The annexes contains a more general presentation of supercritical fluids (7.1.3).


39

SCCO2 is limited to apolar product, even if this difficulty might be overcome by using a modifier

such as methanol. SF chromatography was also applied to the preparation of drug (Fuchs et al.

1992; Perrut 1994). The columns used for SF chromatography are the same as for HPLC.

The readers who would like to know more about this field are advised to consult the book

Preparative Enantioselective Chromatography (Cox 2005) and Techniques in Preparative Chiral

Separation (Subramaniam 2001). Both have a section on SF chiral chromatography.

2.1.3.2.Resolution by selective crystallisation:

conglomerates and racemates.

a) b)

Figure 12: Binary mixture melting point diagram for a conglomerate-forming pair of

enantiomers (a) and a racemate-forming (b).

When a racemic mixture is liquid it forms a homogenous phase that is not optically active.

Solid enantiomers present two main behaviours: conglomerate or racemate. If they form

conglomerate they crystallise separately with different symmetry and this was exploited by Pasteur

for the resolution of tartaric acid (Pasteur 1922-1939). However, most enantiomers don’t form a

conglomerate but rather a racemate (Jacques et al. 1981): The racemate is a compound of the two

enantiomers in an equimolar ratio5 whereas in the case of conglomerate no such compound is stable

and the enantiomers crystallise separately as shown on Figure 12a). Most enantiomers are

racemate.

A conglomerate can be separated by differential crystallisation, the crystallisation of the

desired enantiomer is promoted by seeding (Jacques et al. 1981) as used for the preparation of α-

methyl-DOPA by Merk (Collins et al. 1992). A very interesting enantioseparation method

combines with racemisation gave enantiomerically pure product. It is a new concept and was

applied N-(2-Methylbenzylidene)phenylglycine Amide (Noorduin et al. 2008a; Noorduin et al.

5 except for the rare case of liquid solutions.


40

2008b) A saturated solution of this amine containing a fraction of solid crystal is ground until a

only one crystal is present Figure 13. This is among the convincing explanations for the origin of

asymmetry in life but, to the point of view of the chemical engineer, a process based on such

method has an impressive efficiency: virtually no waste is generated.

Figure 13: Chemical and

physical equilibria in the

racemization for N-(2-

methylbenzylidene)phenyl

glycine amide

1: N-(2-methylbenzylidene)phenylglycine amide, for this figure only. The enantiomers

have different colors, blue and red.

Reproduced from (Noorduin et al. 2008a)

Unlike congomerate-forming chiral molecule, racemate-forming enantiomers yields a

racemate only when crystallised and a non-equimolar mixture of both (ee≠0) can be purified only if

the enantiomeric excess is over the ee of the eutectic mixture: in this case the enantiomer

precipitates before the racemate (Jacques et al. 1981; Lorenz et al. 2006). This is why a resolution

of a racemate should yield, prior to recrystallisation, a mixture of racemate and enantiomer which

ee is higher than the eutectic’s (Wilen et al. 1977). However, the purification by recrystallisation is

a supplementary step that provokes the use of supplementary solvent, its evaporation if it is wanted

to recycle the molecule that did not crystallise. Consequently, a resolution should yield the

enantiomer with an ee as high as possible. If the ee is high enough no further refinement is needed.

If it is too low a recrystallisation is required and the higher the ee, the less material to recycle is

produced.

2.1.3.3.Resolution by formation of diastereoisomers

A common strategy for the separation of a racemate is the use of a resolving agent that is

asymmetric. A resolving agent is a compound that is added to the targeted product for its

resolution. It is generally removed from the product at the end of the separation and recycled for

another use and, to this point view, comparable to a protecting agent in organic chemistry. The

resolving agent forms a diastereoisomer with each enantiomer. The two diastereoisomers possess

different physical properties and consequently can be separated. The different interaction between

the enantiomers and the resolving agent that are used for the resolution are covalent (the present


41

part and 2.1.3.5) or ionic (see 2.1.3.4) bonds or weaker as intermolecular bonds (see 2.1.3.6). The

chiral molecule often belongs to the so-called chiral pool: this is a set of molecule mostly from

natural origin (but not only) which are readily available and cheap. They are often used as the

building block that introduce the asymmetry in a synthesis or as resolving agent. The chiral pool

includes aminoacids, hydroxyl acid as tartaric acid (TA), carbohydrates, alkaloids and so on

(Collins et al. 1992). Examples of resolution applied to the resolution of our targeted enantiomer,

R-1-phenylethanol and trans-1,2-cyclohexanediols, will be given in 2.1.3.5 and 2.3.1.

At the moment we would like to give and example of resolution of Rac-CHD by formation

of covalent bound with a chiral resolving agent. The method was presented in a publication

(Chatterjee et al. 2007) and summarised in Figure 14, the resolving agent is (S)-O-Acetylmandelic

Acid which forms an ester with CHD. The two diastereoisomeric esters are separated by

chromatography (or preferential crystallisation).

AcOH

Ph

COOH

OH

OH

OH

OH

OR

OH

OH

RO

OR

OH

OH

RO

OH

OH

OH

OH

(R,R)-(-)-CHD

(S,S)-(+)-CHD

2 + esterification

+

hydrolysis

hydrolysis

(R,R)-(-)-CHDee=96%Y=32%

(S,S)-(+)-CHDee=97%Y=32%

Silica gel

chromatography

Figure 14: The resolution of rac-CHD by Chatterjee.

A practical way to measure the quality of a resolution is to use the selectivity of Fogassy

parameter that combines the yield and the ee in one value. Equation 2 is used when only one of

the enantiomers is required and Equation 3 when both are. When there is no resolution, S and F

are nil. At best, S and F can be 0.5 and 1, respectively. They can be calculated with the data from

Chatterjee’s method of resolution S= 0.31 for SSCHD and F=0.62.

eeYS * Equation 2

eeYeeYF 2211 ** Equation 3

Racemisation

An elegant method of exceeding the maximum theoretical yield of 50 % is to treat the

unwanted enantiomer with a catalytic amount of a substance which leads to its racemisation as

illustrated in Figure 15. In this figure a general scheme of resolution combined with racemisation is


42

presented while other examples will be given the text. This method allows to overcome the

limitation of resolution to a yield of 50 % and the yield can be theoretically as high as 100 %. The

method of separation of the enantiomers of Figure 10 or Figure 13 were two good first examples of

resolution combined with racemisation. Another will be given in 2.1.3.5 for phenylethanol.

Figure 15: Separation of enantiomers (+)-A and (-)-

A combined with racemisation.

Pure (+)-A is obtained for all (-)-A is transformed

into (+)-A. S can be superior to 0.5.

2.1.3.4.Resolution of enantiomer by formation of

diastereoisomeric salt.

The importance of this method was given in the introduction and examples are given in the

Table 2. Many reviews and books exist on this topic that remains the main technique of production

of enantiomers despite the importance that asymmetric synthesis and biocatalysis took and the fact

it can be applied only to molecules that present acidobasic properties. The following works can be

mentioned: Enantiomers, racemates, and resolution (Jacques et al. 1981), Strategies in

optical resolutions (Wilen et al. 1977); Optical resolution via diastereoisomeric salt

formation (Kozma 2002), Strategies in optical resolution: a practical guide (Faigl et al.

2008), Chirality in industry (Collins et al. 1992), Stereochemistry of organic compounds

(Eliel et al. 1994), and Optical resolution procedures of chiral compounds (Newman 1978-

1984). Those text books helped me for the general presentation given below.

(+/-)-Anon stereoselective

separation

reaction

racemisation

(-)-A

(+)-A


43

Table 2: Examples of pharmaceuticals resolved by diastereomeric crystallisation in the

process

(reproduced from (Collins et al. 1992))

Pharmaceutical Resolving agent

Ampicillin D-camphorsulphonic acid Ethambutol L-tartaric acid (natural form) TA Chloramphenicol D-camphorsulphonic acid Dextropropoxyphene D-camphorsulphonic acid Dexbrompheniramine D-phenylsuccinic acid Fosfomycin R-(+)-phenethylamine Thiamphenicol L-tartaric acid (natural form) Naproxen Cinchonidine Diltiazem R-(+)-phenethylamine

In Figure 16, the most common resolution procedure is shown. It presents the resolution of

a base (DL)-B, whose enantiomer (L)-B is desired. The resolving agent is a chiral acid (L)-HA,

whose pKa is lower than BH+ , i.e. HA is an acid which is stronger than (DL)-BH+: (DL)-B and HA

react together and form salts that precipitate. In the most general case the separation of the salts is

performed by the selective crystallisation of the less soluble salt, n-salt. Other method of separation

of the diastereoisomeric salt exists as chromatography. The salts are then decomposed to free the

enantiomers. The decomposition is often realised with a base (or acid) that are strong enough to

take the place of an ion in the diastereoisomeric salt (In the example of Figure 16 NaOH reacts with

the n-salt to give (L)-B). To the perspective of the atom economy, any kind of waste should be

minimised but to the point of view of the profitability of the process, the separation of the free

enantiomer and the recovery of the resolving agent are essential. Indeed the loss of enantiomer or

resolving agent can be an important cost and the waste should be treated. It is difficult to give a

general method to do it (This is to say to recover (L)-HA and (L)-B in the figure below.). However,

an example of such a process for the case of a resolution with TA, a very common polar acidic

resolving agent, is given in 2.3.4 Figure 29. The workout of aqueous solution is generally done by

liquid-liquid extraction and energy intensive differential crystallisations. These steps of separation

or purification are based on simple acidobasic chemical reactions and equilibria and is the source of

much waste and energy consumption along the chemical process. TA is so cheap a resolving agent

that it is sometimes disposed of. There is also the possibility to use ion exchange resins to this

purpose or chromatographic method.


44

(D)-B + (L)-B + 2 (L)-HA (D)-BH+,(L)-A- + (L)-BH+,(L)-A- racemate resolving

agentn-salt p-salt

n-salt p-salt

separation

decomposition of the salt by NaOH

(L)-B

Na+,(L)-A-

Figure 16: Common resolution of a base B by the acid (L)-HA.

Formation of 2 diastereoisomeric salts, their separation and the decomposition of a

diastereoisomeric salt by a strong base, NaOH in this case.

Partial crystallisation has several advantages: it is relatively simple and flexible as

appropriate to intermittent batch production and only standard equipment is required. However it

can use many tanks for the storage of mother liquor and presents disposal problem because much

waste can finally be produced (Collins et al. 1992) It is possible for bases and acids only and when

the pair of salts forms a conglomerate6 but no compound nor liquid solution (or if a liquid solution

occurs it should be of a limited extend). The ee of the eutectic point must be high (Es on Figure 17

b)). A first estimation of the selectivity (Equation 4) can be derived from the eutectectic

composition xeu of the binary melting point phase diagram (Figure 17 a)), that is generally closed to

eutectic composition. The investigation of the position of the eutectic point by DSC allows an easy

and fast screening of the resolving agent (Kozma et al. 1992). The procedure is simple: a DSC scan

of an equimolar mixture of the salt p and salt n is run. So it is not necessary at this moment to have

the enantiomerically pure compounds. This scan will provide 4 values that are the temperature of

the eutectic and its enthalpy of melting and the temperature of the liquidus and its melting enthalpy:

these are the two points marked on Figure 17 for xn=0.5. With a calculation based on the Schröder-

van Laar equation, cf Equation 5 (Jacques et al. 1981), the selectivity can be evaluated (Madarász

et al. 1994) by extrapolating the eutectic composition xeu.

eu

eu

xxF

1

21 Equation 4

6 The same way that two enantiomers can form a racemic compound as shown in 2.1.3.2.


45

)11()1ln(1

12 TTR

Hx

)11(ln2

22 TTR

Hx

Equation 5: Schröder-van Laar equations

0.0 0.5 1.0

Xn

T

Salt pXeu Salt n

a)

b)

Figure 17 : a) Binary melting point phase diagram of salt p and salp n and b) their

solubility diagram.

Figure 17 b) is the solubility diagram of the salt (B-HA+) in the solvent used for

crystallisation. The line PEN is the limit of solubility of the salt, there is only a liquid phase above

it. E is the eutectic composition (Es on b) is the same xeu on b)), at which the liquid is in

equilibrium with the two salts (in the domain salt p-salt-E) where A is found. In the two last

triangular domains of the ternary P-E-salt p and N-E-salt n (where C is), the mother liquor, the

liquid that contains the unprecipitated diastereoisomeric salt, is in equilibrium with one salt. The

amount of solvent is very important during the crystallisation of the diastereomeric salts M. when

the amount of solvent is increasing from A to C the salt p disappears, the quantity of precipitated

salt n then decreases. The optimum of a resolution is found at the point B (eu

eu

xxF

1

21) where

the maximum amount of salt n has precipitated without containing any salt p. The mother liquor at

this point at the composition E, so some salt n is lost with the mother liquor and contaminated the

salt p. Consequently, the more on the left E is the better the resolution is. The lower E is the less

solvent should be used.


46

Molar ratio

The molar ration (Equation 6) is defined as the ratio between resolving and agent and

enantiomer.

nn

enantiomer

agentresolvingmr _

Equation 6: general definition of the molar ratio, mr

In the case of the most classic resolution, ie by formation of two diastereoisomeric salts, p

and n, 1 mol of resolving agent was used for 1 mol of racemic compound (molar ratio=1). Both

Enantiomers are thoroughly transformed into salt. After the mixing of the two molecules, no

species remains as a molecule (without mentioning the case where the molecule are able to

exchange more than an electron, as for TA) it was developed a technique where a non

stoechiometric amount of resolving agent is added to the racemate.

(D)-B + (L)-B + (L)-HA (D)-BH+,(L)-A- + (L)-Bracemate resolving

agentn-salt free enantiomer

Figure 18 : Resolution with a molar ratio of 0.5

The goal is to allow a better separation of the enantiomers not based on the difference of

solubility between the salts but between a salt, the more stable, and the enantiomer that did not

react. In the most typical case mr=0.5, half quantity of resolving agent is added to the enantiomers,

the method is named Pope and Peachey in reference of the two researcher who were the first to use

it. This supposed, and this is most favourable case, that all the resolving agent reacts with only one

of the enantiomers. For instance and in the most common (and favourable) case, the formed salt

precipitates completely, and the unreacted enantiomers remains in solution as shown on Figure 18.

Actually several equilibriums take place: acido basic reaction and precipitation (Ács et al. 1985).

The system of equation was solved and the optimal molar ratio is found slightly above 0.5: it

corresponds to the quantity of resolving agent for which all the enantiomer has reacted, this is to

say half the quantity of enantiomers plus a generally neglectable portion that corresponds to the

resolving agent consumed by the other equilibrium (Kozma 2002; Faigl et al. 2008).

A resolution can also run with half an equivalent of resolving agent (molar ratio of 0.5) but

also half an equivalent of a stronger base or acid, so that both enantiomers form a salt but with very

dissimilar physical properties, for instance solubility in different solvents which the further

separation can be based on. The case of the resolution of 1-(4-Fluorophenyl)-2-Methylamino)-

propane with tartaric acid (TA) is detailed in Optical resolution via diastereoisomeric salt

formation (Kozma 2002).


47

mr=0.5 and evaporation of the solvent

Several problems can arise while trying to separate the n-salt from the free enantiomer. The

mother liquor that is obtained after filtering out the precipitate may contain a variable amount of

impurities, which are unprecipitated salts, unreacted resolving agent and unwanted enantiomer.

This is why an alternative technique was developed where the solvent is evaporated. After half an

equivalent of resolving agent is added to the system and the salt n is formed the solvent is

evaporated until a mixture of solid is obtained: mostly salt and free enantiomer.

At this point two strategies are possible. The first is to pursue the distillation, and the free

enantiomer is the next compound to get out of the column. In some case a further increase in

temperature can lead to the decomposition of the salt immediately followed by it. This method

allows a very simple separation of the enantiomer as well as the regeneration of the resolving

agent. However it implies that the enantiomer is distillable, the salt releases the enantiomer before

degrading. Thus, amphetamine was resolved with its hemiamide with phtalic acid (Figure 19).

After distillation of the first enantiomer the temperature is raised and diastereoisomeric salt

decomposes: the hemiphtalate is transformed into a phthalimide (loss of CO2 and cyclisation) and

release the second enantiomer (Kozma 2002). The resolving agent can be recovered by careful

hydrolysis.

Figure 19: Resolution of amphetamine by distillation

Reproduced from Optical resolution via diastereoisomeric salt formation (Kozma 2002).

The second strategy is the following: the free enantiomer is extracted from the solid by a

solvent that presents a better selectivity for this system than the solvent used for the acidobasic

reaction with resolving agent. At the end of the extraction only diastereoisomeric salt is left and it

can be treated by an acid or a base for recovering the opposite enantiomer and the resolving agent

as previously described. For all the reasons given in 7.1.3 SCCO2 was particularly suitable solvent

for the extraction of the free enantiomer (Fogassy et al. 1994). This technique that combines


48

diastereomeric salt formation, evaporation of the solvent and SFE was used for the extraction of

(±)-cis- and (±)-trans-permetric acids when resolved by R-α-phenylethylamine and S-2-

kenzylaminobutan-1-ol (Simándi et al. 1998), of tetramisole with DBTA (Keszei et al. 1999;

Székely et al. 2002), of ibuprofene with R-phenylethylamine (Molnar et al. 2006), N-

methylamphetamine with DBTA and O,O’-di-p-toluyltartaric acid (Kmecz et al. 2007). This

technique afforded better results than using the more classical method of partial crystallisation,

notably due to the adjustable selectivity of the solvent.

2.1.3.5.Other methods of resolution of alcohols

Resolution by formation of diastereoisomeric salts is impossible for product which cannot

give or accept a proton. This is the case of alcohols. Nonetheless, they are often derivatised for this

purpose. The salt forming derivative used to this purpose are phtalates, oxalates, glycolic acid

esters, hemisulphate (Jacques et al. 1981). Phthalate derivatisation is presented on Figure 20. The

phthalate is often resolved with brucine. The alcohol is released by saponification.

O

O

OCOOH

O

OR(+/-)ROH (+/-)+

Figure 20: Formation of phtalate derivatives, useful intermediates for resolution of alcohol

by diastereomeric salt formation

A second strategy based on esterification with a carboxylic acid as a resolving agent was

presented in 2.1.3.3 on Figure 14. A third is to form a covalent bond between one of enantiomers

and an achiral molecule selectively, the asymmetry is introduced via a chiral catalyst such a

biocatalyst (see 2.1.2). An interesting instance of this method is the stereoselective acylation of

alcohol via a lipase. Only one of the enantiomer is acylated and then separated from the other. This

reaction is run in non-aqueous solvent because water would lead to the hydrolysis of the ester, an

unfavourable equilibrium. Vinyl acetate is often used for reacting with an alcohol because no water

is formed but formaldehyde that renders the reaction irreversible (as in Figure 21). However the use

of salt hydrate allows to control the water activity and to absorb the water formed during the

esterification (Halling 1992; Zacharis et al. 1997) and other techniques exists (Rosell et al. 1996).


49

OH

OH

O

OOH

O CH3

O

O

R-phenylethanol

S-phenylethanol

+

S-phenylethanol

R-phenylethanyl acetate

+

Figure 21: Resolution of chiral alcohol by enzyme catalysed acylatation

Rac-1-phenylethanol was resolved using this technique over Candida antarctica lipase

B in ionic liquids (Figure 21). The result was good as the enantiomeric was very high 99% and the

yield was 50%. A similar process with Pseudomonas cepacia lipase was developed in SCCO2 and

combined with the racemisation of the unreacted R-1-phenylethanol with the metal catalyst or the

acid catalyst Nafion SAC 13 (Benaissi et al. 2009), fairly good enantiomeric excesses and yield

were achieved, up to 95% and 85%, respectively. (R)-1-phenylethanyl acetate is sold as a fragrance

and has a floral, fresh, green note (Sheldon et al. 2007).

There is a parallel between this method of resolution and the resolution via partial

diastereomer formation, in both cases the concept of molar ratio can be applied as vinyl acetate

associated with a lipase is equivalent to a chiral resolving agent.

2.1.3.6.Resolution of enantiomers by formation of a

diastereoisomeric cocrystal with tartaric acid instead of a

salt.

TA and the related compounds have been widely used as an acidic resolving agent, their

price and availability made them the preferential source of chirality not only for resolution but also

as a catalyst for asymmetric catalysis (Collins et al. 1992). A comprehensive review about the uses

of TA and its main derivatives, diacyls (dibenzoyl- or O,O’-di-p-toluyltartaric acid) and anhydrids

for resolution exists (Synoradzki et al. 2008). Another advantage of TA is that (-)-tartaric acid, the

unnatural form, is available but more expensive.

The use of tartaric acid is not restricted to the resolution of bases. Indeed, in some cases

where even the existence of a salt was expected, IR spectroscopy did not show an exchange of

proton, and the formed compound was a complex or co-crystal rather than a salt (Nemák et al.


50

1996). No ionic but a strong system of hydrogen bonds were responsible of the co-crystallisation,

(+)-tartaric acid (TA) being a good hydrogen bond acceptor and donor. This opened the way for the

resolution of compounds without basic properties with (+)-tartaric acid derivatives and the

screening of some alcohols with O,O′-dibenzoyl-(2R,3R)-tartaric acid was realised (Kassai et al.

2000). DBTA forms also cocrystal with ethers (Szczepańska et al. 1995) and a trans-

bicyclodiamine (Hatano et al. 1994). Among TA derivatives DBTA is by far the most used for

forming diastereomeric cocrystals. And, to my best knowledge, no such compound was found with

TA.

In most cases the resolution is performed by precipitation of the diastereoisomeric

compounds, which the free enantiomer is then separated from. To my best knowledge, the

resolution of enantiomer with TA derivative that is based on a soluble complex rather than a crystal

is rare, although and example exist mentioned in (Kozma 2002) who referred to enantioselective

extraction and the partition of salts of chiral bases between water and solvents containing a

lipophilic tartaric acid ester (Prelog et al. 1983). The application is partition chromatography.

TA derivatives are obviously not the only compounds which form cocrystal susceptible to

help resolving enantiomeric pairs. The presentation of supramolecular compound was done in the

treatise Crystalline Molecular Complexes and Compounds (Herbstein 2005) and especially the

chapter 12 Hydrogen bonded molecular complexes and compounds (Vol 2) is of our interest. H

bond are undoubtedly the most widespread of the specific interactions linking molecules with

suitable functional groups together in the solid state and in the gas and liquid phases. The

enzymatic molecular complexes are examples. The cocrystal that involved TA derivative are made

of weak bonding, especially H-bond and Van der Waals interaction instead of the ionic bonds that

was responsible for the formation of the diastereiosomeric salt. The term cocrystal was chosen as

very general for this kind of compound. Herbstein proposes a more precise nomenclature based on

the crystallographic structure. So a difference is made between the rare case of molecular

complexes AB where A is mostly bounded to other A molecule by H bonds and forms only one

with B and the common case of molecular compound where the framework structures, with the two

components in alternating array The arrays may extend in zero, one, two or three dimensions

(Herbstein 2005).

In relation to the production of enantiomerically pure (R)-1-phenylethanol, an very

interesting method of resolution based on inclusion complex was described: 1-phenylethanol was

subject of a screening for resolution with a cocrystal of (1S,2R)-2-amino-1,2-diphenylethanol and

benzoic acid that possesses three-dimensionally dissymmetric cavities (Kobayashi et al. 2004). An

interesting ee of 87% was obtained for RPE, as well as a yield of 91% (to be divided by two if

referred to (±)-1-phenylethanol instead of RPE). Another example with the resolving agent (R,R)-


51

1,2-cyclohexanediamine (for CHD) will be given in 2.3.4.

The resolution by diastereomeric cocrystal formation seems to be theoretically usable to

more compounds than only acid or base if compared with salts. In practice this is not true because

it is difficult to find a resolving agent which forms a diastereomeric cocrystal and whose

cocrystallisation with the targeted enantiomer is stereoselective enough.

The separation of the cocrystal from the uncocrystallised enantiomer by extraction or

selective crystallisation is more difficult when the crystal structure is stabilised by H bonds which

are less strong than ionic bondings. That’s why the extraction with SCCO2 is so valuable in this

case, it allows to adjust finely the extraction fluid to the optimised selectivity by playing on

pressure and temperature.


52

2.2. Conversion of acetophenone to R-1-phenylethanol using alcohol

dehydrogenase from Lactobacillus brevis (LBADH)

2.2.1. Alcohol dehydrogenases require a coenzyme NADH

and NADPH that must be regenerated.

2.2.1.1.Generalities about the coenzymes

Table 3 : Coenzyme and their associated group

Coenzyme Abbreviation Group transferred

nicotine adenine dinucelotide

NAD - partly composed of niacin Hydride (H-)

nicotine adenine dinucelotide phosphate

NADP -Partly composed of niacin Hydride (H-)

flavine adenine dinucelotide

FAD Partly composed of riboflavin (vit. B2)

electron (hydrogen atom)

coenzyme A CoA Acyl groups

Coenzyme Q CoQ electrons (hydrogen atom)

thiamine pyrophosphate thiamine (vit. B1) Aldehydes

pyridoxal phosphate pyridoxine (vit B6) amino groups

3-phosphoadenosine-5’-phosphosulfate PAPS Sulphate

Biotin Biotin carbon dioxide

carbamide coenzymes vit. B12 alkyl groups

Contrary to the enzymes which require only their amino acid residues, some needs a co-

factor or a coenzyme to be active. The co-factors are inorganic ions such as Fe2+, Mg2+, Mn2+, Zn2+

and so on. For instance, Mg2+ is necessary for glucose 6-phosphatase or Cu2+ for cytochrome

oxidase. A coenzyme is a complex organic or metalloorganic molecule and brings a group that is

exchanged between the substrate and the product of the reaction (see Table 3). For instance ATP

and ADP are important coenzyme related to phosphate transfers and, thus, energy exchanges

(Nelson et al. 2004).

From now on, only the case of nicotine adenine dinucelotides NAD(P)(H) will be

envisaged. Alcohol dehydrogenases (ADH) catalyse the reaction shown on Figure 22 and requires

the coenzyme: NAD(P)H. NAD(P)H can be used for the production of other products such as


53

aminoacids, ketones from α,β-unsaturated carbonyl compounds, via D-amino acid dehydrogenase,

ene reductases respectively. The structure NADP is shown on Figure 23.

R

O

R' R R'

HOH+ NAD(P)H + H+ + NAD(P)+

ADH

* Figure 22: General activity of an alcohol dehydrogenase (ADH)

Figure 23 : NADP+: Nicotinamide adenine dinucleotide

The group that is given by NAD(P)H to the ketone to form an alcohol is an hydride (H-) (cf

Figure 24) and this reaction is consequently a reduction. To an industrial point view, the main

interest is that the alcohol is asymmetric.

Figure 24 : NAD is a hydrid acceptor and NADH a hydrid donnor.

“H-“ is used symbolically (it is not stable).

NAD is a coenzyme and its presence is necessary for the reaction to occur: no other

hydride exchanger is accepted by the enzyme. But to a certain point of view, regarding to the


54

kinetic of the process or the structural information concerning the enzyme-coenzyme bond, NAD

can be considered as a substrate for the reaction. This bond is not so strong and nicotinamidic

coenzymes are found in solution free from the enzyme. The KM(NAD)=0,01-0,1mmol/L for

commercially available formate dehydrogenase (Cordes et al. 1994). For acetophenone reduction

by NADPH by Lactobacillus brevis (LBADH): KM(acetophenone)=0.85 mmol/L and

KM(NADPH)=0.16 mmol/L (Hummel 1997). It is not the case for every coenzyme. As an example, the

heme in hemoproteins is strongly bound to the apoenzyme through a single coordination bond

between the heme iron and an amino acid side-chain. Actually the bound between the NAD(H)(P)

and the apoenzyme, the enzyme without the coenzyme, is established by weak bonds, specific (H-

bonds) and non-specific bonds (Niefind et al. 2003; Schlieben et al. 2005) for LBADH. Most

NAD(P)-dependent enzymes follow compulsory-order ternary-complex mechanism, which means

that the coenzyme-enzyme complex is formed at first and then it binds the substrate (Cornish-

Bowden 1995).

The coenzymes are expensive chemicals. Most of them are available on the market but

their price is high. As a matter of fact, their structure is complicated and their synthesis is

consequently expensive. (They are synthesised using biocatalytic steps or by purification of cell

extracts.) Table 4 gives some example of the prices of the nicotine adenine dinucleotide

coenzymes.

Table 4 : Price of some nicotine adenine dinucleotides (from Jülich Chiral Solution

GmbH’s product portfolio (february 2007))

Coenzyme Formula and molecular weight (g/mol)

Price Price by mol of exchangeable Hydride

NAD C21H27N7O14P2 × 3 H2O 717.47

850 €/250g 2,500 €/mol

NADH C21H29N7O14P2Na2

709.4 1150 €/100g 8,200 €/mol

NADP C21H26N7O17P3Na2 787.4

1330 €/100g 10,500 €/mol

NADPH C21H26N7O17P3Na4

833.4 1450 €/10g 120,000 €/mol

At those prices no process is viable if the coenzymes are used stoichiometrically. That’s

why the coenzymes need to be regenerated. This means that the form of the coenzyme which is

used for the synthesis, NAD(P)H, is given back by a second chemical reaction from NAD(P), see

Figure 25. The number of cycles undertaken by the coenzyme, or the number of time the coenzyme

has been regenerated is called “coenzyme total turnover number”. This concept is symmetrical to


55

any turnover used for the description of a catalyst, an enzyme for instance: actually in this case the

coenzyme is also a catalyst, it does not appear in the chemical equation of the overall reaction.

Figure 25 : Coenzyme regeneration in the

case of a reduction with NAD(P)H.

There are three main ways to

regenerate coenzymes this is to say how to

perform the “second reaction” of Figure 25:

Regeneration by a second enzyme that depends on the same coenzyme: a second enzyme

and its associated substrate(s) is added to the reaction medium.

Regeneration by the same enzyme as for the synthesis of the product: a sacrificial reactant

is added to the reaction medium.

Chemical method and electrochemical method.

At the moment the last case is rare, the coenzyme stability being still an issue but the two

first techniques are used.

2.2.1.2.Regeneration by a second enzyme

The system that is the most used for the regeneration of NAD(P)H is based on formate

dehydrogenase:

O

OH+ NAD(P)H + NAD(P)+CO2

Equation 7: Formate dehydrogenase

Formate Dehydrogenase (FDH) is the most used system for NADH regeneration in

bienzymatic system (Eckstein et al. 2004a). Indeed, the equilibrium constant of the reaction

presented in Equation 7, which gives NADH back, has an equilibrium constant of 15000 (Liese et

al. 2006), so that the reaction is almost irreversible because the by-product of the reaction, CO2, is

easily separated as a gas. This is not the case of the by-product of every regeneration method

(compare with Glucose 6-phosphate). Formate is an available and very cheap product. Its

consumption over the reaction leads to a modification of the pH, which should be controlled

therefore. Properties and price of a NAD-specific FDH and a NADP-specific is given in Table 5.

NAD(P)H NAD(P)+

PRODUCTSUBSTRATEfirst reaction

second reaction


56

Table 5 : Some properties of Formate Dehydrogenase (FDH) (Product Portfolio 2007)

Coenzyme NAD-specific FDH

From Candida boidinii (E. coli recombinant)

NADP-specific FDH from Pseudomonas spec. 101 (E. coli

mutant recombinant)

KM(formate) 13 mmol/L 12 mmol/L

KM(NAD(P)) 0.09 mmol/L 0.29 mmol/L

References (Schütte et al. 1976; Cordes et al. 1994) (Tishkov et al. 1993)

Price 38€/kU 500€/kU

As it can be seen in the table, NADPH regeneration is not as good as for NADH. First the

enzyme is more expensive, and the KM(NAD(P)) is higher.

Other way of regenerating NAD(P)H are given in the Equation 8 (Gu et al. 1990), Equation

9 and Equation 10 (Van Der Donk et al. 2003).

+ NAD(P)H + H+ + NAD(P)+gluconolactone glucose Equation 8: regeneration with glucose dehydrogenase.

+ NAD(P)H + H+ Glucose-6-phosphate + NAD(P)+6-phospho-D-gluconate

Equation 9: regeneration with Glucose-6-P dehydrogenase.

+ NAD(P)H + H+ Lactate + NAD(P)+pyruvate

Equation 10 : regeneration with Lactose dehydrogenase.

2.2.1.3.Regeneration by the same enzyme: sacrificial

substrate method.

This is the technique that was used in the experimental work. Only one enzyme is used

during the process. A second substrate which is called “sacrificial substrate” and possesses a group

belonging to the same family as the product is added to the system. The second substrate is not

transformed into a valuable product: it is “sacrificed”. In our case, R-phenylethanol is produced

from acetophenone and NADPH according to the first equation of Figure 26. NADPH is

regenerated using isopropanol as sacrificial substrate. It undergoes the reaction opposite to the

substrate’s (second equation of Figure 26), which gives NADPH back. In the last line of the Figure

26, the final yield is given where both NADP and LBADH are catalysts.


57

Figure 26 : Reactions catalysed by LBADH, conversion of acetophenone and regeneration

of NADPH by addition of isopropanol.

Isopropanol is in a large excess so that the equilibrium of the reaction is favourable to the

production of R-1-phenylethanol: The final equilibrium of such a reaction is determined by

Equation 11. Enzyme catalysed reactions were used for the determination of thermodynamic

equilibrium in different solvents including SCCO2 (Tewari et al. 2005).

aaaa

lisopropanoACP

acetoneRPEKeq

Equation 11: Thermodynamic equilibrium of the conversion of ACP into RPE.

To increase the final concentration of the product, different possibilities exist:

Increasing the sacrificial substrate initial concentration (but it can lead to enzyme inhibition

or deactivation),

Removing the sacrificial product during the reaction. As an example, when isopropanol is

used as sacrificial substrate in the case of ketone reduction, acetone, the sacrificial product,

can be eliminated by evaporation (acetone is more volatile than isopropanol) or by

pervaporation (for instance, the reaction media is in contact with a membrane which has a

reduced pressure on its other side (Stillger et al. 2002)).

2.2.2. Different studies with alcohol dehydrogenase from

Lactobacillus brevis

LBADH was chosen because of its availability, low price, the rare orientation it gives to its

products, and promising application in gas phase reactor. Moreover, no trial of synthesis in organic

media has ever been published.


58

The discovery of LBADH went back to the preliminary work on LBADH (Hoshino 1960).

The enzyme which will be used for this study was latter patented by Hummel (Hummel et al. 1997)

and also presented in a publication (Hummel 1997). The gene coding for Alcohol dehydrogenase

(ADH) from Lactobacillus brevis (LBADH) was expressed in E. coli. LBADH is commercialised

by Jülich Chiral Solution GmbH, Codexis, Jülich, Germany (Codexis 2008), as a lyophilised

powder (crude enzyme preparation).

The structure of LBADH was resolved and the crystallographic structure of LBADH was

presented in three works (Niefind et al. 2000; Niefind et al. 2003; Schlieben et al. 2005): it is a

homotetramere bound by Mg2+, its molecular weight is 4*26,6 kDa=106.4 kDa. It catalyses the

conversion of acetophenone (ACP) to R-1-phenylethanol (RPE) using the coenzyme

NADP/NADPH (first equation in Figure 26). LBADH presents a high enantiomeric excess for the

synthesis of R-1-phenylethanol (RPE) which is generally enantiopure. More generally, LBADH has

a broad substrate range and the synthesis could be therefore applied or adapted to another product,

notably substrates having a relatively high molecular weight. LBADH accept other carbonyls as

substrate, such as acetophenone derivatives, propiophenones, aliphatic open chain ketones, 2- and

3-ketoesters, cyclic ketones and so on (Hummel 1997). LBADH give the anti-Prelog orientation to

their products, as well as LKADH (Lactobacillus kefir). This is an advantage because most ADHs

lead to Prelog orientation, for instance HLADH (from horse liver), TBADH or ADH T (from

Thermoanaerobacter brockii), and YADH (Bakers’ yeast from Saccharomyces cerevisiae).

The specific activity of LBADH is 490 U/mg (Hummel 1997). No complete kinetic study

of LBADH is available. Nonetheless, pieces of kinetic data are available: Hummel measured the

following kinetic parameters for LBADH: Km(acetophenone)=0.85 mM and Km(NADPH)=0.16 mM

(Hummel 1997). In a publication (Schlieben et al. 2005) two LBADH are investigated: the wild-

type and a mutant. Only the values concerning the wild-type are given here: for the NADP

reduction Km(NADP)=0.015 mM, kcat=4.4 s-1 (substrate saturation), and for the NADPH oxidation

Km(NADPH)=0.04 mM, kcat=38.1 s-1 (substrate saturation). For the acetophenone reduction by

NADPH, Km(acetophenone)=2.8 mM, kcat=44.5 s-1 (cosubstrate saturation), and for the oxidation of

phenylethanol by NADP Km(phenylethanol)=2.9 mM, kcat=5.4 s-1. A simple kinetic model is proposed

(initial rate only) for the study of substrate inhibition (butan-2-one as the substrate) (Schumacher et

al. 2006). The issue was to find out whether a cosolvent could limit the inhibition. An important

deactivation due to the cosolvent was noticed.

An unusal purification of LBADH based on extraction in an ionic liquid was patented

(Dreyer et al. 2008b; Dreyer et al. 2008a). This enzyme was recently further improved by

engineering (Ching et al. 2008). Lactobacillus brevis was also engineered for the production of

ethanol out of biomass (Liu et al. 2007).


59

The stability of LBADH was presented in most publications presenting a process involving

it. A more specific research were done about its stability with a cosolvent as acetonitrile or 1,4-

dioxane in water (Schumacher et al. 2006). The deactivation of LBADH in biphasic system

water/organic solvent for storage was investigated (Villela Filho et al. 2003).

LBADH was used in many processes. In most cases it was used dissolved in water, as well

as the coenzyme. Some processes were based on whole cell with regeneration by sacrificial

isopropanol and removal of acetone in continuous flow (Schroer et al. 2007b), with formate

dehydrogenase (Ernst et al. 2005). The different methods of regeneration are compared in (Schroer

et al. 2007a) and the method with isopropanol gave the best results. In a case the enzyme was

coimmobilised on an amino-epoxy support plus treatment with glycine, mercaptoethanol

glutardialdehyde and applied to a plug-flow reactor (Hildebrand et al. 2006) and the stability of the

catalyst was much improved (This case of immobilisation is detailed in 2.1.2.5). Crosslinked

enzyme particles were made using a polyaldehyde prepared from a polysaccharide as the water-

soluble crosslinking agent (Mateo et al. 2004). Stabilisation with ionic liquid were also investigated

(Braeutigam et al. 2007; Dreyer et al. 2008a). A biphasic system of ionic liquid and water can

present an advantageous coefficient of partition that improve the yield of the conversion (Eckstein

et al. 2004b).

Most patented processes involve an aqueous phase which is extracted at the end of the

reaction and, sometimes continuously. Never a dense gas was used for this purpose. Isopropanol

regenerates NADPH, for instance: (Mueller et al. 2003; Peschko et al. 2005; Groeger et al. 2006;

Meudt et al. 2006; Peschko et al. 2006; Pfaller et al. 2007; Yasohara et al. 2007; Groeger et al.

2008).

No process was run in the absence of water except for gas phase reaction run at

atmospheric pressure (Ferloni et al. 2004; Trivedi et al. 2005; Trivedi et al. 2006) presented in next

paragraph.

2.2.3. Alcohol dehydrogenase in non aqueous solvent.

The general case of enzyme-catalysed reaction in organic solvent was treated in 2.1.2.7. In

the case of coenzyme-dependent reactions, permanent complexes between enzyme and coenzyme

are formed in organic solvent; hence no dissociation takes place, even at low affinity as in the case

of LBADH and NADP. LBADH was successfully immobilised on glass beads using sugar as a

stabiliser (Ferloni et al. 2004; Trivedi et al. 2005; Trivedi et al. 2006) and a residual activity

superior to 300 % was observed probably because of “structural changes in the enzyme molecule

during the drying process” (Trivedi et al. 2005). The method for co-immobilising the enzyme


60

and the coenzyme on glass beads that will be used herein was adapted from this work

about gas-phase continuous reaction.

Several studies about ADH in non aqueous media were carried out. Except from Deetz’s

preliminary study (Deetz et al. 1988), no successful way for coenzyme regeneration with a second

enzyme was proposed : the regeneration is undertaken by the same enzyme with a sacrificial

substrate. The conversion was catalysed by HLADH (Horse Liver alcohol dehydrogenase, an NAD

dependent ADH) in isopropyl ether, butyl acetate, chloroform (Grunwald et al. 1986). The reaction

was catalysed by YADH (Yeast ADH) co-lyophilised with NAD in heptane (Yang et al. 1993). The

stability of YADH and TBADH (Thermoanaerobacter brockii ADH) in different organic solvents

(n-dodecane, n-octane, toluene, and pyridine) was measured (Miroliaei et al. 2002). Fluorinated

NAD was used with HLADH in SCCO2 (Panza et al. 2002) . The conversion and the total turnover

was investigated with YADH and HLADH (Snijder-Lambers et al. 1991) on different carrier and in

different solvents. The influence of the water activity was tested (Adlercreutz 1991; Jonsson et al.

1999; Jönsson et al. 1999a) in hexane with TBADH and HLADH and kinetic constants evaluated.

The equilibrium between 2 pairs of related alcohol and ketone (for instance isopropanol and

acetone or 2-pentanol and 2-pentanone) was studied using a commercially available preparation of

ADH and coenzyme (Tewari et al. 2005). Polymere (ethyl cellulose)-aided solubilisation of

HLADH and NAD is tested in different solvents at different water activities with different enzyme-

coenzyme ratios (Virto et al. 1995). Generally YADH presents the advantage of its price and

availability while it deactivates rapidly. HLADH and TBADH are more stable and TON as high as

a million were obtained with them. Many studies showed that the water activity the most

favourable to the initial is the highest (Deetz et al. 1988; Yang et al. 1993; Virto et al. 1995;

Jönsson et al. 1999b). Indeed the maximum of activity of HLADH is observed when a monolayer

of water surrounds it (about thousand water molecule) (Adlercreutz 1996). However, only few

studies investigated the variation of stability according to the water activity (Yang et al. 1993). In

this study a high water activity led to a faster deactivation, but the low reaction TON is

counterbalanced by the increase in the reaction rate.

The hydrogenation of acetophenone by coimmobilised LBADH and NADP in gas phase

(Ferloni et al. 2004) gave very good results (high turnover, enantiomeric excess and space-time-

yield.) NADPH was regenerated by isopropanol, while the enzyme and the coenzyme were

coimmobilised on glass beads. The conditions of co-immobilisation were optimised: the choice of

the support, the temperature and pressure of drying, the amount of NAPD and the addition of

stabiliser (sucrose) (Trivedi et al. 2005). The influence of the water activity was investigated and

the results fit the previous well: when the water activity is higher, the initial rate is higher and the

half life is smaller. The effect of the temperature was also investigated and an increase in


61

temperature increased the reaction rate but decreased the half-life of LBADH. They eventually

chose to perform the reaction at 30°C, water activity at 0.55: excellent TON as high as 4 million

were thus afforded (Trivedi et al. 2006).

Dense gases have been used as solvents for enzyme-catalysed reactions, with lipases in

most cases (Habulin et al. 2007; Knez 2009). Whereas ADH were relatively often tested in non

aqueous solvent (Klibanov 2003), processes with ADH in non-aqueous solvents are rather rare due

to their instability in them (Lavandera et al. 2008) and they were seldom tested in dense gases

(Matsuda et al. 2000; Panza 2001; Tewari et al. 2005). Matsuda set up a process with immobilised

resting cell of Geotrichum candidum in SCCO2 but the biocatalyst was too prone to deactivation

(Matsuda et al. 2003).

2.2.4. Goal of this work concerning LBADH in dense gases

As showed previously, very few processes based on ADH were run in dense gases and little

is known about their stability in those fluids. Our laboratory has an expertise in enzyme-catalysed

reaction in non-aqueous solvent and especially dense gases. However, processes using ADHs were

not treated at the moment. So an important part of the practical work is the development of

protocols in relation to this family of enzyme, as techniques of immobilisation, reaction setup in

gas phases. The goal was to address those issues: whether a conversion is possible, which

enantiomeric excess presents the product, what stability the enzyme has in this medium. LBADH,

as a first target, was tested.

2.3. Resolution via the formation of diastereomeric complexes with (+)-

tartaric acid followed by extraction with supercritical carbon

dioxide applied (±)-trans-1,2-cyclohexanediol

(R,R)-trans-1,2-cyclohexanediol (RRCHD) and (S,S)-trans-1,2-cyclohexanediol

(SSCHD) are important building blocks or chiral auxiliaries (Groaning et al. 1998; Tanaka et al.

2001; Tiecco et al. 2003; Wojaczynska et al. 2008). This bibliographical review was limited to

CHD and resolution. Rac-CHD is a cheap product that is almost free of cis-1,2-cyclohexanediol.

Indeed the organists possess several anti addition to produce CHD from cyclohexene (Clayden et

al. 2001). It is interesting to mention that enantioselective synthesis of CHD were described, as

example the asymmetric hydrogenation of cyclohexane-1,2-dione over cinchonidine-modified

platinum (Sonderegger et al. 2003). Biocatalytic methods exists for the synthesis of an enantiomer

of CHD, the hydrolysis of the corresponding epoxyde by an epoxide hydrolase (Chiappe et al.

2007). Many chromatographical methods were described but in an analytical purpose in most

cases.


62

2.3.1. Different methods of resolution of (±)-trans-1,2-

cyclohexanediol based on the formation of a covalent bond.

Numerous methods of resolution of (R,R)-CHD and (S,S)-CHD have been developed. The

majority of them involved the formation or the hydrolysis of an ester. Those resolution techniques

found a source of asymmetry in a biocatalyst, a chiral reagent or a chiral non biological catalyst.

The biocatalytic resolution techniques based on an ester bound are :

Selective hydrolysis: hydrolysis of their racemic acetates and chloroacetates in the

presence of a highly selective ester hydrolase from Pseudomonas sp. (Laumen et al. 1989),

by fermentation with Rhizopus nigricans (Kawai et al. 1981), monoacetates were also

preparated. from the racemic diacetates by lipase-catalyzed hydrolysis (Bodai et al. 2003)

and the 2 step hydrolysis of diacetate CHD by porcine liver esterase (PLE) and then lipase

from Pseudomonas cepacia (Caron et al. 1991).

By transesterification with Pseudomonas cepacia lipase (Kaga et al. 1998) or by fungal

lipases (Bodai et al. 2003).

By a selective esterification: acylation by lipase YS (from Pseudomonas fluorescens)

(Naemurz et al. 1995), lipase from Pseudomonas cepacia, and lipase from Candida rugosa

(Kazlauskas et al. 1991).

The resolution was also realized by esterification with chiral carboxylic acid such as o-

acetyl mandelic acid (Chatterjee et al. 2007) or Ma NP acid (Kasai et al. 2004). Chiral catalysts

were used in those cases: acylation catalysed by Cu(II)(borabox) (Mazet et al. 2006) or by amine-

phosphinite bifunctional organocatalysis derived from quinidine (Mizuta et al. 2006) and

monobenzoylation catalyzed by organotin compounds (Iwasaki et al. 2000).

Dispiroketals were used for the resolution of CHD, a chiral diketone was used as a

protecting group for CHD by formation of 1,2-diacetal.

Other resolution techniques were developed: Dispiroketals selectively reacted with CHD

(Ley et al. 1996a; Ley et al. 1996b). A chiral diketone was used as a protecting group for CHD by

formation of 1,2-diacetal (Lenz et al. 1998). The hydroboration-oxidation with

diisopinocampheylborane of benzyl or diphenylmethyl vinyl ethers, followed by cleavage

(Peterson et al. 1988), the kinetic resolution of CHD by Bacillus stearothermophilus diacetyl

reductase (Bortolini et al. 1998).

2.3.2. Resolution of (±)-trans-1,2-cyclohexanediol by selective

formation of a cocrystal

CHD was resolved (Kawashima et al. 1991) by reacting with an other resolving agent

(R,R)-1,2-cyclohexanediamine that formed a co-crystal with SSCHD preferentially. This method


63

gave SSCHD in a yield of 73% based on the enantiomer present in the racemic compound. Its

optical purity was 67%.

The structure of the co-crystal between SSCHD and (R,R)-1,2-cyclohexanediamine was

established and discussed in publication by Hanessian (Hanessian et al. 1994; Hanessian et al.

1995; Hanessian et al. 1999). Other co-crystals were found between RRCHD and (R,R)-2,3-

diaminobutane (Hanessian et al. 1999) or N-methylmorphine-N-oxide (Chanzy et al. 1982). These

structures present an inner core based on hydrogen bonds between the diamine and diol moieties,

which are responsible for the stereoselectivity of the crystallisation and the geometry of the crystal,

and an outer region formed of the hydrophobic groups.

Inconvenient separation of the product from the resolving agent.

A matter with the method of resolution with (R,R)-1,2-cyclohexanediamine

(Kawashima et al. 1991) is that the diol and the diamine have comparable solubility in most

solvents and the decomposition of the cocrystal and the separation of (R,R)-1,2-

cyclohexanediamine from CHD required a silica-gel short column. The problem was

encountered in the case of the resolution of ibuprofene with the resolving agent R-(+)-

phenylethylamine, where the separation of the salt and the free enantiomer was done by SFE

(SCCO2) (Molnar et al. 2006): in both case the resolving agent and the racemic compound

present similar polarities and, thus, solubility. The SFE of the free enantiomer lead to an

extract that is polluted with R-(+)-phenylethylamine that was removed by liquid-liquid

extraction: the acidic aqueous solution that allows to have R-(+)-phenylethylamine as a

hydrophilic cation and ibuprofen as a neutral acid that is extracted in an apolar organic

solvent. In both cases (Kawashima et al. 1991; Molnar et al. 2006), the extracts presented

good enantiomeric excesses but the extra step of purification to remove the resolving agent

induces disadvantages: the resolving agent is lost (in one case absorbed on silica gel and in

the second dissolved into liquid solution) and its recovery is possible but expensive, possibly

energy-intensive. Tartaric acid (TA) comparatively presents an important advantage: its

solubility is extremely low in SCCO2.

Our group already studied the resolution of 1,2-disubstituted cyclohexane using TA

derivative (Székely et al. 2004) based on the screening of the resolution of secondary alcohols

with DBTA (Illés et al. 2002). In the case of the resolution of trans-2-chloro-cyclohexanediol

with DBTA it was possible to decompose the diastereomeric cocrystal in the extractor at

moderate temperature and to extract the released (S,S)-enantiomer.


64

2.3.3. Physical properties of CHD.

Table 6: Melting point and structural data for crystalline phases of trans-1,2-

cyclohexanediol and references.

Tfus (K)

Reference(s) for Tfus

Space group, cell parameters: a b c (Å) α β γ (°)

Refcode in CSD (Allen 2002),

PDF file No. from ICDD T(K)

RacCHD Stable

polymorph

376–377

Verkade et al. (1928) quoted in (Lloyd et

al. 2007)

Pbca 7.885 19.301 8.498

90 90 90

ZZZKPE01, 02-064-1664

ca 295 K (Sillanpaa et al. 1984)

377 Lettré & Lerch

(1952) quoted in (Lloyd et al. 2007)

Pbca 7.888 19.333 8.501

90 90 90

ZZZKPE02, 02-073-8658

ca 295 K (Jones et al. 1989)

376.4 (Leitao et al. 2002) Pbca

7.885 19.301 8.498 90 90 90

ZZZKPE04 02-093-6744

299 K (Lloyd et al. 2007)

377 White (1931) quoted in (Lloyd et al. 2007)

RacCHD Metastable polymorph

C12/c1 18.578 10.007 7.272

90 96.32 90 subliming at RT

ZZZKPE06 299 K (Lloyd et al. 2007)

RRCHD or SSCHD Stable

polymorph

382.5 (Leitao et al. 2002) P3121

10.191 10.191 10.821 90 90 120

PIWXIK, 215 K (Hanessian et al. 1994)

P3221

10.229 10.229 10.909 90 90 120

RIHMAX01 02-093-3042

299 K (Lloyd et al. 2007) RRCHD or

SSCHD Metatable polymorph

352.8 (Leitao et al. 2001)

CHD was the subject of several publications because of its industrial importance and also

different features very interesting to a theoretic point of view (see

Table 6). The pure enantiomers crystallised according to two different forms that melts at

382.5 K and 352.8 K and the crystal are generally a mixture of the two forms whose difference in

melting can be observed (Leitao et al. 2001). Only the stable polymorph’s structure was determined

(Hanessian et al. 1994; Lloyd et al. 2007). A racemic compound (racCHD) of (R,R)- and (S,S)-1,2-

cyclohexanediol is present under two polymorphic forms, as well. The stable polymorph’s structure

was determined by three research groups (Sillanpaa et al. 1984; Jones et al. 1989; Lloyd et al.

2007) and the metastable’s by one (Lloyd et al. 2007).


65

The solid-liquid melting phase diagram for mixtures of RRCHD and SSCHD was

investigated by DSC methods (Leitao et al. 2002). This binary system although resembled a type of

melting phase diagram including a racemic compound (Jacques et al. 1981), but had an unusual

feature. Usually, a racemic compound, as a co-crystal of both enantiomers, can form a eutectic

mixture, i.e. crystal conglomerate with one of the enantiomers as presented Figure 27 a).

a) usual case

b) the case of RR and SSCHD

Figure 27: Comparison of 2 solid-liquid melting phase diagrams a) a usual for a couple of

enantiomers and b) CHD’s case with a solid solution.

Actually, in this case, eutectic compositions were found at about molar fractions of XSSCHD

= 0.2 and 0.8. Furthermore the corresponding eutectic temperature of Teu=371 K was observed in

the composition ranges 0 < XSS-CHD < 0.2 and 0.8 < XSS-CHD < 1, but at intermediate composition,

0.2 < XSS-CHD < 0.8, the endothermic heat effect of eutectic melting does not occur (Leitao et al.

2002), see Figure 27b. This phenomenon was explained by Leitao et al. as the proof of the

formation of a solid solution for compositions within the latter wide range around the 1:1 molar

ratio. A metastable phase was also found below eutectic temperature (xSS-CHD < 0.2 and xSS-CHD >

0.8), as the continuation of the liquidus curve of solid solution (Leitao et al. 2002). In this figure the

liquidus and solidus of the solid solution are not distinguished. Indeed DSC measurements showed

only a broad peak probably because these two effects occurred at very closed temperature by DSC.

The interval of composition where the liquid solution is found rather than two compounds

according to the temperature is not precisely known. However, according to their results to Leitao,

the solid solution should occupy the interval 0.2-0.8 for all the studied temperatures because no

transition was observed below.


66

2.3.4. Method of resolution of CHD by formation of a cocrystal

CHD-TA and optimisation of the parameters of extraction:

temperature and pressure.

This part was subject of the paper I co-authored with Peter Molnar (Molnar et al. 2008) and

is included to his PhD thesis (Molnár 2009).

Resolving agents of TA’s family were screened and TA was eventually found to be the

best. The method of resolution of CHD with TA and extraction with SCCO2 is schematically

represented on Figure 28.The source of asymmetry in this method of resolution comes from TA.

This method consists in three steps:

1. The selective cocrystallisation.

2. The extraction by SCCO2.

3. The raffination of the residuum, decomposition of the co-crystal.

OHOH

O

OH OH

H OH

OH

OH

OH

OHOH

OH

OH

OHOH

OHO

OH OH

H OH

(R,R)-(+)-tartaric acid

(S,S)-(+)-CHD

(S,S)-(+)-CHD(R,R)-(-)-CHD

+ +

,

evaporation of the solvent (ethanol)

Co-crystal to be raffinated

Extracted inSCCO2

Figure 28: Principle of the resolution of (±)-CHD by co-crystal formation followed by an

extraction in SCCO2 (mr=0.5)

The selective cocrystallisation consists in the dissolution of racCHD and TA in two

separate ethanolic solutions that are then mixed. The solution gets cloudy, which presumably

shows that the cocrystal has precipitated – this interpretation was the most natural at this moment

of the research but will reveal wrong as shown in the experimental part (and more precisely in

4.2.4.3). An inert filtering agent, perfil, is added to it. Ethanol is then evaporated. A cocrystal is

preferentially formed between RRCHD and TA: it “captures” RRCHD but no or little SSCHD

because only no or only little cocrystal between SSCHD and TA is formed. The SS enantiomer

remains “free”, this is to say not “cocrystallised” or “uncocrystallised” and is due to be extracted

with SCCO2. The quantity of TA in relation to the quantity of CHD is an important factor during

the extraction and the molar ration (mr) is defined as :


67

nn

CHD

TAmr

Equation 12: definition of the molar ratio, mr

At the moment only the more usual mr of 0.5 was studied. This molar ratio gives the best

result in the ideal case. This is to stay when the perfect cocrystallisation occurs between 1 mol of

CHD (0.5 mol SSCHD + 0.5 mol RRCHD) and 0.5 mol of TA, 0.5 mol of CoC is formed and 0.5

mol of SSCHD are free (ideal case).

The free CHD, mostly SSCHD, is extracted with SCCO2. What is left over in the extractor

mostly consists in the cocrystal presumably. Those methods, that will be used herein again, are

detailed in 4.1.4 and 4.1.4.2.

The residuum is treated in order to recover RRCHD. The decomposition of the cocrystal or

raffination is based on the acidic properties of tartaric acid: the tartaric acid treated with a basic

aqueous solution. Water was then rotoevaporated and CHD was extracted by chloroform. The

detailed protocol (precisely given in 4.1.4.3) is presented in Figure 29. This method presents

several disadvantages: it includes many steps, is time-consuming, uses much solvent, generates

wastes.

Figure 29: Decomposition of the residuum.

A perfect extraction selectivity, in this case, means that a) the partial cocrystallisation is

perfect so that CoC contains all and only RRCHD consequently all free CHD is SSCHD b) all and


68

only SSCHD is extracted. This extracts presents no RRCHD released by CoC so the cocrystal

RRCHD-TA does not decompose at all over the supercritical extraction. The yield of extraction is

50%. c) No CoC is lost and the decomposition of CoC is total. So the yield of the decomposition is

50%. The temperature and the pressure play an important role in the extraction. Indeed, the

solubility of CHD in SCCO27, the speed of decomposition of the cocrystal varies with temperature

and pressure. The density, viscosity and diffusivity changes with the temperature as well. The study

of the influence of the pressure and temperature of extraction on the F-parameters (Equation 3) was

the subject of an experimental design (Molnar et al. 2008) and the best resolutions within the range

of the experimental design gave F about 0.6 and were achieved under the following conditions: P =

10 MPa, T = 63 °C; or P =20 MPa, T = 33 °C (see Table 7).

Table 7: Comparasion of different results obtained for the resolution of CHD

According to the result published in (Molnár 2009). The “middle point” (150 bar, 48°C) was

repeated 4 times, the standard deviation is given.

Yext (%) eeext (%) Yraf (%) eeraf (%) F 0≤F≤1

Ideal resolution with mr=0.5 50 100 50 100 1

Middle point mr=0.5,

Pext=150 bar Text=48°C

54.0

±1 55.5

±3

31.5

±3

76.4

±4

0.54

±0.025

mr=0.5, Pext=100 bar Text=63°C

47.8 61.4 36.9 91.9 0.61 Experimental

resolution

mr=0.5, Pext=200 bar Text=33°C

50.6 62.1 33.7 81.9 0.59

2.3.5. Issues concerning the resolution of CHD by

cocrystallisation and SFE to be addressed in the present

work.

The resolution of CHD by cocrystallisation and SFE is performed in three steps as shown

previously (in 2.3.4). The second steps, the extraction, was optimised and is not the concern of the

present work. The last step (Figure 29) presented several disadvantages: it is time consuming and

7 It should not be forgotten that the solubility of CHD depend on its enantiomeric excess.


69

has very bad contribution to the E-factor, while all the previous steps are essentially green. It is

necessary to develop an alternative method for the decomposition of the cocrystal.

The eeext seemed to be limited at about 60% (cf Table 7). It should be investigated if this is

due to a thermodynamic limitation during the sample preparation or another matter arising during

the sample preparation and/or the SFE. And generally, it is aimed at a better knowledge of what

occurred during the extraction. Thus other issues arise: what can be learnt by analytical techniques

and what is the structure of the cocrystal RRCHD - TA responsible for the extraction the

stereoselectivity and what is the structure of its counterpart cocrystal of SSCD – TA. One of the

goal is to observe which kind of data can be afforded by DSC and to check if it corroborates with

other techniques such as XRD.

Conversion of acetophenone to R-1-phenylethanol

70

3. Conversion of acetophenone to R-1-phenylethanol

3.1. Materials and methods

3.1.1. Materials

3.1.1.1.Reagent

NADPH and NADP were provided by Jülich chiral solution GmbH, Codexis, Jülich,

Germany (Codexis 2008). Heptane, acetone, isopropanol, phosphate buffer solution (pH = 7), were

provided by Merck, and MgCl2 was added so that [Mg2+]=1 mM. MgCl2.6H2O, orange silica gel

were provided by Riedel-deHaën. Acetophenone, acetophenone standard, and Na2CO3.10H2O were

provided by Fluka. Decane was provided by Aldrich. Sucrose, glass beads (425-600 µm), R- and S-

1-phenylethanol standards were provided by Sigma.

3.1.1.2.Biocatalyst

LBADH expressed in E. coli, was provided by Jülich Chiral Solution GmbH, Codexis,

Jülich, Germany as a lyophilised powder (crude enzyme preparation).

2 different batches of lyophilised powder were used. The first was used for the conversion

of ACP into RPE. The provider indicated that the activity before shipping was of 9.3 U/mg while

we measured it at 8 U/mg at the reception. From the value of the specific activity we deduce that

the content of active enzyme of this enzyme preparation was about 1.6% and that 1 mg of this

preparation contains 0.15 µmol. This value was used for calculation of TONE.

The second batch was used for the study of deactivations. Its activity was 69 U/mg before

shipping and 21 after. The meaning of U and the way to measure it is given in 3.1.2.4.

3.1.2. Methods

3.1.2.1.Preparation of the biocatalyst

The preparation is indicated for the catalyst used for the reactions in propane. The catalyst

for reactions in heptane was prepared according to the same method but with different quantity of

material. The details are given in Table 16 in annex 7.2. The co-immobilisation of LBADH and

NAPD on glass beads was done the following way: enzyme (25 mg) and coenzyme (11 mg) were

diluted in the buffer in a beaker. The solution was stirred for 10 minutes at 4°C. The support, 5 g of

glass beads, was then added. This mixture was stirred for 1 hour at 4°C. The beaker was placed into


71

a desiccator at the pressure of about 50 kPa for 2 hours and then at the pressure of about 10 kPa

until it dried. This protocol was adapted from the work done in gas phase (Trivedi et al. 2005).

3.1.2.2.High-pressure view cell

The reactions at high-pressure were performed in a high-pressure reactor with sapphire

windows which is presented in Figure 30. This apparatus provided by NWA GmbH (Lörrach,

Germany) (NWA (http://www.nwa-highpressure.de)) consists of a large piston, equipped with a

mechanical stirrer, heater and thermostat. It is closed by a sapphire window at each extremity, and

can be seen through. The back sapphire window is mobile and the volume of the cell thus variable

(from 60 to 30 cm3). The pressure inside the cell is controlled by the operator. The back sapphire

window is pushed by hydraulic oil whose pressure is set and balances the pressure inside the cell.

Pressurized propane was supplied by a high-pressure membrane pump (PM-101, NWA

GmbH, Lörrach, Germany). When the catalyst was immobilised on glass beads it was introduced

into the opened cell. The cell was assembled and the liquid reactants, isopropanol and

acetophenone with decane, the internal standard (as well as the aqueous phase in the case of the

biphasic system) were injected through an opening at the top of the cell. The cell, which had been

sealed, was filled with dense gas until the gas phase was present only in a small amount. When the

desired temperature was reached the pressure was adjusted by changing the position of the back

sapphire window. Sampling without depressurisation was made possible with an autosampler of the

type of HPLC injector. The volume of the cell was reduced by the volume of the sample as the

movable sapphire windows moved towards the other extremity. The samples were sprayed into a

test tube which was then rinsed with hexane. This solution was analysed by Gas chromatograph.


72

Figure 30 : Scheme and picture of the high-pressure reactor.

1 liquid propane, 2 high-pressure pump, 3 one-way valve, 4 needle valve at the inlet of the

view cell, 5 needle valve (sampling of the lower phase) 6 static sapphire window (the view

cell can be opened at this level), 7 mobile sapphire (It slides inside the cell), 8 mechanical

stirrer, 9 temperature indicator and temperature regulation, 10 heaters, 11 pressure

indicator, 12 autosampler, 13 safety rupture disk, 14 air, 15 compressor, 16 air pressure

regulator, 17 hydraulic oil system, 18 hydraulic oil, 19 pressure regulator, 20 safety pane.

3.1.2.3.Reaction with LBADH

The reaction was performed using the regeneration method described in 2.2.1.3. It relies on

the “sacrificial substrate” isopropanol. This reaction was run in different media: water, heptane,

propane with coimmobilised catalysts and in the biphasic system water-propane.

The reaction run in water at 30°C was catalysed by 8 mg NADP and 15 mg LBADH.

50mL of buffer contained 10 mmol of acetophenone 1.5 mol of isopropanol. The reaction started

by adding the 2 solid catalysts that dissolved immediately. The evolution of the concentration of

ACP and RPE was measured by GC. The samples taken from the reaction mixture regularly were

extracted with 5 times with their volume of ethyl acetate The organic phase is then dried with

molecular sieve (4 Å) prior to GC analysis.


73

Reactions were performed in heptane. The conditions varied and are given in Table 16 in

annex 7.2. The reaction started when the prepared biocatalyst (see 3.1.2.1) was added to the

reaction mixture (heptane, ACP and ISP) followed by water.

The reaction of acetophenone with isopropanol catalysed by LBADH and NADP was

investigated in two media that involved high-pressure: in propane, the enzyme and coenzyme being

co-immobilised and in a biphasic system water/propane (enzyme and coenzyme dissolved in the

aqueous phase). Reactions catalysed by immobilised biocatalyst were run in dense propane.

1.5 mmol of acetophenone and 25 mmol of isopropanol reacted at 3 MPa and 30 °C in propane

inside the “high-pressure view cell”. It was checked before the reaction that it forms a

homogeneous mixture. The water activity was set with Na2SO4 10/0 (Zacharis et al. 1997). The

hydrate Na2SO4.10H2O can release water while the anhydrous Na2SO4 can absorb some.

Equilibrium took place so that the activity coefficient of water was buffered at 0.8.

In the case of the reactions in the biphasic system water/dense propane the reactions were

catalysed by 25 mg LBADH, 10 mg NADP dissolved in water, to which 7 g isopropanol and 0.18 g

acetophenone was added and, then, liquid propane. The temperature was set at 30 °C and the

pressures at 30 bar. The concentration of ACP and RPE was measured in the propane phase.

3.1.2.4.ADH activity test

The activity of ADH solution was measured according to the protocol given by the

provider. The method is based on chemical reaction presented at the first line on Figure 26. The

disappearance of NADPH in the presence of acetophenone and ADH was quantified by UV-Vis

spectrometry. The molar extinction coefficient of NADPH is 6220 M-1cm-1 at 340 nm while NADP

does not absorb at this wavelength. 20 µL of NADPH (9.5 mM) was added to 970 µL of a solution

of acetophenone (11 mM) in phosphate buffer (pH = 7) containing 1 mM of Mg2+ at 30°C. The

reaction started when 10 µL of a solution containing the enzyme was added to it. The absorbance

was regularly measured over the first minute. The activity of ADHs is given in the unit U. 1 U

corresponds to the disappearance of 1 µmol/min of NADPH.

The measurement of the stability of an enzyme preparation consists of measuring its

remaining activity after incubation in certain conditions. Hence, samples from the enzyme

preparation are taken regularly and the rest of their activity is measured. When the deactivation of

solid preparation was investigated the samples were dissolved in the appropriate quantity of buffer

before activity test.


74

Figure 31: Example of a test of enzyme

activity

3.1.2.5.Autoclave for incubation of biocatalyst

A set-up was developed for the measurement of the deactivation of LBADH in two media

involving high-pressure, propane and propane/aqueous phase. LBADH was treated in a

thermostated autoclave presented in Figure 32. When the enzyme was dissolved in aqueous

solution, the liquid sample was introduced in the autoclave at the desired temperature. The seal was

then screwed and the propane pumped into the autoclave. After the sample was taken through the

capillary, a small amount of propane was added to maintain the pressure to balance the decrease in

pressure. For the activity measurement of the lyophilised powder, the autoclave was disassembled

in order to take samples.

Figure 32: Autoclave for the measurement of

ADH deactivation in propane

1 liquid propane, 2 aqueous solution of ADH, 3

liquid propane from high-pressure pump, 4 blow

disk, 5 manometer, 6 thermocouple, 7

sampling, 8 seal, 9 magnetic stirrer. It is represented the case of a biphasic system.

In case of biocatalyst under the form of a solid

no stirrer was used.

3.1.3. Analytical methods

The solution were analysed by Gas chromatograph HP 5890 equipped with the integrator

HP 3392A and the column BetadexTM 120 30 m × 0.25 mm, 0.25 µm film (Supelco). The

quantification and the calibration were done using decane as an internal standard.


75

3.2. Results and discussion

3.2.1. Reaction in water

The conditions for this reaction (i.e. concentration of acetophenone, isopropanol and

NADP) were taken from an article dedicated to the optimisation of the conversion of acetophenone

to S-1-phenylethanol by ADH T (Findrik et al. 2005). In this case the concentration of ISP was

very high so that it acted not only as a sacrificial substrate but also as a cosolvent that allowed the

dilution of ACP, a compound rather insoluble in water. The reaction catalysed by LBADH (Figure

33) exhibited a very good yield (98.5 %) and enantioselectivity (RPE was enantiopure). The initial

rate is high (0.25 mmol/(L.min), 12.5 µmol/min, or 0.68 µmol/(min.mgE)). The total turnover

related to the enzyme was TONE = 150 and related to the coenzyme TONCoE = 50. We cannot

conclude from this experiment that the TON were limited to those values neither they could be

higher.

Figure 33: Conversion of ACP to RPE in water

3.2.2. Preliminary test in heptane

We decided to use heptane as a solvent because it was shown that this solvent is suitable

for bioconversion with coimmobilised ADH and coenzyme (Snijder-Lambers et al. 1991; Yang et

al. 1993). Reactions at high-pressure require expensive equipment and special safety conditions so

the early development of the co-immobilisation with usual organic solvent is preferable. The

reactions run in heptane are presented in Table 16 of annex 7.2. As an example the bioconversion is

shown below in Figure 34.


76

Figure 34: Reaction in heptane with co-immobilised catalyst (reaction G)

In each case RPE was enantiopure, but the conversion was limited (small yield) as well as

the total turnovers (<100). The constant K was calculated for each reaction, according to the

equation:

][*][][*][

ISPACPacetoneRPEK

Equation 13

K did not reach a limit, Keq (see 2.2.1.3) and as the co-immobilised catalyst could not be

reused for a second conversion it was deduced that the TONs were probably limited because of the

deactivation of the enzyme.

The biocatalyst required to be co-immobilised to show some activity. Mixed powders of

LBADH and NAPD without any further preparation led to no conversion. Such an experiment was

also performed in dense propane and gave the same result. This confirms the need for a delicate

preparation of the catalyst for reaction in medium where they are not soluble.

3.2.3. Reaction at high-pressure

The reaction of acetophenone with isopropanol catalysed by LBADH and NADP was

investigated in two media : in propane, the enzyme and coenzyme being co-immobilised and in a

biphasic system water/propane (enzyme and coenzyme dissolved in the aqueous phase).


77

3.2.3.1.Reaction in propane with co-immobilised catalyst

When the reation is run with catalysts immobilised on a solid carrier it is important that the

reaction medium presents only a single liquid phase. In case there is no good solubility of the

reactants and products in the solvent, a second liquid phase could provoke the deactivation of the

enzyme by a more polar phase. Indeed, it could strip off the essential water or inhibit the enzyme

by a too high concentration of the substrate. Mixtures with the substrates and products at different

concentrations with propane at 30°C and 30 bar were observed in the view cell (described in

3.1.2.2) in order to check how many phases were present (see results in annexe 7.3 Table 17). In

every case the miscibility was perfect and no second phase was seen except for a gaseous at low

pressure.

The water activity was buffered –set at a certain activity- using salt hydrates. The water

activity can be set at 0.8 using Na2SO4.10H2O and Na2SO4 ((Zacharis et al. 1997) and more

generally (Halling 1992)). This high water activity could influence the miscibility between the

different substances: a water-rich phase could appear. However such a second phase was not

observed.

The catalyst and the salt pair were placed separately in folded filter paper into the view

cell. The initial rate in propane when enzyme and co-enzyme were co-immobilised was about

0.05 µmol/min/mgenzyme . As an example for comparison, it was found with TBADH (ADH from

Thermoanaerobium brockii) initial rate up to 0.15 µmol/min/mgenzyme at 25°C for the reduction of

2-pentanone in hexane after optimising the water activity (Jönsson et al. 1998). The comparison is

interesting because the activity of the lyophilised ADH powder used for preparation of the catalyst

were comparable, 7.3 U/mg in their case, 8 U/mg in our case. So the recovery of LBADH activity

in propane after immobilisation (by comparison with the activity test presented in 3.1.2.4, so in

water) was less than 1 %. Low recoveries in organic media of the same order of magnitude were

found for YADH (ADH from yeast) but about 100% with Horse liver ADH (HLADH) (Snijder-

Lambers et al. 1991). The TON of our reaction related to the enzyme was 180 and related to the

coenzyme 50 and the yield was 45 %. To compare, other studies showed that after an optimised

immobilisation LBADH gave TON above 106 in aqueous solvent in a continuous reactor. If certain

studies in organic solvents presented such low order of magnitude for TON, other presented values

as high as 106 (Grunwald et al. 1986; Snijder-Lambers et al. 1991) (both with HLADH). A study of

hydrogenation done in dense gases with a biocatalyst based on ADH have shown limited yield

(Matsuda et al. 2003). Contrary to our results, the immobilisation of LBADH with a similar method

gave good results with gas phase continuous flow reactor: the TONs were above 106 (Ferloni et al.

2004; Trivedi et al. 2006). As the stability of the enzyme was already an important issue in heptane,

the investigation of the stability of LBADH was undertaken and the results are presented in 3.2.4.2.


78

Figure 35: Bioconversion in propane with immobilised catalyst

3.2.3.2.Reaction in biphasic system propane-water

The bioconversion in propane was run at three different pressures, 3, 10 and 20 MPa. The

evolution of the concentration of ACP and RPE are given in Figure 36, Figure 37 and Figure 38

and these results are summarised in Table 8.

Figure 36: Bioconversion in the biphasic system water/dense propane at 100 bar


79



The syntheses in the biphasic system gave high yield of about 90 %. The initial rate , 0.1-

0.2 µmol/min/mgenzyme., were higher than in the previous case but lower than in water. The yield

was about 90 %. The fact that the initial rate is lower than in water can be explain easily:

the mass transfer in a stirred single phase is higher than in the biphasic system. The total

turnover was about 80 related to coenzyme and 300 related to enzyme. The yield was satisfactory

but the catalyst could not be reused. LBADH seemed to have a limited stability in the biphasic

system. That’s why the study of its deactivation in this medium was needed.


80

Table 8: Result of the three conversions run in biphasic systems.

Pressure (bar) 30 100 200

Temperature (°C) 30 30 30

Yield (%) 93 Not determined 85

Enantiomeric excess (%) >99 >99 >99

µmol/min 5 3.6 3.3 Initial rate

µmol/(min/mgE) 0.2 0.15 0.13

Total turnover/ tetramere 320 Not determined 285

Total turnover/ active sites 80 Not determined 80

Total turnover/ coenzyme 100 Not determined 93

3.2.4. Deactivation of LBADH

The deactivation of LBADH was measured at 4 different conditions, in a water solution, as

a native powder at atmospheric pressure and in dense propane, finally in the biphasic system

water/propane.

3.2.4.1.Deactivation of “untreated” LBADH

The deactivation of LBADH was measured in an aqueous phase and at atmospheric

conditions. The plot of the deactivation of the enzyme, the remaining activity according to the

incubation time is presented in Figure 39 and Figure 40 , respectively.

Figure 39: Deactivation of an aqueous

solution of LBADH at atmospheric

pressure and 36°C.

Figure 40: Deactivation of LBADH in

powder form at atmospheric conditions.


81

The half-life of the enzyme in a phosphate buffer (pH = 7) was about 1.5 hours at 36 °C.

When lyophilised LBADH powder was incubated at atmospheric conditions a fast deactivation

occurred at first but its activity remained almost constant at 6 U/mgenzyme after 20 hours at 30 °C.

The activity decreased fast at 40 °C and, after 7 hours, more slowly. The deactivation is faster

when the enzyme is dissolved into water than when treated as a powder. Solvation in water gives

enzymes more plasticity and LBADH can unfold and deactivate more rapidly. Lack of water

renders its unfolding harder (Fágáin 1995), the good stability LBADH possesses as a solid is

exploited for storage and gas phase processes.

3.2.4.2.Deactivation of LBADH in propane

Figure 41: Deactivation of the preparations

of LBADH in propane at 30 bar.

The deactivation of LBAHD in

propane is shown in Figure 41. The trend of

this deactivation was similar to the

deactivation at atmospheric condition. A fast

decrease took place at first and was followed by a quasi steady remaining activity. This special

phenomenon was observed when lyophilised LBADH preparation was incubated at atmospheric

pressure (Figure 40) or in dense propane (Figure 41). After a fast deactivation, a portion of activity

remained. Such a residual activity is usually explained by the presence of different enzyme

populations with different stability. The work by Rees and Halling seems to provide a good

explanation for the presence of the residual activity (Rees et al. 2001). They elegantly demonstrated

that “the lyophilized powders contain different populations of protein molecules. Some are

relatively exposed, whereas others are protected by contact with other protein molecules and/or

other components of the biocatalyst powder”. Enzymes inside the particles are less sensitive to

deactivation than those on surface. Finally, the most relevant to the description of stability of non-

immobilised LBADH is the slope at the beginning of deactivation test and not the residual activity.

We noticed that LBADH (at the surface of the lyophilised powder) is prone to deactivate whether it

is in dense propane or at atmospheric conditions.

3.2.4.3.Reaction in biphasic systems propane-water

The deactivation of the enzyme was measured in the biphasic system water/dense propane

(Figure 42). The half-life of an aqueous solution of LBADH in phosphate buffer in contact with

dense propane was 6.5 hours at 29°C, and 0.2 hours at 35°C. At 40°C, the second sample of the


82

enzyme solution taken after 25 minutes presented no activity. When comparing with the result

obtained in aqueous phase we conclude: the stability of the enzyme in an aqueous buffer was

lowered when propane was added. The enzyme was not only deactivated in an aqueous solvent but

also at the interface with propane (Halling 1994). The deactivation of LBADH in biphasic systems

at 4°C was the subject of another study and half-life from 1.5 hours with dicloromethane up to

2000 hours in tert-buthyl methyl ether. Indeed, the stability was enhanced by this solvent.

Cyclohexane and nonane gave half-life of 4 and 9 hours, respectively (Villela Filho et al. 2003).

This is comparable with the fast deactivation in water/propane medium. The deactivation of ADH

at the interface between aqueous solution and organic solvent is usually fast and a method for

overcoming it is the use of membranes that prevent the contact between enzyme and organic

solvent (Kruse et al. 1996).

Figure 42: Deactivation of an

aqueous solution of LBADH in a

biphasic system with dense

propane at 30 bar.

3.3. Conclusion and future work

Protocols for the testing of the potential of ADHs in dense gases were successfully

developed and applied to LBADH and the conversion of acetophenone to R-1-phenylethanol.

Synthesis and deactivation determination were done in propane, enzyme and coenzyme being co-

immobilised, and in biphasic system water/propane where enzyme and coenzyme are solubilised in

an aqueous phase. LBADH presented a high enantioselectivity but also a fast deactivation in those

media.

Due to this fast deactivation LBADH is probably not a good candidate for performing this

bioconversion industrially. However protocols are ready and other ADH could be tested such as

ADH T or HLADH which stability was shown to be higher in organic solvent.

Resolution of (±)-trans-1,2-cyclohexanediol

83

4. Resolution of (±)-trans-1,2-cyclohexanediol via the formation

of diastereomeric complexes with (+)-tartaric acid followed by

extraction with supercritical carbon dioxide.

4.1. Materials and methods

4.1.1. Materials.

(+)-Tartaric acid (>99.5%, ref. 251380), RacCHD (>96%) was provided by Sigma-Aldrich

Corp. RacCHD (>96%, ref. 29005) and SSCHD (>99% (sum of enantiomers), ref. 29003) were

provided by Fluka. RRCHD (>99%, 421790) was provided by Aldrich. Pure ethanol was provided

by Reanal Ltd, (Budapest, Hungary). Hungary. CO2 (99.5 w/w% pure) was supplied by Linde Ltd.

Perfil 250™ (expanded and milled perlite for use as a filtering aid with specific surface area of 2.9

m2/g) was kindly given by Baumit Co.

4.1.2. Determination of the structure of the co-crystal.

The transparent monocrystals of the co-crystal (1) of TA and (R,R)-CHD were grown by a

slow evaporation at room temperature of an equimolar solution of TA and (R,R)-CHD in water and

ethanol (1:1). The measurement and the calculation of the sructure that lead to the resolution of the

structure were done by Petra Bombics and are presented in annex 7.4.

4.1.3. Supercritical fluid extractor.

The supercritical fluid extractor is presented in Figure 43. A more detailed presentation of

the plant was given in (Simándi et al. 1998). The extractions were run at a semi-preparatory scale

(gram-scale) with extraction pressure from 10 to 20 MPa (25 MPa being the pressure limit of the

equipment). The separator pressure was set 4 MPa and at the temperature of 40°C. Both extractor

and separator had a volume of 25 mL and were thermostated via their heating jackets where

thermostated water circulated. The liquid carbon dioxide at -10°C was supplied to the constant flow

diaphragm Lewa® pump via the by-pass vessel. There were two working modes for the extractor:

in the stand-by mode CO2 remained in the closed by-pass circuit that included the by-pass vessel,

the cooler and the pump (the fluid circulates according to the blue arrows on Figure 43). When the

extraction was due to start, the by-pass valve was closed and the extractor valve opened: this is the

extraction mode. The direction of the fluids in the extraction mode is indicated by the orange

arrows. After the extractor valve the liquid CO2 was heated up to the desired extraction temperature

and, consequently became supercritical if the temperature was higher than Tc. The CO2 expanded.

The pressures inside the extractor and inside the first separator were manually controlled by


84

playing on the control valves 1 and 2. The pressure dropped dramatically at the outlet of the control

valve 1: the separator pressure was set below Pc and CO2 is gaseous. Most of the extracts is

insoluble in gaseous CO2 and precipitated in the separator, mostly, and in the pipe between the

control valve 1 and the separator. The expansion of CO2 imposed the thermostating of the valve

and the heating of the pipe between the control valve 1 and the extractor with a heat gun. Two

supplementary separators kept at atmospheric pressure and room temperature were added after the

main for insuring that no extract powder were released in the atmosphere (only one is represented

on Figure 43). A flow meter measured the flow of CO2 at atmospheric condition before the exhaust

pipe.

Figure 43: Supercritical fluid extractor

On this figure typical conditions of extraction are indicated. The blue arrow indicates the

recirculation of the fluid in the bypass when the extractor is in the stand-by mode, whereas

the orange indicates the direction of the flow of CO2 during an extraction.

The plant involving high-pressure, risk assessment was a permanent concern while

conceiving and using the plant. The temperature of the extractor and the separator cannot exceed a

temperature set on the water thermostat. The equipment was tested by the providers to support a

pressure higher than 25 MPa. The pressure is maintained below 25 MPa by safety blow disk that

burst above this pressure. Two are placed on the plant. The first is in the bypass and in case the

pressure after the pump exceed 250 bar it bursts and the liquid CO2 is released into the bypass

vessel that buffer the pressure. The second is placed before the extractor. The SCCO2 would be

released directly to the exhaust pipe in case the pressure exceeds 25 MPa. This might happen in

case the extractor or the control valve or the pipe between the extractor and the separator get


85

cogged with precipitated extract, for instance or if the operator is too slow to open the control valve

1 in case of increasing pressure.

4.1.4. Resolution of CHD with TA and SFE

The resolution of CHD by formation of a cocrystal with TA followed by supercritical

extraction consists in three steps: the sample preparation, the supercritical fluid extraction and the

raffination by an alkaline treatment. The general principle of an extraction was described in 2.3.4.

4.1.4.1.Sample preparation.

For a typical sample due to be extracted by SFE, i.e the sample in this work (unless the

contrary is stated) and also the samples used in previous works (Molnar et al. 2008; Molnár 2009),

were prepared according to this protocol: 1g of CHD was dissolved in 15 mL ethanol and 0.646 g

TA, as well. These two solutions were mixed and then 1.5 g Perfil 250tm might be added to it.

Ethanol was evaporated at 40-50°C in a rotoevaporator at a reduced pressure of about 160 Torr and

the solid was scratched out from the round-bottomed flask with a spatula and let to dry out

overnight at room temperature and atmospheric pressure in an open Petri dish.

The samples of the binary mixtures of TA and enantiomeric CHDs were prepared by

evaporation of the solvent from the ethanolic solutions of CHD and TA at reduced pressure (about

160 Torr) and 40-45°C when no other condition is specified.

4.1.4.2.Supercritical fluid extraction.

The extractions performed on a sample as described below were carried out at pressure

from 100 to 200 bar and temperature from 33 to 63°C with an average flow of CO2 of about 20

g/min and the total weight of CO2 was 470 g. It was checked by a second extraction in the same

condition but with 180 g of CO2 that the “extractable” part of the CHD which mostly corresponds

to the uncomplexed (S,S)-CHD had been extracted. Indeed only a small amount, inferior to 20 mg,

was extracted in this second step. When other conditions are applied to the extraction they are

mentioned.

The yield of extraction of Yext was calculated as the loss of weigh of extractor during the

considered extraction divided by the weight of CHD initially put into the extractor. In order to

allowg a comparison between the different extractions, the amount of CO2 is given as the weight of

CO2 divided by the weight of CHD contained initially in the extractor. This value is CO2rel defined

as:


86

Equation 14: definition of the relative weight of CO2 during an extraction

ww

initial

CHD

CrelCO 022

4.1.4.3.Raffination by alkaline treatment

A schematic presentation of the alkaline treatment of the residuum, the remainder of the

sample after extraction, is shown on Figure 29. The raffinate was removed from the extractor and

40 ml of methanol was added to it, in order to dissolve the remaining diastereomeric complexes.

After 1 h of stirring, the inert support was filtered and the methanol was evaporated at 40-50 °C in

30 mbar vacuum. The complex was decomposed by a saturated aqueous solution of Na2CO3 (6 ml)

during the stirring for 15 min. The water then was evaporated from the residue at 70–80°C in

30 mbar vacuum. The solid particles were crushed and 20 ml of trichloromethane was added to it.

After 30 min of stirring, the sodium-tartrate salt was filtered out and the organic phase was

evaporated. The product obtained contains the RRCHD in excess.

The yield of raffination Yraf was calculated as the weight of CHD isolated after the

rotoevaporation of chloroform divided by weight of CHD initially put into the extractor.

4.1.5. Analytical methods

The GC analysis were done with an Agilent 4890D chromatograph using Hydrodex-β-6-

TBDPM column (25 m × 0.25 mm × 0.25 µm film with permethylated β-cyclodextrin, Macherey &

Nagel, No.: 21519/11). The analysis was performed at isotherm conditions (130 °C), the carrier gas

was helium, the split ratio 1:50, detector: FID at 250 °C, injector temperature at 250 °C.

FTIR spectra of cocrystal (1) was measured by Excalibur Series FTS 3000 (Biorad) FTIR

spectrophotometer in KBr between 700 and 4000 cm-1.

Powder X-ray diffraction patterns were recorded with a PANalytical X’pert Pro MDP X-

ray diffractometer using CuKα and Ni filter.

Differential scanning calorimetry (DSC) measurements were performed using a Modulated

DSC 2920 device (TA Instruments). The samples (1-2 mg) were measured in sealed Al-pans at a

heating rate of 10 K/min. Simultaneous thermogravimetry and differencial thermal analysis

(TG/DTA) tests were conducted using an STD 2960 Simultaneous TG/DTA equipment (TA

Instruments), a heating rate of 10 K/min, open Pt crucibles and an air purge of 130 ml/min.


87

4.2. Results and discussion

4.2.1. Characterisation of the cocrystal

4.2.1.1.The structure of the co-crystal.

The crystal system of co-crystal 1 is orthorhombic, the space group is P212121 (No 19),

having one TA and one CHD molecules in the asymmetric unit (Z=4, Z’=1) (Figure 44). Detailed

crystallographic data, beyond the parameters of data collection, structure determination and

refinement are presented in Table 9. The orientation of the molecule is given by RRCHD and

(2R,3R)-(+)-tartaric acid which were the starting material. Both alcoholic oxygen atoms in the

CHD are in equatorial positions. The hydrogen atomic positions of the alcoholic and acidic OH-s

were determined by difference Fourier calculations (for measurement and calculation see 7.4)

Figure 44: ORTEP diagram (Spek 2003) of the CHD-TA co-crystal (1)

Represented at 50 % probability level, heteroatoms are shaded. The chiral centres are C2

R, C3 R, C21 R and C26 R, respectively.


88

Table 9: Summary of crystallographic data, data collections, structure determination and

refinement for CHD-TA co-crystal (1).

Formula C6 H12 O2, C4 H6 O6

Formula Weight 266.24 Crystal System Orthorhombic Space group P212121 (No. 19)

a, b, c [Angstrom] 6.7033(13) 7.2643(16) 24.863(5) V [Ang**3] 1210.7(4)

Z 4 D(calc) [g/cm**3] 1.461 Mu(MoKa) [ /mm ] 0.128

F(000) 568 Crystal Size [mm] 0.20 x 0.20 x 0.20 Temperature (K) 295

Radiation [Angstrom] MoKa 0.71073 Theta Min-Max [°] 3.2, 26.4

Dataset -8: 8 -9: 9 -31: 31 Tot., Uniq. Data, R(int) 25226, 2468, 0.121

Observed data [I > 2.0 sigma(I)] 2216 Nref, Npar 2468, 170 R, wR2, S 0.0640, 0.1369, 1.05

Max. and Av. Shift/Error 0.00, 0.00 Flack x -0.5(18)

Min. and Max. Resd. Dens. [e/Ang^3] -0.26, 0.28 * w = 1/[\s^2^(Fo^2^)+(0.0222P)^2^+1.1376P] where P=(Fo^2^+2Fc^2^)/3

The two moieties of the co-crystal contain six donors and eight acceptors of hydrogen

bond. Thus a rather complex hydrogen bonding pattern (Table 10) is constructed in the crystal

structure (Figure 45). A two dimensional sheet is formed in the ab crystallographic plane with the

width of c/2. The 2D hydrogen bonded sheet is like a “double sandwich”. The inner part is

constructed from TA-s (Figure 46). It is “covered” on both upper and bottom side by CHD-s

(Figure 45a and b). Thus the inner part is hydrophilic, the outer coat is hydrophobic.


89

Table 10: Intermolecular interactions in the crystal structure of CHD-TA co-crystal (1).

D-H...A D-H (Å) H...A (Å) D...A (Å) D-H...A (°) symmetry operation

O2 -H2O ...O12 0.8200 2.3000 2.615(3) 104.00 Intra O3 -H3O ...O41 0.8200 2.3800 2.707(3) 105.00 Intra O12-H12O...O21 0.8200 1.7700 2.593(3) 176.00 Within asym unit O26-H26O...O11 0.8200 2.1300 2.923(3) 162.00 Within asym unit O2 -H2O ...O26 0.8200 2.1000 2.873(3) 157.00 x,-1+y,z O3 -H3O ...O41 0.8200 2.0200 2.740(4) 146.00 1/2+x,-1/2-y,-z O21-H21O...O3 0.8200 1.9700 2.786(3) 171.00 1/2+x,1/2-y,-z O26-H26O...O41 0.8200 2.5000 2.882(3) 109.00 1/2+x,1/2-y,-z O42-H42O...O26 0.8200 1.9900 2.773(3) 160.00 -1+x,-1+y,z C2 -H2 ...O21 0.9800 2.3900 3.280(4) 150.00 -1+x,y,z

There are two intramolecular hydrogen bonded loops in TA stabilizing the conformation of

the molecule: …O12-C1-C2-O2-H2O… and …O41-C4-C3-O3-H3O… both are S(5) by the graph set

analysis (Grell et al. 2000). Within the asymmetric unit the two constituents are hydrogen bonded

forming a R22(9) homodromic loop: …O11=C1-O12-H12O…O21-C21-C26-O26-H26…. A further

hydrogen bonded loop exists between CHD and TA: …H21O-O21-C21-C26-O26-H26O…O41=C4-

C3-O3… which is heterodromic R22(10). In the inner part of the “sandwich” the TA molecules are

connected by strong intermolecular interaction to each other (Figure 46) and to CHD molecules.

There is no hydrogen bond between CHD molecules. One TA molecule within the sheet is

connected to four other TA-s, to three TA-s directly, to one TA via a CHD. One TA molecule is

connected to four CHD molecules with six hydrogen bonds. The hydrogen bond loops are: …H3O-

O3-C3-C2-O2-H2O…O26-H26…O41… R23(9) and …O41-C4-C3-O3-H3O…O41=C4-O42-

H42…O26-H26… R33(11). Finally, there is a weak C-H…O type interaction also, C2-H2…O21 within

the sheet.


90

A

B

C

Figure 45: The two dimensional infinite hydrogen bonded plane of the CHD-TA co-crystal

(1).

View from the a, b crystallographic axis, respectively, are side views, while view from the

c crystallographic axis is a perpendicular view to the sheet. TA is coloured red, while CHD

is blue. Hydrogen atoms are omitted for clarity(Macrae et al. 2006).


91

Figure 46 : The inner TA layer of the sheet

(Macrae et al. 2006) presenting its

hydrogen bonding system of co-crystal 1.

View from the c crystallographic axis.

Hydrogen atoms are omitted for clarity.

The experimental powder X-ray diffraction pattern of sample with 1:1 molar ratio of (R,R)-

CHD and (R,R)-TA has been checked with comparison with the simulated powder diffraction

pattern of co-crystal CHD-TA (CoC) generated from the single crystal data presented above: the

agreement was good as shown on Figure 47.

Figure 47: Theoretical (in black) and

experimental (in red) diffraction pattern of

CoC

4.2.1.2.Characterisation of the co-crystal TA-RRCHD

The FT-IR spectrum of the co-crystal (1) has been measured and presented an interesting

feature (Figure 48). A splitting of the single carboxyl carbonyl C=O stretching vibration (at 1740

cm-1) of pure TA, occurred in to two C=O bands at 1738 and 1698 cm-1 in the spectrum of co-

crystal 1. The former band (1736 cm-1) belongs certainly to the C4O41O42H42 carboxylic group

despite its carbonyl oxygen O41 is involved in three of relatively strong hydrogen bonds as

acceptor, nevertheless its H12 proton is kept also relatively strongly, based on donor – acceptor


92

distances (Table 2). The latter band (at 1698 cm-1), meanwhile, should belong to C1O11O12H12

carboxylic group, whose O11 oxygen is involved only in a weak hydrogen bond, and the H12

proton is loosely kept and very intensely shared with O21 oxygen of O21H21 hydroxyl group of

CHD. The quasi-anionic feature (which stabilized with strong intramolecular hydrogen bond of

O12 to O2 of O2H2 hydroxyl group) resulted in large decrease of carbonyl frequency.

Figure 48: FTIR: spectrum of the CHD-TA co-crystal Coc.

The DSC analysis of CHD-TA co-crystal (1) in sealed Al-pan showed a melting point at

133.2°C and its enthalpy of fusion was 56.7 kJ/mol.

Figure 49: DSC melting peak of the pure

CHD-TA co-crystal CoC in sealed Al-pan at

10°C/min (mass 2.59 mg).


93

Figure 50: Simultaneous TG/DTA curves of the

pure CHD-TA co-crystal 1

(open Pt crucible, air flow of 130 ml/min, heating

rate 10°C/min, mass 8.68 mg).

In an open Pt crucible of the simultaneous

TG/DTA apparatus, the co-crystal 1 shows the same

melting point (133.2 °C). Mass loss of 3 % in the TG

curve shows some sublimation of CHD from 100°C to

the melting point. A further evaporation of CHD seems

to be overlapping with the decomposition process of

tartaric acid above 170°C.

Remark: Growing crystal from mixture of SSCHD and TA gave monocrystal large enough to be

submitted to single crystal X-ray diffraction measurement. Only SSCHD enantiomer and TA

crystals were found.

4.2.2. Decomposition of the CoC in situ

It was shown previously many disadvantages of the technique of decomposition of the

residuum of the extraction that consists mostly in CoC and perfil (this will be demonstrated in

4.2.3.1). It was intended to decompose the cocrystal “in situ” in a two step extraction. The first step

is exactly the same as in 2.3.4 and follow by the second step that is run at higher pressure and

temperature. The idea was that at a temperature and pressure that are high enough CoC will

decompose by releasing RRCHD that is extracted by CO2, TA being insoluble in SCCO2. A

cocrystal involves less strong interaction between the two molecules than a salt. That’s why such a

method of decomposition seemed possible. Also an experiment in a view cell (cf Figure 30)

allowed observing the behaviour of a monocrystal of CoC in SCCO2. At 93°C and 20 MPa the

shape of the crystal was changing. This indicates that the RRCHD is extracted from it and that TA

forms from CoC(this was demonstrated by XRD).

22

SCCOSolidSCCO RRCHDTACoC

Equation 15: decomposition of CoC in SCCO2.

Experiment in the view cell was a visual proof for the reaction written above and it was

decided to test this by an extraction. This was proven possible by a preliminary work at an


94

extraction in two steps (curve 1 in Figure 51) : the sample was prepared in the condition of 4.1.4

but with 50% more CHD, TA, perfil and ethanol (solvent), ie in the molar ratio 0.5 with 1.5 g of

CHD. The first extraction step was run at 33°C and 100 bar. Y1 was equal to 43% with ee=58%.

After taking the sample and rinsing the pipes the extraction (second extraction step) is continued at

a higher pressure and temperature, 20 MPa and 93°C. The yield of the second extraction (Y2) is

calculated as Y1 is. Y2 was found equal to 52% and ee2=54%. This result was promising in the

sense that the overall yield of extraction is high (95% whereas it is about 85% with the raffination

step) even if the ee2 is lower than result found with the raffination method described before.

Figure 51: Extraction curves at different temperature and pressure, study of the

decomposition of CoC in situ.

This experiment was reproduced but with an extraction performed at 15 MPa and 48°C cf

Figure 51 curve 2). Y1 was equal to 52%. The second extraction step was performed at temperature

and pressure of 95°C and 200 bar, both the maximum values possible within the safety range of the

equipment. The result was very good: Y2=43% and eeext2=91%. This method of decomposition is a

clear improvement. The Y2 is higher than Yraf and eeext2 higher than eeraf. Consequently, in this

case a higher F parameter is obtained F=0.68 (while in the middle point F=0.54±0.025). The fact

Yext2 is higher than Yraff. The first reason is that the raffination of the residuum had probably

quite a low yield, ie much CHD is lost over the process. The second is that in the case of the


95

raffination by alkaline treatment Yraf is calculated on the base of the weight of CHD isolated

contrary to the Yext2 which is calculated the same way as Yext, based on the decrease in weight of

the extractor. This is a “maximised” yield because it does not include all the lost of extract inherent

to any extraction process and so does not represent the quantity of CHD isolated. ee2 is higher than

eeraf probably because the sublimation of CHD might be stereoselective and induce a variation of

ee. Indeed, the raffination includes an evaporation of water at reduced pressure and temperature as

high 80°C and CHD is known to sublimate easily. It was discovered that many substances’

enantiomeric excess varies during sublimation(Fletcher et al. 2007). The sublimate usually presents

a higher enantiomeric excess and, consequently, the residue’s ee decreases. The reason might be

that the excess of enantiomer crystallises in a form that retains the molecule less by comparison to

the racemic. Another explanation is that the molecule in the gas phase forms clusters which are

more or less stable if the molecule presents the same orientations or not (Borho et al. 2001; Perry et

al. 2007). 1,2-diols structure is favourable to the formation of dimers which stability is gas phase is

different if they are formed of twice the same enantiomers (RR and RR or SS and SS) or of one

another (RR and SS). (Interestingly, this might be another explanation for the origin of asymmetry

in life: amino acid brought by meteorite would present an enantiomeric enrichment by sublimation

in the atmosphere.)

The second step of the extraction that corresponds to the decomposition of CoC was tested

at 20 MPa and two other temperatures, 73 and 83°C, see extraction curves 3 and 4 in Figure 51.

The extraction took place at a very similar rate in those two last cases whereas at 93 °C the

extraction was completed with less solvent. Both extractions were achieved in high yields, 96 %

and >99 %.

The advantages of SCCO2 were applied here for the decomposition of CoC in situ. The

insolubility of TA in CO2 is an advantage here again: because other solvent, as chloroform or

water, are able to decompose CoC but they also dissolve a part of TA. High yield were obtained.

The ee of the second extract are rather high but the eeext are still limited. Can analytical method

explain the limitation on ee1 and ee2? Which phases are encountered during the extraction process?

Remark: Even if SCCO2 is a privileged solvent for its “greenness” (see in the Annex 7.1.3) other

were tested. The solubility of CHD is so low in hexane that too much would be necessary. Diethyl

ether and chloroform posed two problems, they presented little selectivity and they dissolved a

fraction of TA.


96

4.2.3. Description of the extraction

4.2.3.1.Monitoring the evolution of the content of the

extractor by XRD and fractionning

The different constituents of the material present in the extractor are distinguishable by

XRD but their precise quantification is not feasible and amorphous phases are invisible. The

experimental diffractograms of the different pure substances are given in Figure 52.

We wanted to take advantage of XRD for understanding what occurred during an

extraction better. This experience consisted in taking sample of the material partially extracted to

see which phase got dissolved and taken away from the extractor with CO2. The extraction curve is

plotted in the Figure 53 at each points of the extraction where a sample is indicated on this figure

the extractor was dismantled, the material left into the extractor was mixed and consequently

correspond to the average content of the extractor and 2 samples were taken in order to limit the

problem due to sampling. They were similar in every case this is why only one diffractogram was

presented in Figure 55 for each sampling. As the samples taken for XRD were a significant part of

the material to be extracted, some CHD is not extracted, lost with the sample for XRD and the

aspect of the extraction curve is modified, actually flattened at the end. That’s why a corrected

curve was plotted (Figure 54) that shows the course of the extraction as if no sample had been

taken.


97

Figure 52: Experimental powder pattern of the compound involved in the resolution

system.

ANA 12 is defined in Table 18 of the annex 7.6. It presents the peak of the “X” compound

whose case is treated in 4.2.4.3.

Figure 53 : Loss of

weight of the extractor

according to the weight

of CO2 and sampling.

The spline line is only

indicative.


98

Figure 54: Theoretical

loss of weight of the

extractor if no sample

had been taken. This

figure does not show

more information than

the previous but has

the advantage to show

which aspect the

extraction curve would

have if no sample had

been taken.

The sample to be extracted is done as in 4.1.4, mr=0.5 but with 1,5g CHD. One initial

sample was taken: Sample 0 from the material to be extracted. In a first step, the sample was

extracted with CO2 at 200 bar and 33°C. This corresponded to the best extraction condition as

demonstrated in the previous work (Molnar et al. 2008). 2 samples were taken, the first (Sample 1’)

at the bottom of the extractor where CO2 had entered the extractor first, as it went up in it and the

other after mixing the content of the extractor (Sample 1). After reassembling the extractor and

pursuing the extraction another sample, Sample 2, was taken the same way as Sample 1. A last

sample was taken for the first step of the extraction when the decrease in weight of the extractor,

that is equivalent to the quantity of extracted CHD, was smaller than 20 mg CHD for 200g of CO2:

it corresponded to the end of the first step of the extraction.

The complex was decomposed by a second extraction as in 4.2.2 (95°C and 200 bar). Two

samples were taken in the course of this extraction, the first in its middle of the second extraction

and the other at the end when 360 g of CO2 extracted less than 10mg of CHD.


99

Figure 55: Diffractograms of the different samples from the material inside the extractor

over the extraction.

The origin of the most important peaks is indicated. When two peaks cannot be

distinguished both origins are given.

The analysis of the sample results allowed a good understanding of the phenomenon taking

place over the extraction. The different diffractograms are shown in the Figure 55, where the

substances responsible for the different peaks are indicated.

The initial material (sample 0 before extraction) contained enantiomeric CHD, (S,S)-CHD,

the cocrystal plus racemic CHD and traces of other material. The presence of racemate indicated

that the cocrystallisation of TA with CHD was partial only because a total cocrystallisation would

lead to only CoC and SSCHD (enantiomer). Consequently, the preparation of the sample is an

important factor that lowered the ee of the first extract and will be treated in more details later.

Over the first extraction step the enantiomer, SSCHD, and the racemate were dissolved and

withdrawn by CO2 as the progressive decrease in the intensity of their peaks indicated. In the last

sample, sample 3, the peaks referring to the enantiomer and the racemate had completely

disappeared. SSCHD could not be distinguished from racemate by extraction as shown by the

sample 1’ at the bottom of the extractor whose diffraction pattern was similar to Sample 3’s (except


100

there is less TA) that was fully extracted at the beginning of the extraction already. The difference

in solubility between the liquid solution of racCHD and of the enantiomer was not important

enough to allow a separation of those two phases by extraction.

Very little TA can be seen in the sample 0 and the quantity of TA increased over the

extraction. At the end of the first extraction, after the enantiomer and the racemate are fully

extracted, the cocrystal decomposed slowly, leaving more TA which cannot be extracted due to its

insolubility in CO2. The decomposition of the cocrystal continued in the second step of the

extraction faster because of the higher temperature, as it was shown in Figure 51. Indeed, the

amount of TA increased from sample 3 to sample 5 while the cocrystal disappeared until

completion for sample 5 that is actually only composed of TA and perfil. So the rest of a resolution

of CHD by cocrystallisation and SFE can be recycled for a new resolution: the quantity of waste is

considerably low and the resolving agent does not require any expensive regeneration.

Thus the decomposition of CoC started from the first extraction steps on and this can be

compared to the first experiment of decomposition in two steps in 4.2.2 (see Figure 51). During the

first extraction step fractions, whose ee was measured, were taken. The ee are given on Figure 56,

curve 2 and 5. The trend followed what we found by XRD: at the beginning mostly free CHD was

extracted. The ee’s of the fractions taken during the first extraction (excess of SSCHD) decreased

when most free CHD had been extracted because the contribution in RRCHD by the decomposition

of CoC was more important. The extraction of free CHD as racemate or pure enantiomer seemed

total according to the diffractogram as we cannot see their peaks any longer in Sample 3. However,

an incomplete extraction of free CHD in the first extraction cannot be excluded based on XRD.

Indeed the sensibility of this method is limited and only the compounds at the surface of the

powder diffract X-rays are seen; so crystals of free CHD covered up by CoC or TA or trapped into

Perfil pores are likely to be unseen and to be last portion of free CHD to be extracted.

To summarise, the ee of the first extracts was relatively low due to two phenomena. The

first is the only partial formation of cocrystal, some TA remained not cocrystallised, as well as

RRCHD that forms racemate with SSCHD. The second is the slow decomposition of the cocrystal

that was the most visible at the end of the first extraction but probably occurred all over in different

proportions.

Remark: there is a difference between the evolution of the ee of curve 2 and 5. In the first case the

ee did not simply decrease but increased and then decreased. The explanation for such an evolution

might be a fine difference if the solubilities of the two free CHD species: SolCHD and SSCHD.


101

Figure 56: Different fractions during extraction, their enantiomeric excesses

Curve 2 and 3 were fully presented in Figure 51. Only the fraction of the curve where ee

were measured are given. In the case of the extraction 3 the first step of extraction lead to

an enantiomeric excess of 63 %. The ee indicated for certain segment of the extraction

curve corresponds to the ee of this fraction.

The second extraction step has shown good results because the whole cocrystal was

extracted and only TA was left at the end of it (Sample 5). The limitation of the ee of the second

extraction step at about 90% can have several reasons.

A first explanation is that a small amount of free CHD unseen with XRD had not been

completely extracted at the first extraction or, from another point of view, the impurity of SSCHD

found in the second extraction are due the incompleteness of the first extraction. The ee increase

over the second extraction step because the rest of free CHD (mostly RRCHD) is mostly extracted

at the beginning of the second step.

A second extraction is that some other cocrystal is formed between SSCHD and TA or the

co-crystal could also intake a small amount of SSCHD in addition to RRCHD i.e. CoC forms a

solid solution, possibly a lattice compound where a little proportion of RRCHD can be replaced by

SSCHD. This happens very often in the case of diastereomeric salt and a common limitation to this

resolution method (Jacques et al. 1981; Kozma 2002). The reason for the variation of ee over the


102

second step could be the following: The rest of unextracted free enantiomer is extracted before the

complete decomposition/extraction of CoC than at lower temperature and pressure or the fact that a

CoC containing a portion of (or only) SSCHD would be less stable than CoC with pure RRCHD.

4.2.3.2.Improving the enantiomeric excesses by leaving

off an intermediate fraction

The evolution of the ee over the extraction indicates that it is lower at the end of the first

extraction and at the beginning of the second extraction. That’s why it was intended to perform an

extraction similar to the previous except that the intermediate fraction between the first and the

second extract is left off. The first extraction gave a fraction rich in SSCHD whose ee1=75 % and

the second rich in RRCHD ee2=91 %. The intermediate fraction presents a sligh ee of 14 % and

accound for 15-20% of the sample. Removing this small fraction allow an improvement of the ee

but reduces Y1 and Y2 (they could not be calculated precisely.). However it should be noted that

this step does not lead to the extra waste because in an industrial process this intermediate fraction

would be recycled in another resolution as will be shown later in 4.2.5.2. However, the

improvement of ee1 is limited.

4.2.4. Sample preparation

4.2.4.1.Investigation of the binaries RRCHD-TA and

SSCHD-TA. Coroboration by XRD.

The stereoselective formation of a cocrystal is the base of this resolution. The difference in

stability between the two cocrystals of SSCHD-TA and RRCHD-TA will partially determine the

quality of a resolution: if one is much more stable only a small amount of the other is formed and

the resolution is good while if their stability is similar no resolution occurs. At this point we know

that the cocrystal RRCHD-TA is more stable than SSCHD-TA (otherwise the resolution would not

take place). The formation of a cocrystal SSCHD-TA is questionable as we could not grow a

monocrystal of it and, concerning this issue, we have investigated both binary systems between

RRTA and RRCHD or SSCHD with DSC and powder X-ray diffraction. This work served as a

background for thermal investigation of the ternary RRCHD-SSCHD-TA. This work was presented

in an article (Thorey et al.). The melting point and enthalpy of fusion of the initial chemicals

applied in preparation of binary mixtures are listed in Table 11, while the temperatures of the

observed thermal heat effect(s) and the initial crystalline phase composition of various mixtures at

room temperature are given in Table 12 for some examples.


103

Table 11: Melting point and enthalpy of fusion of the applied chemicals

Initial chemicals

Tfus observed by DSC K, (°C)

Hfus measured by DSC (kJ/mol)

Reference Tfus (K) Ref.

(±)-CHD

372 (99) (small pre-melting

endothermic peak at 87°C, as well)

34.4 376.4 (Leitao et al. 2002)

(R,R)-CHD 381 (108) 18.0 382.5 (Leitao et al.

2002) (S,S)-CHD 377 (104) 15.0 Ibid ibid

(R,R)-tartaric acid 444 (171) (decomposes) 22.4 (estimated)

(Martin Britto Dhasa et al. 2007; Stanton et al. 2008;

Takata et al. 2008)

Co-crystal (1) 406 (133) 56.7 This work

Table 12: DSC and XRD data of binary mixtures in the ternary system

Binary mixtures Molar ratio

Eutectic temperature Teu K,

(°C)

Liquidus temperature Teu K (°C)

Crystalline phases present

(XRD)*

1 (R,R)-CHD: (R,R)-TA 3:1 373.3 (100.2) Ca. 396 (123) (R,R)-CHD and

Co-crystal (1)

2 (RR)-CHD: (R,R)-TA 1:3 404.7 (131.6) Ca. 433 (160) Co-crystal (1)

and (R,R)-TA

3 (S,S)-CHD: (R,R)-TA 3:1 359.7 (86.6) - (S,S)-CHD and

(R,R)-TA

4 (S,S)-CHD: (R,R)-TA 1:1 358.6 (85.5) Ca. 410 (137) (S,S)-CHD and

(R,R)-TA

5 (S,S)-CHD: (R,R)-TA 1:3 253.6 (80.5) Ca. 433 (160) (R,R)-TA and (S,S)-

CHD

The XRD profile of the samples formed from the 1:3 and 3:1 binary mixtures of (R,R)-

CHD and (R,R)-TA corresponded to the 1:1 mixtures of co-crystal (1) and (R,R)-CHD or (R,R)-TA,

respectively. These samples showed eutectic melting behavior, eutectic temperatures were found

by DSC lower than the melting point of co-crystal (1): one between (R,R)-CHD and co-crystal (1)

at 100°C (Fig. 8a); and the second between co-crystal (1) and (R,R)-TA at 131.6°C. DSC showed,

in all the cases of the binary (S,S)-CHD – (R,R)-TA system, a constant eutectic temperature of 85-

86°C. It is lower than the melting point of (S,S)-CHD and the eutectic temperature of (1) and TA or

(1) and (R,R)-CHD. This indicates that no co-crystal is formed between (S,S)-CHD and (R,R)-TA.

This result is corroborated by the XRD profile of the sample for the binary mixtures of (R,R)-TA


104

and (S,S)-CHD that presented 1:3, 1:1, and 3:1 only patterns of (R,R)-tartaric acid (PDF No. 00-

033-1883; 00-020-1901; 00-031-1911) and (S,S)-CHD (PDF No. 02-093-3042), and no other

reflections occurred.

After the eutectic melting, an elongated dissolution of the excess phase has been continued,

as it is shown for both the 1:1 molar ratio of (R,R)-CHD and co-crystal (1) or (S,S)-CHD and (R,R)-

TA in Fig. 8a and 8b, respectively. In case of 3:1 molar ratio of (S,S)-CHD and (R,R)-TA, only the

single eutectic melting effect occurred, i.e. it represents almost a eutectic composition between

(S,S)-CHD and (R,R)-TA.

Figure 57: DSC curve of binary mixture

corresponding to 1:1 molar ratio of a) (R,R)-

CHD and co-crystal (1) and b) (S,S)-CHD and

(R,R)-TA, both exhibiting eutectic melting

behavior.

For this kind of study, DSC is very

efficient. Indeed only few scans were sufficient to

demonstrate that no cocrystal and, moreover, are

corroborated by XRD.

It is possible to represent the binary phase

diagram TA-SSCHD: see Figure 58 with the

calculated liquidus and eutectic temperature (for

detail see the related article (Thorey et al.)). The

eutectic temperature is well defined and rather constant but the liquidus is a bit low compared to

the calculated value. It can be argued that the rather high experimental value of the liquidus is due

to the fact that SSCHD sublimate easily so that a part of it is lost for dissolving TA so that the point

should be actually shifted to the left as if there was less SSCHD in the sample.

Figure 58: Melting binary phase

diagram SSCHD TA.


105

a)

b)

Figure 59: The problematic melting point phase diagram of RRCHD-TA.

a) Binary RRCHD and TA with formation of CoC. First set of data with the interpretation

with two eutectic points.

b) More data with the interpretation with a eutectic point (RRCHD-CoC) and a peritectic

point (CoC-TA). The open circle corresponds to the end of the melting of the unique peak

of the DSC scan of CoC.

An example of DSC analysis will be given for XTA = 84% in Figure 61: DSC analysis of a

sample of composition RRCHD:TA 16:84.

A binary diagram RRCHD-TA with two eutectics, RRCHD-CoC and CoC-TA,

corresponds to the interpretation given above and is presented in the Figure 59 (a) with the

calculated liquidus and eutectic point. The value of the eutectic melting temperature of the binary

CoC-TA is very close to the melting point of pure CoC and the fitting is generally bad. An

explanation can be similar to the one given for the binary SSCHD-TA: the sublimation of CHD

causes its non interaction with TA and so suppresses the eutectic melting between TA and CoC. So

we added some points to the binary to confirm the trend. The points are given in the Figure 59 (b).

Another interpretation of the binary RRCHD-TA can be based on the presence of a peritectic point

between CoC and TA or incongruent melting of CoC rather than a eutectic. This could explain why

the first melting in the CoC-TA region equals the melting of CoC: this melting would correspond to

the decomposition of CoC. However this explanation has its flaws, especially regarding the melting

of CoC that is neat and so does not seem to take place parallely to the crystallisation of a fraction

TA followed by its dissolution into the melt. It is also possible that transitions are missing by the

kinetic of the transition that ends with an amorphous phase unable to crystallise. For instance, if the

interpretation with a peritectic point is exact it is possible that the crystallisation of TA from molten


106

CoC is difficult and so is of a limited extend. The presence of amorphous and metastable phase will

be raised in the next part.

The thermoanalytical study allowed us to determine that no cocrystal was formed between

TA and SSCHD, while a complete knowledge of the binary RRCHD-TA revealed complicated and

subject of different interpretations. Complex issues such as amorphicity and sublimation arise.

4.2.4.2.Investigation of the cocrystallisation by XRD,

ternary phase diagram.

The goal of this study is to give some tool for the fast determination of suitability of a

sample for resolution of CHD and also to know more about the limitation on eeext. We have shown

that the limitation was of two types: the first is the slow decomposition of CoC, the second the not

total formation of CoC. The question we desire to answer here is: is the partial cocrystallisation is

due to thermodynamical reasons or rather kinetic? In case the ee is thermodynamically limited the

low ee cannot be overcome whereas if the reason is kinetic more appropriate conditions of

crystallisation can lead to improvements and effort are worse spending on this issue. The other

reason why the understanding of the phase diagram is important is to know whether a non-racemic

CHD can be resolved a second time for enantiomeric enrichment.

Examples on the analysis of diffractograms were presented in 4.2.3.1 and 4.2.4.1. The

spectrum of CoC fits with the simulated powder pattern simulated (Table 6) well. It should be

noted that the racemate has a structure that fits with the metastable racemic CHD (ZZZKPE06) and

not with the more stable structure. The racemate CHD corresponds to the liquid solution that is

formed between the enantiomers for ee<60 % according to Leitao’s work (Leitao et al. 2002) (this

will be again verified on sample ANA4 and ANA5 or ANA19 in Annex 7.6). It is possible that the

less stable polymorph of racCHD is the only which can accept an ee, this is to say that can form a

solid solution. In Sample 0 (Figure 55) the solid solution is in equilibrium with the SSCHD, so the

solid solution is saturated with SSCHD and if Leitao’s values for the limit of the solid solution is

good the solid solution contains mostly SSCHD (ee=60%) and consequently the free enantiomer

(solid solution + pure enantiomer) has an enantiomeric excess superior to 60 %. A more detailed

presentation of the solid solution that forms RacCHD is given in annex 7.5.2.1 because of it

implication for the ternary phase diagram. The phase diagram presented in Figure 60 a) is the more

favourable to the resolution of CHD by TA because CoC is formed in every part of the diagram and

the sample for mr=0.5 is composed only of CoC and SSCHD. We call this type of ternary phase

diagram “solid solution and CoC>solCHD” because it includes the formation of a solid solution

and CoC is always more stable than SolCHD. We thought that the partial crystallisation can be


107

explained by a ternary phase diagram between TA-SSCHD-RRCHD (Figure 60 b), solid solution

and CoC=solCHDlim) where an area is found where the CoC formation is less favourable than the

formation of SolCHD . The demonstration of the theoretical existence of those phase diagrams is

presented in appendix 7.5.

a)

b)

Figure 60: Two ternary phase diagrams TA-SSCHD-RRCHD

a) “solid solution and CoC>solCHD” ternary phase diagram.

b) “solid solution and CoC=solCHDlim” ternary phase diagram.

To this purpose samples were prepared and analysed by XRD. Their composition and the

phases observed by XRD is given in Annex 7.6. ANA 8 correspond to the cocrystal grown for

powder XRD, DSC, TGDTA and IR in 4.2.1. The sample ANA20-21-22 and the sample ANA7

and 8 were used for the binaries of 4.2.4.1 (ANNA20-21-22 are similar to ANA1-2-3 but the

quality of the analytical SSCHD used in the 20-21-22 was higher and the DSC scan gave better

results.). In those case the interpretation is straightforward: TA forms CoC with RRCHD (and the

excess of TA or RRCHD crystallises by its own) but not with SSCHD as shown in 4.2.4.1. We

should remark that the proportion between the peaks of TA, enantiomer and CoC follows

proportion of the different compound reasonably well.

The interpretation of the diffractogram of the next samples was more complicated. The

formation of CoC for mixture of TA and CHD with an excess of SSCHD was difficult (ANA23-24-

25) which sees to fit better to the case of “solid solution CoC=solCHDlim” ternary phase diagram.

Mixture of racemCHD with different amount of TA shows results that might confirm this

interpretation as little enantiomer is shown (ANA 13 to 17). The samples done with CHD

presenting an excess of RRCHD and different amount of TA (ANA 10-11-12, right side of he

phase diagram) gave at first surprising results as new compound was formed, named “X”, while in


108

the repetition of this experiment (ANA26-27-28) no X was found but CoC did not form in every

case (“X” is a metastable compound and will be treated below in 4.2.4.3). In fact, several

experiments showed that no CoC was formed in area where it was expected to form in every type

of phase diagram as ANA13-25. Actually many samples presented too low quantities of CoC

compared to what is expected as ANA14. Also many experiments present SSCHD in the region

where, in the second type of diagram, CoC is in equilibrium with SolCHD, as for sample 0 of

Figure 55 for instance. The crystallisation of the enantiomerically pure fraction of SSCHD is

actually a problem as is shown on Table 11 where SSCHD melting enthalpy, low compared to the

value found for RRCHD or in the literature, is due to a low crystallinity. Consequently, if this

interpretation is right, the samples on the line TA-racemCHD of the ternary (ANA 13 to17) present

no or little SSCHD not because they belong to the second type of ternary but rather because the

SSCHD is present as an amorphous phase. Actually the fact the cocrystallisation is mediated by a

metastable phase is a common phenomenon reported for the preparation of pharmaceuticals

(Jayasankar et al. 2006). The presence of an liquid, amorpheous phase is obvious for many samples

which are sticky and difficult to spread out for on the sample holder for XRD. Moreover at the end

rotoevaporation a vitreous phase is present from which the crystals germinates and grow slowly.

The cocrystallisation might be impeded by the high viscosity of this amorphous phase. For all those

reasons we think the formation of CoC is prevented because the kinetic of its growth is smaller

than TA or SolCHD’s and the liquid phase remains amorphous, as will be shown below. Other

proofs of the hypothesis stating that the limited formation of CoC is due to kinetic problems and

not thermodynamical problems will come from the result of the last part of this chapter. During the

preparation of a sample containing racCHD and TA in mr=0.5 we could isolate a fraction of

amorphous phase and analyse it by GC: it presented an ee of SSCHD close to 50%. This finding

corroborates the hypothesis of the amorphous phase which contains the excess of SSCHD well.

The composition of this amorphous phase is close to the ternary eutectic composition that contains

also a small amount of TA: the composition of a ternary eutectic between CoC, SSCHD and

RacCHD was found equal to XCoC = 0.051 xSSCHD = 0.542 xracCHD = 0.405 (calculated by Janos

Madarasz see (Thorey et al.))

The investigation of the ternary phase diagram indicates us that the solid solution is formed

but cannot give a clear answer about the thermodynamic equilibrium as it presents difficulties to be

reached and is of the type of “solid solution CoC>solCHD” ternary phase diagram Figure 60 a)

while some “trend” of “solid solution and CoC=solCHDlim” ternary phase diagram (Figure 60 b)) is

seen because of the low difference of stability of the different species (see annex 7.6). The research

of condition of crystallisation that are the most favourable to the formation of CoC was

investigated and the results are presented below in 4.2.4.4.


109

The study of molar ratio is generally important for the optimisation of a resolution

technique. In our case this study is particularly important for two reasons: SSCHD-TA cocrystal

does not exist, so an excess of TA should not decrease the eeext and an excess of TA might allow a

better formation of CoC. The study of the variation of the molar ratio will be done in 4.2.5.1. at the

moment 2 problems left untreated have to be considered:

4.2.4.3.Two issues raised by the XRD studies: sodium

hydrogen tartrate (NaTA) and metastable compound.

Sodium hydrogen tartrate (NaTA)

When the ethanolic solution of TA and the solution of racCHD were mixed together

precipitation occurred immediately. The first interpretation proposed was that this precipitate is

CoC because the concentration of RRCHD and TA would have been above the molar solubility of

CoC. However if TA was the compound that precipitates it might explain why TA is found in the

solid sample whereas it is not expected : TA would be not stable but precipitates faster than CoC

from the ethanolic solution and a fraction of TA would not be available to transform into CoC.

To test this hypothesis a large sample in the molar ratio 0.5 was prepared. As usual a

precipitate formed after mixing the CHD and TA solutions; it was filtrated out. The analysis by

XRD showed that, although its pattern looked liked TA it was the anhydrous salt catena-(hydrogen-

(+)-tartrato)-sodium. This was demonstrated by dissolving the precipate into distillated water, and

analysing the salt obtained after evaporation: it was hydrated hydrogen tartrato sodium.

Interestingly the solid formed after filtering out the salt and evaporating ethanol gave no trace of

TA or NaTA. We first thought that sodium impurity came from TA but using a higher grade TA

gave the same result. Eventually the sodium impurities came from the technical CHD and reacted

with TA during the formation of Coc. The equation of the reaction that took place in case if the

impurities are sodium hydroxide (only used as an example-it could be another base) is given in

Equation 16.

NaOHNaOH 6542664 OHCOHC

Equation 16

That’s why the next experiments with racCHD were conducted with racCHD extracted

with SCCO2 because SCCO2 does not dissolve salts generally. Solution of this purified CHD did

not lead to any precipitation when mixed with solution of TA. Also a resolution run with this

material instead of technical CHD gave better results: The yield (Y1+Y2) are higher simply


110

because initially there were more CHD compared with a technical preparation. eeext are higher this

is probably explained by the fact there is more TA for forming Coc i.e. the molar ratio was superior

to the previous experiments (see 4.2.5.1). An example is given latter in Table 13.

“X” compound

A compound called “X” was found in some cases see annex 7.6. This compound is

identified without ambiguity by XRD as it possesses peak distinguishable from the other at about

2θ = 8.85, 12.2, and 16.8. The diffractogram of ANA 12 presented in Figure 55 possesses the peak

of X. The structure of this compound is not known as it could not be isolated. The metastability

was demonstrated by different methods. Sample presenting X’s peaks lost them after been heated at

temperature above 100°C. The XRD pattern of the sample CHDG (see annex 7.6) were measured

at different temperatures and X disappeared at about 90°C. Eventually, a DSC scan of a sample of

the binary RRCHD-TA (xTA=0.84) see Figure 61. X melted at about 98°C and recrystallised

afterwards. This behaviour is typical for a metastable compound (Jacques et al. 1981; Leitao et al.

2001; Leitao et al. 2002).

Figure 61: DSC analysis of a sample of

composition RRCHD:TA 16:84 featuring

the metastable compound “X”

4.2.4.4.Conditions of crystallisation/sample preparation

We showed in 4.2.4.1 that the problems linked to the preparation of the samples are of a

kinetic nature rather than thermodynamic. In this paragraph we would like to show experiments

that, first of all, confirm the fact that the limitation is of a kinetic nature and, secondly, which

conditions play important roles.

As indicated in the bibliographic review, some diastereoisomeric salts for resolution were

prepared by precipitation into the melt without adding a solvent (the enantiomeric pair and the

resolving agent were heated together until they melted). In some cases higher F parameters were

afforded and it also saves solvent. We intended it for the system of resolution TA CHD. It was


111

impossible to prepare the CoC from the melt because of the sublimation of CHD: CHD tends to

precipitated at the top of the flask where melting was intended. Also such an operation requires

high temperature (>120°C) that could be deleterious effect on the compounds. Other solvents were

tested, water was not satisfactory because it was difficult to evaporate and lead to low yield. Less

polar solvents than ethanol could not dissolve TA. We eventually focused on ethanol because it

gave the best results, it allows a fairly good dissolution of those rather polar compounds, its

evaporation is easy and finally ethanol is also a green solvent.

The preparation of the sample was done at different temperatures, different molar ratios

and with and without perfil. Examples of diffractograms are presented on Figure 62 where the

peaks that can be focused at for a quick evaluation of the diagram are indicated. CoC being the

species responsible for the resolution its easily noticeable peak is the most important. The data are

collected in the annex 7.7 in Table 22.

It shows that a complete formation CoC where no RRCHD is left uncocrystallised (this is

to say no SolCHD) is possible as in the case of “40°C with perfil”. Another point is shown by this

study the moment of the sample preparation when the solvent is rotoevaporated is not the only

determinant step of the crystallisation: the latter evaporation of ethanol remaining after

rotoevaporation in an open Petri dish gave good results whereas sample let to evaporate in a round-

bottomed flask gave only little formation of CoC. Indeed the cocrystallisation did not occur in

some cases (55°C) and only SolCHD and TA were detected. A too slow or partial evaporation of

ethanol might be the reason. In two cases the crystallinity was really poor and surprisingly in one

case solid was mostly made of CoC and SolCHD (30°C rbf) while in the other it was SolCHD and

TA (40°C rbf). The conditions that may hinder CoC formation are not well understood, and perfil

did not seem to have a very determinant influence. The problem is a matter of difference in the

crystal growth kinetics of the different species.


112

Figure 62: Diffractograms of sample of mr=0.5 evaporated at different temperature

Only a characteristic peak useful to the determination and evaluation of the quantity of a

compound is marked. The preparation of the sample at 40°C was done with and without

perfil.

The study of crystallinity is often done by DSC that could give quantitative results. Indeed

DSC is the technique the most used for determination of crystallinity especially in the case of

polymers. The study of the melting of samples at mr=0.5 gives different peaks that difference in

size can be explained by the fact that the crystallinity of certain phase is more or less important.

However the presence of many phases (CoC, enantiomer (SSCHD), solCHD and TA) renders this

type of analysis delicate. Sampling is difficult for this kind of sample and the present of an achiral

support as perfil will lead to smaller sample. The interpretation of the scans is difficult because

many parameters influence the size and the temperature of the peaks: quantity of uncrystalised

RRCHD, quantity of precipitated TA, sublimation of CHD, crystal size. It is also important to be

able to run the study with technical CHD for the optimisation of an industrial process and DSC is

very sensitive to the purity of the material. A DSC scan at 10°C/min (usual scan speed used over

this work) was eventually run showed that the eutectic melting was not separated well from the


113

subsequent dissolving of CoC into it. For all those reason the study of crystallinity by DSC looks

tedious and uncertain even if we cannot assume that valuable results would not arise from such a

study.

XRD analysis has the advantage of rapidity and easiness of analysis and interpretation

without destroying the samples. This method allowed us, first of all, to validate our hypothesis of

4.2.4.2. The second indication given by XRD is semi-quantitative: the sample is the best when the

peak of CoC is the highest compared to solCHD and TA. The best results were obtained by

rotoevaporation of ethanol at 40°C followed by evaporation of ethanol overnight from a Petri dish.

It is important to remember that XRD does not show amorpheous phases and their presence is an

important source of limitation of the resolution process.

Concerning the preparation of the sample due to be extracted for the resolution of CHD

with TA, we identified the limitations on the ee1 and found empirical condition the most favourable

to a good resolution. However, the obtained extracts have ees too low for commercialisation and

should be improved.

4.2.5. Toward enantiopure products, molar ratio, double

extraction

The aim of this last part concerning the resolution is the practical use of the more

theoretical research presented above. They are based on two principles, the first is the variation of

the molar ratio that can be realised due to the fact that no cocrystal is formed between RRCHD and

TA, the second is to perform the resolution twice on the same material with (+)- or (-)-TA.

4.2.5.1.Extraction with molar ratios varying

The concept of molar ratio, nn

CHD

TAmr , was introduced in 2.1.3.4 and in the case of the

resolution of CHD by TA in 2.3.4.

The first extraction step was performed at 20 MPa and 33°C. The second was based on the

acidobasic treatment shown in 4.1.4.3. The result of these series of resolution is presented in Figure

63 and Figure 64.


114

Figure 63: Yield and F parameter with varying molar ratio

Figure 64: Enantiomeic excess with varying molar ratio


115

If the phase diagram of Figure 60 a) is correct, the CHD that is extracted corresponds to all

the CHD that is uncocrystalised consequently:

if mr<0.5 : mrYext 1 and mr

mreeext

1

if mr>0.5: Yext=0.5 and 1extee .

Yext-1Yraff and 1Rafee . All those equation are plotted on the graphs to allow

comparison. The theoretical Yext fits with the points well, while Yraf is a bit below the theoretical

curve. This is probably due to the loss of material over the raffination of the residuum, as stated in

4.2.2. The theoretical eeext is not so bad for small mr values but for mr>0.4 it goes wrong.

Explanations were proposed for this in 4.2.3.1. The last value of eeraf stay close to 0.8 and shows a

decrease for the small value of mr (this could fit with a phase diagram where CoC intakes some

SSCHD see 7.5.2.2) and then slightly increase maybe because of a better crystallisation (this can

also explain the increase in eeext). The point for mr=0.1 is very low and this might be due to an

uncomplete first extraction step or a matter of crystallisation.

F is increased by increasing mr. At best, it reached 0.8, which represents a good

improvement. However the ee are still too limited and a double resolution seemed necessary. The

interpretation of the resolution with varying molar ratio is enabled by the work done in the previous

part. The corroboration is good.

4.2.5.2.Resolution repeated twice

We did not manage to produce an extract with satisfying ee by a single resolution. In this

context it was intended to increase the enantiopurity by performing a second resolution performed

on 2 typical extracts obtained according to the resolution with TA: mixture 1 that presents an ee of

75% of SSCHD and mixture 2 with an ee of 85% of RRCHD.

ee2 is higher than ee1 so, the same way that TA was used to produce RRCHD of good

enantiopurity, SSTA was used to produce SSCHD: this is to say we expected to obtain SSCHD or

RRCHD of high ee from the second extract only. The further enantiopurification of SSCHD

(mixture 1) and RRCHD (mixture 2) is shown on Figure 65 and Figure 66, respectively. Different

molar ratios were applied and the results are collected in Table 13. The molar ratio is again defined

as nn

CHD

TAmr ( nnn chdrrchdsschd because nnn chdrrchdsschd 2 ) The next figures are

intended to give a visual idea of the quantity of resolving agent compared to the targeted


116

compound. The height of the rectangles is proportion to the quantity of RRCHD and SSCHD or TA

(or SSTA).

Figure 65: Second resolution of mixture 1 presenting an ee of SSCHD

Figure 66: Second resolution of mixture 2 presenting an enantiomeric excess of RRCHD


117

Table 13: Result of the different experiments of further enantioenrichment of mixture 1 and

mixture 2

First extraction Second extraction mr

Y1 ee1 Y2 ee2

RacCHD

ee=0 0.5 50% 87% 50% 79%

0.5 48% 52.4% 52% 96.6% Mixture 1

SSCHD ee=74.6% 0.58 42.6% 39.0% 55.3% 99.6%

0.85 14.1% 4.2% 83.8% 99.3% Mixture 2

RRCHD ee=85.0% 1 7.1% 50% 92.9% 99.2%

In every case, the extraction was complete (Y1 + Y2 close to 100%) and the targeted final

extract presented high ee, superior to 99% in three cases. The crystallisation occurred well and

when TA (or SSTA) was not in excess (last line of Table 13) the formation of CoC was nearly

complete as Y2 was close to mr. This is what was expected as Y2 roughly corresponds to the

quantity of CHD cocrystallised which is egal to the quantity of TA, as long as TA is not in excess.

If we compare those last four extractions to the extraction made with racCHD we noticed

that Xrr2, thecomposition in RRCHD in the second extract, (when TA was used) increased when

Xrr of the sample (and Xrr1 of the first extract) increased. This is the principle which the double

resolution is based on: the successive improvement of the enantiomeric excess. 2 explanations for

the increase of ee2 are given below.

Referring to the limitation indicated on the ee2 before, the two explanation invoked

precedently are both satisfying in this latter case. The first was that some free enantiomer remains

unextracted at the end of the first extraction and is recovered at the second. In the case of a

resolution where the starting material is not a racemic the remaining of the first extraction

recovered in the second would have a higher Xrr so that the obtained second extract would be

contaminated with residue less harming to the enantiopurity of the second extract: ee2 is

consequently higher. The second explanation is that CoC intakes a small amount of SSCHD. This

explanation fits well with the result. Actually equilibrium would take place between the two liquid

solutions of CHD and CoC. The equilibrium would be of that kind: when Xrr of one phase

increases Xrr of the other also does, as proposed in the annex 7.5.2.2 with the example of Figure

76. This explanation is convincing but should be handled with care as no demonstration of the

existence of CoC containing SSCHD was intended at the moment. Generally the exhibition of a


118

solid solution is treated by DSC. This is difficult and should be handled with care because many

wrong interpretations might arise from DSC scan and precise measurement are required for

showing small variations in the melting temperature. Generally it is tedious work and coupling with

microscopic observations might be helpful. The presence of amorphous phase, metastable

compound “X” but also polymorphic forms of racCHD or enantiomeric CHD, the sublimation of

CHD render this task difficult.

A resolution repeated twice using TA and SSTA afforded ee above 99% for the preparation

of RRCHD and SSCHD, respectively, while a simple variation of mr could not. The highly

enantiopure CHD consists in the second extract. These are suitable for industrial use. But if higher

enantiopurity is required further enantioenrichment is feasible by simple fractional recrystallisation

as the ee of the obtained fraction is much higher than the eutectic’s (Jacques et al. 1981). It should

be also added that due to the feasibility of repeated resolution no CHD fraction is lost for any

fraction can be recycled in the next resolution. And the mother liquor used for recrystallisation can

be reused for preparing a new sample with TA or SSTA. The resolution of CHD is thus flexible

and the E-factor of a process of preparation of enantiopure CHD would be very low.

4.3. Conclusion on the resolution of CHD by cocrystallisation and SFE and further plan

The cocrystal responsible for the resolution was characterised and its structure resolved. A

good understanding of the resolution was afforded by analytical method. The fact that SSCHD and

TA do not cocrystallise leads to the development of resolution based on an excess of TA.

Eventually a resolution based on resolutions repeated twice brought products with high

enantiomeric excess above 99%.

Further work

The probable existence of a liquid solution of CoC that allows CoC to intake a small

amount of SSCHD was not demonstrated at the moment. The conditions of preparation of the

initial sample are determinant and seem an interesting matter but the study seemed also very

difficult. However, we have shown that a second resolution overcomes the limitation of a one-step

process.

Beyond those two theoretical issues, it should be emphasised that the resolution of CHD by

TA using SFE afforded almost enantiomerically pure CHD in a very competitive way: no

expensive reagent is needed, the resolution agent is completely recovered, CHD is never wasted as


119

it is recycled in a new resolution. No hazardous solvent is used. Hence a perfectly green process is

ready to scale-up for the commercial production of SSCHD and RRCHD.

Conclusion 120

5. CONCLUSIONS

Two different ways of producing enantiomers were presented in this work. They

correspond to the 2 most important chirotechnologies: the use of a chiral catalyst and resolution by

precipitation of enantiomeric compounds. In both cases high enantiopurity was achieved (ee>99%).

The last important chirotechnlogy is the chromatography and it was used for analysis in this work.

The first method was based on an enzyme-catalysed reaction and involved the development

of novel protocols for testing alcohol dehydrogenases in dense gases. Those protocols were applied

to the alcohol dehydrogenases from Lactobacillus brevis and the production of enantiopure R-1-

phenylethanol. The second consisted in the development of the method of resolution of trans-1,2-

cyclohexanediol by crystallisation with tartaric acid followed by extraction with supercritical

carbon dioxide. The work included an analytical monitoring of the process that afforded a better

understanding and, eventually, the improvement of results with ee superior to 99%. Throughout

this work a special stress was given to green techniques. Indeed alternative solvent as dense gases

were used and the matter of waste minimisation was considered.

The result of the study of production of R-1-phenylethanol is very positive if the

enantiopurity of the obtained product is only considered but the fast deactivation of LBADH

renders an industrial process using this enzyme in dense gases unviable. A set of methods and

protocols were successfully developed. They are available for the screening of other alcohol

dehydrogenases for stereoselective hydrogenation in dense gases.

The resolution of trans-1,2-cyclohexanediol by crystallisation with tartaric acid followed

by extraction with supercritical carbon dioxide is fully developed and perfectly green. Both

enantiomers are separated and their enantiopurity is high. The resolving agent, tartaric acid, is

recovered after extraction. So the next step would be the scaling up of this process.

Litterature 121

6. LITTERATURE

Aaltonen, O. (1999). Chapter 4.9 Enzymatic catalysis. Chemical Synthesis Using Supercritical Fluids. P. G. Jessop and W. Leitner. Weinheim, Wiley-VCH: 480 pp. Ács, M., E. Fogassy and F. Faigl (1985). "A convenient method for optical resolutions via diastereoisomeric salt formation." Tetrahedron 41(12): 2465-2470. Adams, D., P. Dyson and S. Tavener (2004 ). Alternative Reaction Media in Industrial Processes. Chemistry in Alternative Reaction Media D. Adams, P. Dyson and S. Tavener. Chichester, John Wiley & Sons. Adlercreutz, P. (1991). "On the importance of the support material for enzymatic synthesis in organic media - Support effects at controlled water activity." Eur J Biochem 199(3): 609 - 614. Adlercreutz, P. (1996). "Cofactor regeneration in biocatalysis in organic media." Biocatal. Biotransform. 14: 1-30. Allen, F. H. (2002). "The Cambridge Structural Data Base: a quarter of a million crystal structures and rising " Acta Crystallogr., Sect. B: Struct. Sci. 58: 380-388. Anastas, P. and J. Warner (1998). Green Chemistry: Theory and Practice. New York, Oxford University Press. Barbour, L. J. (2001). "X-Seed -- A Software Tool for Supramolecular Crystallography." J. Supramolecular Chemistry 1(4-6): 189-191. Benaissi, K., M. Poliakoff and N. R. Thomas (2009). "Dynamic kinetic resolution of rac-1-phenylethanol in supercritical carbon dioxide." Green Chem. 11: 617 - 621. Bodai, V., O. Orovecz, G. Szakacs, L. Novak and L. Poppe (2003). "Kinetic resolution of trans-2-acetoxycycloalkan-1-ols by lipase-catalyzed enantiomerically selective acylation." Tetrahedron: Asymmetry 14(17): 2605-2612. Bolivar, J. M., L. Wilson, S. A. Ferrarotti, J. M. Guisán, R. Fernández-Lafuente, et al. (2006). "Improvement of the stability of alcohol dehydrogenase by covalent immobilization on glyoxyl-agarose." J. Biotechnol. 125(1): 85-94. Borho, N., T. Häber and M. A. Suhm (2001). "Chiral self-recognition in the gas phase: the case of glycidol dimers." Phys. Chem. Chem. Phys. 3: 1945-1948. Bortolini, O., E. Casanova, G. Fantin, A. Medici, S. Poli, et al. (1998). "Kinetic resolution of vic-diols by Bacillus stearothermophilus diacetyl reductase." Tetrahedron: Asymmetry 9(4): 647-651. Braeutigam, S., S. Bringer-Meyer and D. Weuster-Botz (2007). "Asymmetric whole cell biotransformations in biphasic ionic liquid/water-systems by use of recombinant Escherichia coli with intracellular cofactor regeneration." Tetrahedron: Asymmetry 18(16): 1883-1887. Brundtland, C. G. (1987). Report of the World Commission on Environment and Development: Our Common Future. Oxford, Oxford University Press. Bugg, T. (2004). Introduction to Enzyme and Coenzyme Chemistry. Oxford, UK, Blackwell Publishing Ltd. Cao, L. (2005). Carrier-bound Immobilized Enzymes. Weinheim, WILEY-VCH Verlag. Caron, G. and R. J. Kazlauskas (1991). "An optimized sequential kinetic resolution of trans-1,2-cyclohexanediol." J. Org. Chem. 56(26): 7251-6. Chanzy, H., E. Maia and S. Perez (1982). "Cellulose Organic Solvents. III. The Structure of the N-Methylmorpholine N-Oxide---trans-1,2-Cyclohexanediol Complex." Acta Cryst. B38: 852-855. Chatterjee, A., M. Sasikumar and N. N. Joshi (2007). "Preparation of enantiopure trans-1,2-cyclohexanediol and trans-2-aminocyclohexanol." Synth. Commun. 37(10): 1727-1733.

Litterature 122

Chiappe, C., E. Leandri, B. D. Hammock and C. Morisseau (2007). "Effect of ionic liquids on epoxide hydrolase-catalyzed synthesis of chiral 1,2-diols." Green Chem. 9(2): 162-168. Ching, C., J. M. Gruber, G. W. Huisman, E. Mundorrf and L. M. Newman (2008). Engineering of ketoreductase derived from (R)-specific alcohol dehydrogenases from Lactobacillus brevis and Lactobacillus kefiri and uses for production of chiral alcohols. Application: WO WO, (Codexis, Inc., USA). 92pp. Clark, J. (1999). "Green chemistry: challenges and opportunities." Green Chem. 1: 1-8. Clayden, J., N. Greeves, S. Warren and P. Wothers (2001). Organic chemistry. Oxford, Oxford university press. Codexis (2008). Collins, A. N., C. N. Sheldrake and J. Crosby (1992). Chirality in industry- an overview. Chischester, John Wiley & Sons. Cordes, A. and M. Kula (1994). "Large-scale purification of formate dehydrogenase." Methods Enzymol. 228: 600-8. Corma, A., S. Iborra and A. Velty (2007). "Chemical Routes for the Transformation of Biomass into Chemicals." Chemical Reviews 107(6): 2411-2502. Cornish-Bowden, A. (1995). Analysis of Enzyme Kinetic Data. Oxford, Oxford University Press. Cox, G. B. (2005). Preparative Enantioselective Chromatography. Oxford, UK, Blackwell Publishing. Curzons, A. D., D. J. C. Constable, D. N. Mortimer and V. L. Cunningham (2001). " So you think your process is green, how do you know?—Using principles of sustainability to determine what is green–a corporate perspective." Green Chem. 3: 1–6. Dean, J. R., B. Liu and E. Ludkin (2000). Supercritical Fluid Extraction of Caffeine from Instant Coffee. Supercritical Fluid Methods and Protocols. J. R. Williams and A. A. Clifford. Totowa, Humana Press Inc. Deetz, J. S. and J. D. Rozzell (1988). "Enzymic catalysis by alcohol dehydrogenases in organic solvents." Ann. N. Y. Acad. Sci. 542(Enzyme Eng. 9): 230-4. DeSimone, J. M. (2002). "Practical Approaches to Green Solvents." Science 297(5582): 799-803. DiMasi, J. A., R. W. Hansen and H. G. Grabowski (2003). "The price of innovation: new estimates of drug development costs." Journal of Health Economics 22: 151–185. Dreyer, S. and U. Kragl (2008a). "Ionic liquids for aqueous two-phase extraction and stabilization of enzymes." Biotechnol. Bioeng. 99(6): 1416-1424. Dreyer, S. and U. Kragl (2008b). Use of aqueous two-phase systems with an ionic liquid phase for the extraction of biomolecules. Application: DE DE, (Universitaet Rostock, Germany). 12pp. Eckstein, M., T. Daussmann and U. Kragl (2004a). "Recent Developments in NAD(P)H Regeneration for Enzymatic Reductions in One- and Two-Phase Systems." Biocatal. Biotransform. 22(2): 89 - 96. Eckstein, M., M. Villela Filho, A. Liese and U. Kragl (2004b). "Use of an ionic liquid in a two-phase system to improve an alcohol dehydrogenase-catalyzed reduction." Chem. Commun.(9): 1084-1085. Eliel, E. L., S. H. Wilen and L. N. Mander (1994). Stereochemistry of organic compounds. N.Y., Wiley-Interscience. EPA (2007). Toxics Release Inventory Reporting Year 2007, Public Data Release, Summary of Key Findings. Ernst, M., B. Kaup, M. Mueller, S. Bringer-Meyer and H. Sahm (2005). "Enantioselective reduction of carbonyl compounds by whole-cell biotransformation, combining a formate dehydrogenase and a (R)-specific alcohol dehydrogenase." Appl. Microbiol. Biotechnol. 66(6): 629-634. Fágáin, C. Ó. (1995). "Understanding and increasing protein stability." Biochim. Biophys. Acta 1252(1): 1-14.

Litterature 123

Fahrenkamp-Uppenbrink, J. (2002). "Chemistry goes green." Science 297: 798. Faigl, F., E. Fogassy, M. Nogradi, E. Palovics and J. Schindler (2008). "Strategies in optical resolution: a practical guide." Tetrahedron: Asymmetry 19: 519-536. Fekete, V. (2008). theoretical and expeimental limitations of shear-driven chromatography. Brussel, Vrije Universiteit. Ferloni, C., M. Heinemann, W. Hummel, T. Daussmann and J. Buechs (2004). "Optimization of Enzymatic Gas-Phase Reactions by Increasing the Long-Term Stability of the Catalyst." Biotechnol. Prog. 20(3): 975-978. Findrik, Z., D. Vasic-Racki, S. Lütz, T. Daussmann and C. Wandrey (2005). "Kinetic modeling of acetophenone reduction catalyzed by alcohol dehydrogenase from Thermoanaerobacter sp." Biotechnol. Lett. 27(15): 1087-1095. Flack, H. D. (1983). "On enantiomorph-polarity estimation." Acta Crystallogr., Sect. A: Found. Crystallogr. 39(6): 876-881. Fletcher, S. P., R. B. C. Jagt and B. L. Feringa (2007). "An astrophysically-relevant mechanism for amino acid enantiomer enrichment." Chem. Commun.: 2578–2580. Fogassy, E., M. Ács, T. Szili, B. Simándi and J. Sawinsky (1994). "Molecular chiral recognition in supercritical solvents." Tetrahedron Lett. 35(2): 257-260. Francotte, E. (2005). Chiral stationary phases for preparative enantioselective chromatography. Preparative Enantioselective Chromatography. G. B. Cox. Oxford, UK, Blackwell Publishing. Fuchs, G., L. Doguet, D. Barth and M. Perrut (1992). "Enantiomer fractionation of phosphine oxides by preparative subcritical fluid chromatography." Journal of Chromatography A 623(2): 329-336. Grell, J., J. Bernstein and G. Tinhofer (2000). "Graph-set analysis of hydrogen-bond patterns: some mathematical concepts. Erratum." Acta Crystallogr., Sect. B: Struct. Sci. 56(1): 166-166. Groaning, M. D., B. J. Rowe and C. D. Spilling (1998). "New homochiral cyclic diol ligands for titanium alkoxide catalyzed phosphonylation of aldehydes." Tetrahedron Lett. 39(31): 5485-5488. Groeger, H., F. Chamouleau and C. Hagedorn (2006). Enzymatic reduction of aldehydes for primary alcohol production. Application: WO WO, (Maxens G.m.b.H., Germany). 41 pp. Groeger, H., W. Wienand, C. Rollmann, H. Werner and D. Reichert (2008). Process for preparation of enantiomer-enriched alkylene carbonates. Application: WO WO, (Evonik Degussa GmbH, Germany). 37pp. Groger, H., W. Hummel, S. Buchholz, K. Drauz, T. V. Nguyen, et al. (2003). "Practical Asymmetric Enzymatic Reduction through Discovery of a Dehydrogenase-Compatible Biphasic Reaction Media." Org. Lett. 5(2): 173-176. Grunwald, J., B. Wirz, M. P. Scollar and A. M. Klibanov (1986). "Asymmetric oxidoreductions catalyzed by alcohol dehydrogenase in organic solvents." J. Am. Chem. Soc. 108(21): 6732-4. Gu, K. F. and T. M. S. Chang (1990). "Production of essential L-branchedchain amino acids in bioreactors containing artificial cells immobilized multienzyme systems and dextmn-NAD+." Biotechnol. Bioeng. 36: 263-269. Habulin, M., M. Primozic and Z. Knez (2007). "Supercritical fluids as solvents for enzymatic reactions." Acta Chim. Slov. 54(4): 667-677. Halling, P. J. (1992). "Salt hydrates for water activity control with biocatalysts in organic media " Biotechnology Techniques 6(3): 1573-6784. Halling, P. J. (1994). "Thermodynamic predictions for biocatalysis in nonconventional media: theory, tests, and recommendations for experimental design and analysis." Enzyme Microb. Technol. 16(3): 178-206. Hanessian, S., A. Gomtsyan, M. Simard and S. Roelens (1994). "Molecular Recognition and Self-Assembly by Weak Hydrogen Bonding: Unprecedented Supramolecular Helicate Structures from Diamine/Diol Motifs." J. Am. Chem. Soc. 116(10): 4495-4496.

Litterature 124

Hanessian, S., R. Saladino, R. Margarita and M. Simard (1999). "Supramolecular Chirons Based on Enantiodifferentiating Self-Assembly between Amines and Alcohols (Supraminols)." Chem. Eur. J. 5(7): 2169-2183. Hanessian, S., M. Simard and S. Roelens (1995). "Molecular Recognition and Self-Assembly by Non-amidic Hydrogen Bonding. An Exceptional Assembler of Neutral and Charged Supramolecular Structures." J. Am. Chem. Soc. 117(29): 7630-7645. Hatano, K., T. Takeda and R. Saito (1994). "Optical resolution of trans-bicyclo[2.2.1]heptane-2,3-diamine: chiral recognition in the crystal of its complex with (2R,3R)-O,O-dibenzoyltartaric acid." J. CHEM. SOC. PERKIN TRANS. 2: 579 - 584. Herbstein, F. H. (2005). Crystalline Molecular Complexes and Compounds. NY, Oxford University Press Hildebrand, F. and S. Luetz (2006). "Immobilization of alcohol dehydrogenase from Lactobacillus brevis and its application in a plug-flow reactor." Tetrahedron: Asymmetry 17(23): 3219-3225. Hoshino, K. (1960). "Organism producing isopropanol from acetone. V. Enzymo-logical [studies] on the oxidation-reduction of Lactobacillus brevisvar. hofuensis. 2." Nippon Nogei Kagaku Kaishi 34: 608-15. Hummel, W. (1997). New alcohol dehydrogenases for the synthesis of chiral compounds. New Enzymes for Organic Synthesis Screening, Supply and Engineering Berlin / Heidelberg, Springer. 58: 145-184. Hummel, W. and B. Riebel (1997). Alcohol dehydrogenase of Lactobacillus and its use in the enzymic production of chiral alcohols. EP. EP, (Boehringer Mannheim Gmbh, Germany; Kernforschungsanlage Juelich; Forschungszentrum Juelich GmbH). 34 pp. Illés, R., C. Kassai, G. Pokol, E. Fogassy and D. Kozma (2002). "Thermoananlytical study of O,O'-dibenzoyl-(2R,3R)-tartaric acid Part III. SMC-s formation with chiral secondary alcohol." J. Therm. Anal. Calorim. 68: 679-685. Iwasaki, F., T. Maki, O. Onomura, W. Nakashima and Y. Matsumura (2000). "Chemo- and stereoselective monobenzoylation of 1,2-diols catalyzed by organotin compounds." J. Org. Chem. 65(4): 996-1002. Jacques, J., A. Collet and S. H. Wilen (1981). Enantiomers, racemates, and resolution. New York, John Wiley & Sons. Jayasankar, A., A. Somwangthanaroj, Z. J. Shao and N. Rodriguez-Hornedo (2006). "Cocrystal Formation during Cogrinding and Storage is Mediated by Amorphous Phase." Pharm. Res. 23(10). Jessop, P. G. and W. Leitner (1999). Chemical Synthesis Using Supercritical Fluids. Weinheim, Wiley-VCH Verlag GmbH. Jones, P. G., M. R. Edwards and A. J. Kirby (1989). "Bond length and reactivity: the pinacol rearrangement. 1. Redetermination of the structure of trans-cyclohexane-1,2-diol." Acta Crystallogr., Sect. C: Cryst. Struct. Commun. C45(2): 244-6. Jönsson, A., W. van Breukelen, E. Wehtje, P. Adlercreutz and B. Mattiasson (1998). "the influence of water activity on the enantioselectivity in the enzyme-catalysed reduction of 2-pentanone." J. Mol. Catal. B: Enzym. 5: 273-276. Jonsson, A., E. Wehtje, P. Adlercreutz and B. Mattiasson (1999). "Thermodynamic and kinetic aspects on water vs. organic solvent as reaction media in the enzyme-catalyzed reduction of ketones." Biochimica et Biophysica Acta, Protein Structure and Molecular Enzymology 1430(2): 313-322. Jönsson, Å., E. Wehtje, P. Adlercreutz and B. Mattiasson (1999a). "Thermodynamic and kinetic aspects on water vs. organic solvent as reaction media in the enzyme-catalysed reduction of ketones." Biochimica et Biophysica Acta (BBA) - Protein Structure and Molecular Enzymology 1430(2): 313-322. Jönsson, Å., E. Wehtje, P. Adlercreutz and B. Mattiasson (1999b). "Thermodynamic and kinetic aspects on water vs. organic solvent as reaction media in the enzyme-catalysed reduction of ketones." Biochim. Biophys. Acta 1430(2): 313-322. Kaga, H., Y. Yamauchi, A. Narumi, K. Yokota and T. Kakuchi (1998). "A preparative method for kinetic resolution of trans-1,2-cycloalkanediols through lipase-catalyzed transesterification." Enantiomer 3(2): 203-205.

Litterature 125

Kasai, Y., H. Taji, T. Fujita, Y. Yamamoto, M. Akagi, et al. (2004). "MaNP acid, a powerful chiral molecular tool for preparation of enantiopure alcohols by resolution and determination of their absolute configurations by the 1H NMR anisotropy method." Chirality 16(9): 569-585. Kassai, C., Z. Juvancz, J. Bálint, E. Fogassy and D. Kozma (2000). "Optical Resolution of Racemic Alcohols via Diastereoisomeric Supramolecular Compound Formation with O,O′-Dibenzoyl-(2R,3R)-tartaric Acid." Tetrahedron 56(42): 8355-8359. Kawai, K., M. Imuta and H. Ziffer (1981). "Microbially mediated enantioselective hydrolysis of racemic acetates." Tetrahedron Lett. 22(27): 2527-30. Kawashima, M. and A. Hirayama (1991). "Direct optical resolution of vicinal diols and an a-hydroxy oxime with a vicinal diamine." Chem. Lett.(5): 763-6. Kazlauskas, R. J., A. N. E. Weissfloch, A. T. Rappaport and L. A. Cuccia (1991). "A rule to predict which enantiomer of a secondary alcohol reacts faster in reactions catalyzed by cholesterol esterase, lipase from Pseudomonas cepacia, and lipase from Candida rugosa." J. Org. Chem. 56(8): 2656-65. Kennedy, I. J. and D. E. Jane (2002). stereochemistry in drug design, third edition. Textbook of drug design and discovery. U. Madsen and T. Liljefors. NY, Taylor & Francis Ltd. Kerc, J., S. Srcic, Z. Knez and P. Sencar-Bozic (1999). "Micronization of drugs using supercritical carbon dioxide." Int. J. Pharm. 182(1): 33-39. Kerton, F. M. (2009). Alternative Solvents for Green Chemistry. Cambridge, UK, The Royal Society of Chemistry, Thomas Graham House. Keszei, S., B. Simándi, E. Székely, E. Fogassy, J. Sawinsky, et al. (1999). "Supercritical fluid extraction: a novel method for the resolution of tetramisole." Tetrahedron: Asymmetry 10: 1275-1281. Khetan, S. and T. Collins (2007). "Human pharmaceuticals in the aquatic environment: a challenge to geen chemistry." Chem. Rev. 107: 2319-2364. Klibanov, A. M. (2003). "Asymmetric enzymic oxidoreductions in organic solvents." Curr. Opin. Biotechnol. 14(4): 427-431. Kmecz, I., B. Simándi, E. Székely, J. Lovasz and E. Fogassy (2007). "Application of mixtures of tartaric acid derivatives in resolution via supercritical fluid extraction." Chirality 19: 430-433. Knez, Z. (2009). "Enzymatic reactions in dense gases." The Journal of Supercritical Fluids 47(3): 357-372. Kobayashi, Y., K. Kodama and K. Saigo (2004). "Supramolecular Architecture Consisting of an Enantiopure Amine and an Achiral Carboxylic Acid: Application to the Enantioseparation of Racemic Alcohols." Org. Lett. 6(17): 2941-2944. Kozma, D. (2002). Optical resolution via diastereoisomeric salt formation. London, CRC Press. Kozma, D., G. Pokol and M. Ács (1992). "Calculation of the efficiency of optical resolutions on the basis of the binary phase diagram for the diastereoisomeric salts." J. Chem. Soc., Perkin Trans. 2: 435. Kruse, W., W. Hummel and U. Kragl (1996). "Alcohol-Dehydrogenase-Catalyzed Production Of Chiral Hydrophobic Alcohols.A New Approach Leading To A Nearly Waste-Free Process." Recueil des travaux chimiques des Pays-Bas 115(4): 239-243. Laumen, K., D. Breitgoff, R. Seemayer and M. P. Schneider (1989). "Enantiomerically pure cyclohexanols and cyclohexane-1,2-diol derivatives; chiral auxiliaries and substitutes for (-)-8-phenylmenthol. A facile enzymic route." J. Chem. Soc., Chem. Commun.(3): 148-50. Lavandera, I., A. Kern, M. Schaffenberger, J. Gross, A. Glieder, et al. (2008). "An exceptionally DMSO-tolerant alcohol dehydrogenase for the stereoselective reduction of ketones." ChemSusChem 1(5): 431-436. Leitao, M. L. P., R. A. E. Castro, F. S. Costa and J. S. Redinha (2001). "Phase transitions of 1,2-cyclohexanediol isomers studied by polarised light microscopy and differential thermal analysis." Thermochim. Acta 378(1-2): 117-124.

Litterature 126

Leitao, M. L. P., M. E. Eusebio, T. M. R. Maria and J. S. Redinha (2002). "(Solid + liquid) phase diagram for trans-1,2-cyclohexanediol enantiomer mixtures." J. Chem. Thermodyn. 34(4): 557-568. Lennon, I. and N. Johnson. (2009). "Preformed catalysts." manufacturing chemist, from http://www.manufacturingchemist.com/story.asp?storycode=54987. Lenz, R., S. V. Ley, D. R. Owen and S. L. Warriner (1998). "Stereoelectronic effects in the reactions of conformationally restricted, substituted cyclohexane-1,2-diones with 1,2-diols." Tetrahedron: Asymmetry 9(14): 2471-2480. Ley, S. V. and S. Mio (1996a). "Dispiroketals in synthesis. Part 21. Use of the chiral 2,2'-bis[(halomethyl)dihydropyrans] as new protecting and resolving agents for 1,2-diols." Synlett(8): 789-790. Ley, S. V., S. Mio and B. Meseguer (1996b). "Dispiroketals in synthesis. Part 22. Use of chiral 2,2'-bis[(phenylthiomethyl)dihydropyrans] as new protecting and resolving agents for 1,2-diols." Synlett(8): 791-792,795. Liese, A., K. Seelbach and C. Wandrey (2006). Industrial Biotransformations; Second, Completely Revised and Extended Edition, Wiley-VCH Verlag GmbH & Co. Liu, S., B. S. Dien, N. N. Nichols, K. M. Bischoff, S. R. Hughes, et al. (2007). "Coexpression of pyruvate decarboxylase and alcohol dehydrogenase genes in Lactobacillus brevis." FEMS Microbiol. Lett. 274(2): 291-297. Lloyd, M. A., G. E. Patterson, G. H. Simpson, L. L. Duncan, D. P. King, et al. (2007). "Solid-state compounds of stereoisomers: cis and trans isomers of 1,2-cyclohexanediol and 2,3-tetralindiol." Acta Crystallogr., Sect. B: Struct. Sci. B63(3): 433-447. Lorenz, H., A. Perlberg, D. Sapoundjiev, M. P. Elsner and A. Seidel-Morgenstern (2006). "Crystallization of enantiomers." Chemical Engineering and Processing 45: 863-873. Macrae, C. F., P. R. Edgington, P. McCabe, Elna Pidcock, G. P. Shields, et al. (2006). "Mercury: visualization and analysis of crystal structures." J. Appl. Crystallogr. 39: 453–457. Madarász, J., D. Kozma, G. Pokol, M. Ács and E. Fogassy (1994). "Merit of estimations from DSC measurements for the efficiency of optical resolutions." J. Therm. Anal. Calorim. 42(5): 877-894. Martin Britto Dhasa, S. A., M. Suresha, G. Bhagavannarayanab and S. Natarajan (2007). "Growth and characterization of l-Tartaric acid, an NLO material " J. Cryst. Growth 309(1): 48-52. Mateo, C., L. M. Van Langen and F. Van Rantwijk (2004). Preparation of crosslinked enzyme particles using a polyaldehyde prepared from a polysaccharide as the water-soluble crosslinking agent. WO, (Technische Universiteit Delft, Neth.). 8 pp. Matsuda, T., T. Harada and K. Nakamura (2000). "Alcohol dehydrogenase is active in supercritical carbon dioxide." Chem. Commun.(15): 1367-1368. Matsuda, T., K. Watanabe, T. Kamitanaka, T. Harada and K. Nakamura (2003). "Biocatalytic reduction of ketones by a semi-continuous flow process using supercritical carbon dioxide." Chem. Commun.(10): 1198-9. Mazet, C., S. Roseblade, V. Koehler and A. Pfaltz (2006). "Kinetic Resolution of Diols and Pyridyl Alcohols by Cu(II)(borabox)-Catalyzed Acylation." Org. Lett. 8(9): 1879-1882. McHugh, M. and V. Krukonis (1994). Supercritical Fluid Extraction: Principles and Practice. Boston, Butterworth-Heinemann. Meudt, A., R. Wisdom and C. Boehm (2006). Preparation of chiral oxiranes via alcohol dehydrogenase reduction of a-substituted carbonyls. Application: WO WO, (Archimica G.m.b.H., Germany). 17pp. Miroliaei, M. and M. Nemat-Gorgani (2002). "Effect of organic solvents on stability and activity of two related alcohol dehydrogenases: a comparative study." International Journal of Biochemistry & Cell Biology 34(2): 169-175. Mizuta, S., Y. Ohtsubo, T. Tsuzuki, T. Fujimoto and I. Yamamoto (2006). "Kinetic resolution of d,l-1,2-diols catalyzed by amine-phosphinite bifunctional organocatalysis derived from quinidine." Tetrahedron Lett. 47(47): 8227-8229.

Litterature 127

Molnár, P. (2009). Optical resolution by supercritical fluid extraction. Department of chemical and environmental process engineering. Budapest, BME. Molnar, P., E. Székely, B. Simandi, S. Keszei, J. Lovasz, et al. (2006). "Enantioseparation of ibuprofen by supercritical fluid extraction." J. Supercrit. Fluids 37(3): 384-389. Molnar, P., P. Thorey, G. Bansaghi, E. Szekely, L. Poppe, et al. (2008). "Resolution of racemic trans-1,2-cyclohexanediol with tartaric acid." Tetrahedron: Asymmetry 19(13): 1587-1592. Mueller, M., W. Sauer and G. Laban (2003). Production of (R)- and (S)-8-chloro-6-hydroxy-octanoic-acid-alkyl esters by enzymatic reduction. Application: DE DE, (Viatris G.m.b.H. & Co. KG, Germany). 6 pp. Naemurz, K., R. Fukuda, M. Murata, M. Konishi, K. Hirose, et al. (1995). "Lipase-catalyzed enantioselective acylation of alcohols: a predictive active site model for lipase YS to identify which enantiomer of an alcohol reacts faster in this acylation." Tetrahedron: Asymmetry 6(9): 2385-94. Negawa, M. and F. Shoji (1992). "Optical resolution by simulated moving-bed adsorption technology." Journal of Chromatography A 590(1): 113-117. Nelson, D. L. and M. M. Cox (2004). Lehninger Principles of Biochemistry, W. H. Freeman. Nemák, K., M. Acs, Z. M. Jászay, D. Kozma and E. Fogassy (1996). "Study of the diastereoisomers formed between (N-alkyl)-pipecolic acid-anilides and 2R,3R-tartaric acid or O,O′-dibenzoyl-2R,3R-tartaric acid. Do the tartaric acids form molecular-complexes, instead of salts during optical resolutions?" Tetrahedron 52(5): 1637-1642. Newman, P. (1978-1984). Optical resolution procedures of chiral compounds 1-3. NY, Resolution Information Center. Niefind, K., J. Muller, B. Riebel, W. Hummel and D. Schomburg (2003). "The crystal structure of R-specific alcohol dehydrogenase from Lactobacillus brevis suggests the structural basis of its metal dependency." J. Mol. Biol. 327(2): 317-328. Niefind, K., B. Riebel, J. Muller, W. Hummel and D. Schomburg (2000). "Crystallization and preliminary characterization of crystals of R-alcohol dehydrogenase from Lactobacillus brevis." Acta Crystallographica, Section D: Biological Crystallography D56(12): 1696-1698. Noorduin, W. L., T. Izumi, A. Millemaggi, M. Leeman, H. Meekes, et al. (2008a). "Emergence of a Single Solid Chiral State from a Nearly Racemic Amino Acid Derivative." J. Am. Chem. Soc. 130(4): 1158-1159. Noorduin, W. L., H. Meekes, A. A. C. Bode, W. J. P. van Enckevort, B. Kaptein, et al. (2008b). "Explanation for the Emergence of a Single Chiral Solid State during Attrition-Enhanced Ostwald Ripening: Survival of the Fittest." Cryst. Growth Des. 8(5): 1675-1681. Novak, Z., M. Habulin, V. Krmelj and Z. Knez (2003). "Silica aerogels as supports for lipase catalyzed esterifications at sub- and supercritical conditions." The Journal of Supercritical Fluids 27(2): 169-178. Noyori, R. (2001). asymmetric catalysis: science and opportunities. Nobel lecture. Noyori, R. and T. Ohkuma (2001). "Asymmetric Catalysis by Architectural and Functional Molecular Engineering: Practical Chemo- and Stereoselective Hydrogenation of Ketones." Angewandte Chemie International Edition 40(1): 40-73. Panza, J. L. (2001). "Synthesis of fluorinated NAD as a soluble coenzyme for enzymatic chemistry in fluorous solvents and carbon dioxide." Tetrahedron 58: 4091-4104. Panza, J. L., A. J. Russell and E. J. Beckman (2002). "Enzyme activity using a perfluoropolyether-modified NAD(H) in fluorous solvents and carbon dioxide." ACS Symposium Series 819(Clean Solvents: Alternative Media for Chemical Reactions and Processing): 64-81. Pasteur, L. (1922-1939). Oeuvres de Pasteur / réunies par Pasteur Vallery-Radot. Paris, Masson. Perrin, R. and J.-P. Scharff (2002). Chimie industrielle. Paris, Dunod.

Litterature 128

Perrut, M. (1994). "Advances in supercritical fluid chromatographic processes." Journal of Chromatography A 658(2): 293-313. Perry, R. H., C. Wu, M. Nefliu and R. G. Cooks (2007). "Serine sublimes with spontaneous chiral amplification." Chem. Commun.: 1071–1073. Peschko, C. and J. Stohrer (2005). Enzymatic production of chiral secondary alcohols by enantioselective reduction of ketones. Application: EP EP, (Consortium fuer Elektrochemische Industrie G.m.b.H., Germany). 17 pp. Peschko, C. and J. Stohrer (2006). Process for enantioselective enzymatic reduction of keto compounds. Application: US US, (Consortium fur Elektrochemische Industrie G.m.b.H., Germany). 9pp. Peterson, P. E. and M. Stepanian (1988). "Hydroboration of vinyl ethers with diisopinocampheylborane." J. Org. Chem. 53(9): 1903-7. Pfaller, R. and J. Stohrer (2007). Production of optically acitve secondary alcohols from ketones employing an immobilized carbonyl reductase. Application: EP EP, (Wacker Chemie AG, Germany). 19pp. Phinney, K. W. (2001). Sub- and supercritical fluid chromatography for enantiomer separation. Chiral separation techique. G. Subramaniam. Weinheim, WILEY-VCH Verlag Gmbh. Poliakoff, M., J. M. Fitzpatrick, T. R. Farren and P. T. Anastas (2002). "Green Chemistry: Science and Politics of Change." Science 297(5582): 807-810. Prelog, V., S. Mutakm and K. Kovacevic (1983). "Über die Enantiomerentrennung durch Verteilung zwischen flüssigen Phasen. 3. Mitteilung. Selektivität der lipophilen Weinsäureester für chirale Ammonium-Salze verschiedener Konstitution und Konfiguration." Helv. Chim. Acta 66(7): 2279-2284. Product Portfolio (2007). Juelich, Germany, Juelich chiral solution, Codexis. Ramade, F. (2005). Elément d'écologie appliqué. Paris, Dunod. Rees, D. G. and P. J. Halling (2001). "Chemical modification probes accessibility to organic phase: proteins on surfaces are more exposed than in lyophilized powders." Enzyme Microb. Technol. 28(2-3): 281-292. Rodrigues, A. E. and S. P. Pais (2001). Modelling and simulation in SMB for chiral purification. chiral separation techique. G. Subramaniam. Weinheim, WILEY-VCH Verlag Gmbh. Rosell, C. M., A. M. Vaidya and P. J. Halling (1996). "Continuous in situ water activity control for organic phase biocatalysis in a packed bed hollow fiber reactor." Biotechnol. Bioeng. 49(3): 284-9. Schlieben, N. H., K. Niefind, J. Mueller, B. Riebel, W. Hummel, et al. (2005). "Atomic Resolution Structures of R-specific Alcohol Dehydrogenase from Lactobacillus brevis Provide the Structural Bases of its Substrate and Cosubstrate Specificity." J. Mol. Biol. 349(4): 801-813. Schroer, K., U. Mackfeld, I. A. W. Tan, C. Wandrey, F. Heuser, et al. (2007a). "Continuous asymmetric ketone reduction processes with recombinant Escherichia coli." J. Biotechnol. 132(4): 438-444. Schroer, K., E. Tacha and S. Luetz (2007b). "Process Intensification for Substrate-Coupled Whole Cell Ketone Reduction by In Situ Acetone Removal." Organic Process Research & Development 11(5): 836-841. Schumacher, J., M. Eckstein and U. Kragl (2006). "Influence of water-miscible organic solvents on kinetics and enantioselectivity of the (R)-specific alcohol dehydrogenase from Lactobacillus brevis." Biotechnology Journal 1(5): 574-581. Schütte, H., J. Flossdorf, H. Sahm and M.-R. Kula (1976). "Purification and Properties of Formaldehyde Dehydrogenase and Formate Dehydrogenase from Candida boidinii (p 151-160)." Eur. J. Biochem. 62(1): 151-160. sheldon, R. A. (1993). Chirotechnology: the Industrial Synthesis of Optically Active Compounds. New York, Marcel Dekker. Sheldon, R. A. (2000). "Atom efficiency and catalysis in organic synthesis." Pure Appl. Chem. 72(7): 1233-1246.

Litterature 129

Sheldon, R. A. (2008). "E factors, green chemistry and catalysis: an odyssey." Chem. Commun.: 3352–3365. Sheldon, R. A., I. Arends and U. Hanefeld (2007). 3 Catalytic Reductions. Green Chemistry and Catalysis. R. A. Sheldon, I. Arends and U. Hanefeld. Weinheim, Wiley-VCH. Sheldon, R. A., I. Arends and U. Hanefeld (2007 ). Green Chemistry and Catalysis. Weinheim, WILEY-VCH Verlag GmbH & Co. . Sheldrick, G. M. (1997a). SHELXL-97 Program for the Refinement of Crystal Structures. Göttingen, University of Göttingen. Sheldrick, G. M. (1997b). SHELXS-97 Program for Crystal Structures Solution. Göttingen, University of Göttingen. Sillanpaa, R., M. Leskela and L. Hiltunen (1984). "Crystal and molecular structures of (+-)trans- and cis-1,2-cyclohexanediols." Acta Chem. Scand. Ser. A B38(3): 249-54. Simándi, B., S. Keszei, E. Fogassy, S. Kemény and J. Sawinsky (1998). "Separation of enantiomers by supercritical fluid extraction." J. Supercrit. Fluids 13: 331-336. Snijder-Lambers, A. M., E. N. Vulfson and H. J. Doddema (1991). "Optimization of alcohol dehydrogenase activity and NAD(H) regeneration in organic solvents." Recueil des Travaux Chimiques des Pays-Bas 110(05): 226-30. Sonderegger, O. J., T. Burgi and A. Baiker (2003). "Asymmetric hydrogenation of cyclohexane-1,2-dione over cinchonidine-modified platinum." J. Catal. 215(1): 116-121. Soni, S., J. D. Desai and S. Devi (2001). "Immobilization of yeast alcohol dehydrogenase by entrapment and covalent binding to polymeric supports." J. Appl. Polym. Sci. 82(5): 1299-1305. Spek, A. L. (2003). "Single-crystal structure validation with the program PLATON." J. Appl. Crystallogr. 36: 7-13. Stanton, M. K. and A. Bak (2008). "Physicochemical Properties of Pharmaceutical Co-Crystals: A Case Study of Ten AMG 517 Co-Crystals." Cryst. Growth Des. 8(10): 3856-3862. Stillger, T., M. Bönitz, M. Villela Filho and A. Liese (2002). "Überwindung von thermodynamischen Limitierungen in substratgekoppelten Cofaktorregenerierungsverfahren (Overcoming thermodynamic limitations in substrate-coupled cofactor regeneration processes)." CHEMIE INGENIEUR TECHNIK 74(7): 1035-1039. Subramaniam, G. (2001). Techniques in preparative chiral separation. chiral separation techique. G. Subramaniam. Weinheim, WILEY-VCH Verlag Gmbh. Synoradzki, L., U. Bernas and P. Ruskowski (2008). "Tartaric Acid And Its O-Acyl Derivatives. Part 2. Application Of Tartaric Acid And Of O-Acyl Tartaric Acids And Anhydrides. Resolution Of Racemates." Org. Prep. Proced. Int. 40(2): 163-200. Szczepańska, B., M. Gdaniec and U. Rychlewska (1995). "TheO, O′-dibenzoyl derivative of (R, R)-tartaric acid as a host compound for simple organic ethers." Journal of Inclusion Phenomena and Macrocyclic Chemistry 22(3): 211-219. Székely, E., B. Simándi, R. Illés, P. Molnár, I. Gebefügi, et al. (2004). "Application of supercritical fluid extraction for fractionation of enantiomers." The Journal of Supercritical Fluids 31(1): 33-40. Székely, E., B. Simándi, K. László, E. Fogassy, G. Pokol, et al. (2002). "Effect of achiral support on the resolution of tetramisole by supercritical fluid extraction." Tetrahedron: Asymmetry 13(13): 1429-1434. Takata, N., K. Shiraki, R. Takano, Y. Hayashi and K. Terada (2008). "Cocrystal Screening of Stanolone and Mestanolone Using Slurry Crystallization." Cryst. Growth Des. 8(8): 3032-3037. Tanaka, M., M. Oba, K. Tamai and H. Suemune (2001). "Asymmetric Synthesis of a,a-Disubstituted b-Amino Acids Using (S,S)-Cyclohexane-1,2-diol as a Chiral Auxiliary." J. Org. Chem. 66(8): 2667-2673. Tewari, Y. B., M. M. Schantz, K. W. Phinney and J. D. Rozzell (2005). "A thermodynamic study of the ketoreductase-catalyzed reduction of 2-alkanones in non-aqueous solvents." J. Chem. Thermodyn. 37(1): 89-96. Thayer, A. M. (2008). "Interaction Yields." Chem. Eng. News 86(31): 12-20.

Litterature 130

Thorey, P., P. Bombicz, I. M. Szilágyi, P. Molnár, G. Bánsághi, et al. "Co-crystal of (R,R)-1,2-cyclohexanediol with (R,R)-tartaric acid, a key structure in resolution of the (±)-trans-diol by supercritical extraction, and the related ternary phase system." Thermochim. Acta In Press, Corrected Proof. Tiecco, M., L. Testaferri, F. Marini, S. Sternativo, C. Santi, et al. (2003). "Synthesis of enantiomerically pure 1,4-dioxanes from alkenes promoted by organoselenium reagents." Tetrahedron: Asymmetry 14(9): 1095-1102. Tishkov, V. I., A. G. Galkin, G. N. Marchenko, O. A. Egorova, D. V. Sheluho, et al. (1993). "Catalytic Properties and Stability of a Pseudomonas sp.101 Formate Dehydrogenase Mutants Containing Cys-255-Ser and Cys-255-Met Replacements." Biochem. Biophys. Res. Commun. 192(2): 976-981. Trivedi, A., M. Heinemann, A. C. Spiess, T. Daussmann and J. Buechs (2005). "Optimization of adsorptive immobilization of alcohol dehydrogenases." J. Biosci. Bioeng. 99(4): 340-347. Trivedi, A. H., A. C. Spiess, T. Daussmann and J. Buechs (2006). "Study on Mesophilic and Thermophilic Alcohol Dehydrogenases in Gas-Phase Reaction." Biotechnol. Prog. 22(2): 454-458. Trost, B. M. (1991). "The Atom Economy-A Search for Synthetic Efficiency." Science 254(5037): 1471-1477. van der Donk, W. A. and H. Zhao (2003). "Recent developments in pyridine nucleotide regeneration. ." Curr. Opin. Biotechnol. 14: 421–426. Vasic-Racki, D. (2006). History of Industrial Biotransformations – Dreams and Realities. Industrial Biotransformations; Second, Completely Revised and Extended Edition. A. Liese, K. Seelbach and C. Wandrey, Wiley-VCH Verlag GmbH & Co.: 556 pp. Villela Filho, M., T. Stillger, M. Mueller, A. Liese and C. Wandrey (2003). "Is log P a convenient criterion to guide the choice of solvents for biphasic enzymatic reactions?" Angewandte Chemie, International Edition 42(26): 2993-2996. Villeneuve, M. S. and L. A. Miller (2005). Preparative-scale supercritical fluid chromatography. Preparative Enantioselective Chromatography. G. B. Cox. Oxford, UK, Blackwell Publishing. Virto, C., I. Svensson, P. Adlercreutz and B. Mattiasson (1995). "Catalytic activity of noncovalent complexes of horse liver alcohol dehydrogenase, NAD+ and polymers, dissolved or suspended in organic solvents." Biotechnol. Lett. 17(8): 877-82. Vulfson, E. N., P. J. Halling and H. L. Holland (2001). Enzymes in Nonaqueous Solvents: Methods and Protocols. Totowa, Humana. Wilen, S. H., A. Collet and J. Jacques (1977). "Strategies in optical resolutions." Tetrahedron 33(21): 2725-2736. Wilson, A. J. C. (1992). International Tables for X-ray Crystallography. Dordrecht, Kluwer Academic Publishers. Wojaczynska, E. and J. Skarzewski (2008). "Novel C2-symmetric chiral ligands: enantioselective transformation of cyclic 1,2-diols into 1,2-bis(phenylsulfenyl) and 1,2-bis(phenylselenyl) derivatives." Tetrahedron: Asymmetry 19(5): 593-597. Yang, F. and A. J. Russell (1993). "Optimization of bakers' yeast alcohol dehydrogenase activity in an organic solvent." Biotechnol. Prog. 9(3): 234-41. Yasohara, Y., M. Yano, S. Kawano and N. Kizaki (2007). Method for producing optically active 2-(N-substituted aminomethyl)-3-hydroxybutyric acid ester. Application: WO WO, (Kaneka Corporation, Japan). 32pp. Zacharis, E., I. C. Omar, J. Partridge, D. A. Robb and P. J. Halling (1997). "Selection of salt hydrate pairs for use in water control in enzyme catalysis in organic solvents." Biotechnol. Bioeng. 55(2): 367-374. Zaks, A. and A. M. Klibanov (1984). "Enzymatic Catalysis in Organic Media at 100-Degrees-C." Science 224(4654): 1249-1251. Zaks, A. and A. M. Klibanov (1985). "Enzyme-Catalyzed Processes in Organic-Solvents." Proc. Natl. Acad. Sci. U. S. A. 82(10): 3192-3196.

Litterature 131

Zaks, A. and A. M. Klibanov (1986). "Substrate-Specificity of Enzymes in Organic-Solvents Vs Water Is Reversed." J. Am. Chem. Soc. 108(10): 2767-2768.

Appendix 132

7. APPENDIX

7.1. Challenge of green chemistry

A generally well accepted definition of green chemistry is : “The design, development,

and implementation of chemical processes and products to reduce or eliminate substances

hazardous to human health and the environment.” (Anastas et al. 1998; Poliakoff et al. 2002)

Interestingly, EPA, Environmental Protection Agency (USA), has adopted this definition.

Another definition is: “Green chemistry efficiently utilizes (preferably renewable) raw

materials, eliminates waste and avoids the use of toxic and/or hazardous reagents and

solvents in the manufacture and application of chemical products.(Sheldon et al. 2007 )” This

definition has the advantage of introducing the issues of limiting the waste and saving the raw

material.

It should be noticed that the notion of “endangering” should be considered in a broad

acceptation: danger can be physical (explosition, flammability, …), toxic (or toxicological :

carcinogenic, mutagenic…) or global (ozone depletion, global warming…). The notion of danger

should be enlarged according to the new research and knowledge in safety, toxicity and ecology.

The global warming has only been known for a short period of time and research in toxicology is

improving and reveals hazard where the responsible substances were not even detected few

decades ago (Khetan et al. 2007).

7.1.1. Context of the development of green chemistry

Green chemistry is to connect to the paradigm of the “sustainable development” which was

adopted by the united nation in 1987 as “a development which fits to the needs of the present

day without endangering next generation’s capacity to fulfil their needs.” Report of the World

Commission on Environment and Development: Our Common Future (Brundtland 1987) Figure

67 shows the theoretical set of sustainable development as the merging of considerations which are

social, ecological and economical.

Figure 67: Sustainable development as a confluence

of three domains: social, economy, and environment

Accordingly, sustainable chemistry is defined as:

“Within the broad framework of Sustainable

Development, we should strive to maximise resource

efficiency through activities such as energy and non-renewable resource conservation, risk

Appendix 133

minimisation, pollution prevention, minimisation of waste at all stages of a product life-cycle,

and the development of products that are durable and can be re-used and recycled.

Sustainable Chemistry strives to accomplish these ends through the design, manufacture and

use of efficient and effective, more environmentally benign chemical products and processes.

(Curzons et al. 2001)”

Public opinion about chemical industry is generally negative (Fahrenkamp-Uppenbrink

2002). The strong distrust of chemical industry can stand as a limit to its activity or growth. The

European Chemical Industry Council (CEFIC) survey in 1994 showed that 60% of the general

public had an unfavourable view of the chemical industry and in the USA, a survey carried out for

the Chemical Manufacturers Association (CMA) in 1993 showed that only 26% were favourably

disposed towards the industry (Clark 1999). Many industrial accident got an important media

coverage such as Flixborough disaster (1974, England), Seveso disaster (1976, Italy), Love Canal

disaster, hazardous waste disposal close to a school (from 1978 on, USA), Bhopal Disaster (1984,

India), accident at the Chernobyl nuclear power plant (1986, Soviet Union, Ukraine), Toulouse

AZF, explosion of chemical factory (2001, France). To this, it should be added the arising

consciousness that air (from housing to cities), water (from the ground water to oceans) and soils

are contaminated with chemical wastes is not restricted to the specialists (Ramade 2005). The first

examples of public concern are probably acid rains, which were followed by oil slick or adverse

effects of DDT, and currently global warming. Facing a strong opposition, different actors of

chemical industry have to adopt a “counter-propaganda” strategy. Green chemistry has got this first

aspect: a reassuring concept that targets the non specialist. Green, the colour of chlorophyll, is

opposition to the black colour of the petrol : Total has changed its logo, it is green now.

Legislation has been effective in improving environmental conditions. The waste and their

disposal are better controlled. The maximal level of emission of contaminant in air, rivers, ground

water and so on, has been progressively lowered, as an example we can mentioned the dust or SOx

in the air, the heavy metals or dioxines in the rivers. But toxic materials are still discharged in

considerable amounts. For 2007, the latest year for which data are available, disposal or

other releases of TRI chemicals totalled almost 1.9 million tons from about 22,000 U.S.

facilities submitting approximately 84,900 chemical forms (Epa 2007). Preventing waste is

(sometimes) wrongly seen as a cost without profit. For instance, the USA annually spends $115

billions in 1992 treating this enormous quantity of waste (Clark 1999). The cost provoked by waste

are from different nature as retreatment, recycling or disposal and continuously increases because

of new legislation. To this, should be added the cost of raw material that is not transformed into

valuable products (this factor gets more important with petrol rising price), cost in energy and so

Appendix 134

on. In 1990, USA passed a law called Pollution Prevention Act. It represents a shift, as it promotes

the idea that instead of treating the waste, it is better to avoid them.

7.1.2. 12 principle of green chemistry. Derived conceps.

Those 12 principles were found at first in the book Green Chemistry: Theory and Practice

(Anastas et al. 1998) . There are found in many books and articles thereafter.

1. It is better to prevent waste than to treat or clean up waste after it is formed.

2. Synthetic methods should be designed to maximise the incorporation of all

materials used in the process into the final product.

3. Wherever practicable, synthetic methodologies should be designed to use and

generate substances that possess little or no toxicity to human health and the

environment.

4. Chemical products should be designed to preserve efficiency of function while

reducing toxicity.

5. The use of auxiliary substances (e.g. solvents, separation agents, etc) should be

made unnecessary wherever possible and, innocuous when used.

6. Energy requirements should be recognised for their environmental and economic

impacts and should be minimised. Synthetic methods should be conducted at

ambient temperature and pressure.

7. A raw material or feedstock should be renewable rather depleting wherever

technically and economically practicable.

8. Unnecessary derivatisation (blocking group, protection/deprotection, temporary

modification of physical/chemical processes) should be avoided whenever

possible.

9. Catalytic reagents (as selective as possible) are superior to stoichiometric reagents.

10. Chemical products should be designed so that at the end of their function they do

not persist in the environment and break down into innocuous degradation

products.

11. Analytical methodologies need to be further developed to allow for real-time, in-

process monitoring and control prior to the formation of hazardous substances.

12. Substances and the form of a substance used in a chemical process should be

chosen so as to minimise the potential for chemical accidents, including releases,

explosions and fires.

The chemical syntheses should be designed to prevent waste, leaving no waste to treat or

clean up. In many cases, the improvement must not only be done on a step of the synthesis but the

Appendix 135

whole synthesis should design from start. From this principle, can be deduce other indicators for

the evaluation of a process than the yield: the E-factor and the atom efficiency.

The E factor is the weight of waste divided by the weight of desired product (Equation 17).

It must be minimised. Average E factors according to industry segment are presented in the table

below.

WW

productdesired

wastefactorE_

_ Equation 17: E-factor

Table 14: Waste of the different segment of chemical industry

Adapted from (Poliakoff et al. 2002)

Industry segment Product tonnage Kg waste/kg

product Oil refining 106-108 <0.1

Bulk chemicals 104-106 <1-5

Fine chemical 102-104 5 >50

Pharmaceuticals 10-103 25 >100

The concept of atom utilisation (Sheldon 2000) atom efficiency, or atom economy concept

(Trost 1991) is an extremely useful tool for rapid evaluation of the amount of waste generated by

alternative routes to a specific product. It is calculated by dividing the molecular weight of the

desired product by the sum total of the molecular weights of all substances produced in the

stoichiometric equation for the reaction(s) involved.

iii

productdesired

MnMefficiencyAtom __

Equation 18: atom efficiency

where ni is the stoechiometric coeffient of the molecule i when ndesired product =1.

The comparison is made on a theoretical basis (i.e., 100% chemical yield). The theoretical

E factor is readily derived from the atom efficiency, for example, an atom efficiency of 40%

corresponds to an E factor of 1.5. In practice, the E factor is much higher as the yield is not 100%,

as an excess of reagent(s) is often used, and solvent losses and salt generation in subsequent

neutralisation steps have to be taken into account. More developed tool can also be used as the Life

Cycle Assessment (LCA) or adding indicator as the energy consumed or the quantity of CO2

released by the process.

Appendix 136

We cannot develop every aspects of green chemistry because virtually every different

fields of chemistry are involved. The case of the synthesis of ibuprofene is interesting. The “green”

synthesis is based on two catalysts step (see Figure 68). The influence of the catalyst will find other

examples in the case of biocatalysis (see 2.1.2).

Figure 68: Two

determinant catalytic steps

in the “green” synthesis of

ibuprofen.

Reproduced from (Sheldon

2000)

Other subject of attention is the use of biomass as a feedstock (Corma et al. 2007) which

possess many large scale applications such the conversion of vegetable oil into biofuel. It should be

noticed that in this case as for others nothing is perfect the matter is more the greening of the

chemical industry rather than to obtain a industry labelled “green”: the production of biofuel

generate much glycerol that has not find an use yet and the conversion of biomass into fuel is the

competition with the traditional use of such crop: food supply.

7.1.3. Alternative solvent: supercritical fluids and SCCO2

Solvent is a burning issue due to the large quantity used and the difficult they present to be

contained and recycled. Almost the 15 billion kilograms of organic and halogenated solvent that

are produced worldwide inevitably end up leaching in the environment (Desimone 2002). A

process can be simply improved by the replacement of a solvent by another belonging to the same

family with more favourable properties such as higher boiling point - hence they generate less

waste as vapour - or a lower toxicity. A simple ways of limiting solvent consumption consist in the

recycling or the development of processes using less or even no solvent at all. A process can also

be redesigned in order to implement the use of an alternative solvent. Among them, it can be

mentioned the ionic liquids, the fluorinated phases, or the supercritical fluids.

Supercritical fluids are fluids at temperature and pressure above their critical points (Tc,

Pc) and at a pressure below the solidification (see Figure 69 and Table 15). Properties of these

fluids are unique. Density, viscosity, diffusivity and so on are intermediate between a gas and a

solid. They allow better mass transfer properties than using conventional solvents. The physical

properties of SCF can be tuned by playing on pressure and temperature (Kerton 2009). Some well

known reaction have been running under supercritical conditions for several decades: the Born–

Haber process for ammonia synthesis operates under supercritical conditions as do the

Appendix 137

polymerisation of low density polyethylene (LDPE) (Adams et al. 2004 ) and the synthesis of

methanol (Perrin et al. 2002).

Figure 69: phase diagramm (P,T)

of a fluid

Reproduced from (Kerton 2009)

Table 15: Critical points of fluid presenting an industrial interest.

Compound Tc (◦C) Pc (bar) ρc (kg/m3)

Carbon dioxide 31.3 72.9 468

Nitrogen oxide 36.5 71.4 457

Xenon 16.6 58.8 1155

SF6 45.5 37 734

Ethane 32.4 48 203

Propane 97 42 217

Butane 152 38 225

Pentane 197 33.3 237

Diethyl ether 193.6 41.7 265

Methanol 240.5 78.9 272

Ethanol 243.4 63 276

Ammonia 132.3 111.3 235

Water 374.4 226.8 323

Supercritical carbon dioxide (SCCO2) presents several supplementary advantages. While

its Pc is not so high (72.8 bar) Its Tc is low (31°C) and is compatible with heat sensitive

Appendix 138

compounds. CO2 is an “ideal solvent”, being non toxic, non inflammable, readily accessible, and

less expensive than organic solvents. That’s why it is classified as “Generally recognized as safe”

(GRAS) by the FDA. This solvent is easily separated from the product. When the pressure is

released SCCO2 become a gas where the solubility of most products is very low. Running a

reaction in a dense gas instead of a conventional solvent allows to save energy and to produce

compounds free of solvent. Moreover SCCO2 is a renewable substance readily available in large

amount (product of fermentation).

However it should be emphasised that the high-pressure involved with those fluids requires

a special care about safety and also the equipment is more expensive. Running batch reactions in

SCCO2 is often said impossible because a large volume reactor requires very large walls and,

consequently, an important investment.

Providers of equipment for high-pressure technologies exist and plants using this reaction

medium exist. The most important application SCCO2 concerns the extraction of product meant to

human consumption. One of the first extraction is the decaffeination of coffee (Dean et al. 2000).

Many extraction concern extraction of essential oils (Mchugh et al. 1994). It is also used for

separation techniques, such as resolution of enantiomer (Simándi et al. 1998) and chromatography

(Phinney 2001). Differences in solubilities, dissociation constant, and stability may be more

pronounced than in ordinary solvents and can be adjusted by fine setting of temperature and

pressure.

SCCO2 is also used for the formation of fine particles in a process as expansion from gas

saturated solutions (PGSS). A solute dissolved in SCCO2 is sprayed through a nozzle where the

pressure drops and the CO2 becomes a gas in which the solute is not soluble anymore and,

consequently, precipitates or crystallises (Kerc et al. 1999).

SCCO2 is also used for chemical reaction (Jessop et al. 1999). Thomas Swan & Co (UK)

performs the continuous hydrogenation in this solvent; they take advantage of its ability to dissolve

H2. Its capacity is up to 1000 tonnes per year (Adams et al. 2004 ). This fluid afforded better

selectivity than with conventional solvent because it was possible to control the selectivity

kinetically instead of thermodynamically. The homogeneity of the reaction medium plays an

important role as well as the good mass transport properties which allow good kinetics. Enzyme-

catalysed reaction in this SCCO2 are also the object of much research (Aaltonen 1999; Habulin et

al. 2007; Knez 2009) and industrial process are currently running.

A negative point encountered with SCCO2 is its extremely low polarity and also it weak

solvent power which limits the range of possible solute to apolar compound of relatively low

molecular weight (Kerton 2009). Those difficulties can be overcome by adding a modifier, a low

Appendix 139

molecular weight solvent that is soluble in SCCO2 improves it solvent properties. Ethanol,

methanol are commonly used.

7.2. Reaction run with coimmobilised NADP ad LBADH in non-aqueous solvent, propane and heptane

Table 16: Result of the bioconversion of ACP into RPE in heptane and propane.

WEI

GH

T of

LB

ADH

mg

Wei

ght o

f NAD

P m

g

Con

cent

ratio

n of

AC

P m

M

ratio

isop

ropa

nol-

acet

ophe

none

YIEL

D %

TON

(EN

ZYM

E)

TON

(CO

-EN

ZYM

E)

INIT

IAL

RAT

E µm

ol/min/mgE

cons

tant

K*1

03

A 2.2 13 40 1 3.7 120 3 - 1.6

B 4 5.2 40 0.9 0.8 18 2 - 0.082

C 3.3 4.2 30 1.2 0.0 0 0 0 0

D 2.9 6.2 46 1.1 4.2 100 7 0.2 1.8

E *** *** 40 1 0.0 0 0 0 0

F 2.4 3.2 40 2 8.4 116 13 0.2 4.1

Hep

tane

G 2.4 3.2 40 2 8.8 120 14 0.2 4.5

H 25 11.10 28 16 45.3 180 49 0.056 24 Propane

I 25 11.10 28 16 0 0 0 0 0

The activity of the preparation of LBADH used the experiences reported in Table 16 was

8 U/mg and contained 0.15µmol of active LBADH for 1 mg.

Reaction condition: in 25ml heptane (except for C 38 mL) with 50µL (A, B, D, E) or 30 µL

(F, G) water. In C’s case, no water was added the enzyme and coenzyme were not coimmobilised,

the two powder were introduce in the reactor without any further preparation. For A, B, D, E, F and

G, the catalyst were prepared according to 6.3, 500 mg silica (A, B) coarse glass beads (D, F, G

(7g)) were used with sugar (A (2 mg), F (12 mg)), G (12 mg)) or not (B, D, ). The catalyst for E

was the same as for D reused after a two-day reaction. Reaction H was run in the view cell (about

60 mL) with glass beads (5 g) the water activity was set at 0,8 with salt hydrate Na2SO4, 10/0.

Enzyme was incubated in propane with the salts for 2 hours. Reaction I was run in the same

Appendix 140

conditions as H, except from the catalysts which was made of the two powder with no further

preparation

7.3. Miscibility of ACP, ISP, 1-phenylethanol, acetone and propane

GL : equilibrium gas-liquid

L : liquid phase

Table 17: Miscibility in propane

% in weight of the different compounds

Propane Aceto-

phenone

Phenyl-

ethanol

Iso-

propanol Acetone T °C P (bar) Phase

76 % 24 % 0 % 0 % 0 % 37 12 GL

76 % 24 % 0 % 0 % 0 % 30 30 L

76 % 24 % 0 % 0 % 0 % 30 50 L

61 % 18 % 0 % 21 % 0 % 25 9 GL

61 % 18 % 0 % 21 % 0 % 30 9 L

61 % 18 % 0 % 21 % 0 % 30 52 L

61 % 18 % 0 % 21 % 0 % 30 30 L

79 % 0 % 21 % 0 % 0 % 29 11 GL

79 % 0 % 21 % 0 % 0 % 30 50 L

79 % 0 % 21 % 0 % 0 % 30 30 L

82.5 % 5.5 % 5.5 % 4.5 % 1.5 % 30 10 GL

82.5 % 5.5 % 5.5 % 4.5 % 1.5 % 30 30 L

82.5 % 5.5 % 5.5 % 4.5 % 1.5 % 30 50 L

7.4. Method of determination of the structure of the cocrystal CoC

The selected transparent crystal of 1 for single crystal X-ray diffraction measurement had

the size of 0.55 x 0.55 x 0.34 mm. 1 was mounted on a loop with parathon oil. Cell parameters

were determined by least-squares of all reflections in the whole measured range. Intensity data

Appendix 141

were collected on a RIGAKU RAXIS-RAPID diffractometer (graphite monochromator; Mo-K

radiation, = 0.71073Å). Empirical absorption correction was applied to the data. The structure

was solved by direct methods (Sheldrick 1997b). Anisotropic full-matrix least-squares refinements

(Sheldrick 1997a; Barbour 2001) on F2 for all non-hydrogen atoms were performed. Neutral atomic

scattering factors were taken from the International Tables for X-ray Crystallography (Wilson

1992). Crystallographic data, parameters of data collection, structure solution and refinement can

be found in Table 1. Since there are no strong anomalous scattering centres in the constituents and

the diffraction measurement was performed using Mo-K radiation, the Flack x parameter (Flack

1983) is not reliable. The O-H hydrogen atomic positions could be located in the difference Fourier

maps. Hydrogen atoms were included in structure factor calculations but they were not refined. The

isotropic displacement parameters of the hydrogen atoms were approximated from the U(eq) value

of the atom, to which they were bonded.

Crystallographic data (excluding structure factors) for the cocrystal structure of CoC have

been deposited with the Cambridge Crystallographic Data Centre as supplementary publication

number CCDC 728882.

7.5. Theoretical ternary diagram with a liquid solution

The study of the phase diagram is not only interesting to a theoretical point of view but can

allow to determine in which proportion the sample to be extracted by SFE will give the best yield

and ee. The phase diagram are investigated are room temperature and atmospheric. The hypothesis

is done that over the extraction process the equilibrium between the solid phase is not drastically

modified. This hypothesis should be rather correct as only condensed phase are present and

pressure has little effect on them and the applied temperature is closed to room temperature.

The aim of this part is to show what type can be expected. The different phase diagram are

proposed will the hypothesis on the chemical potential made. The cocrystal exist and forms from an

equimolar mixture of RRCHD and TA. We can conclude that:

TARRCHDCoc Equation 19

No cocrystal is formed between SSCHD and TA (see 4.2.4.1). As RacCHD is formed when

there are RRCHD and SSCHD we can conclude the same way that:

2 RRCHDRacCHD Equation 20

Appendix 142

At first we will not consider the fact that a solid solution is formed between the SSCHD

and RRCHD but a more usual racemic compound.

7.5.1. If no liquid solution between the CHD enantiomers

exists.

The Equation 19 and the Equation 20 are not sufficient for determining the phase diagram:

we do not know if CoC or RacCHD will formed in an equimolar mixture of TA, RRCHD, SSCHD.

We need another hypothesis that is given by the very fact the resolution works and consequently

we know that in the precious case RRCHD is more stabilised by forming a compound with TA

compared to with SSCHD, this is to say:

TARacCHDCoC SSCHD Equation 21

Now the phase diagram may be plotted.

Figure 70: “No solid solution

CoC>racem” phase diagram of

RRCHD, SSCHD and TA

supposing that a racemic

compound is formed and not a

solid solution.

The central point corresponds

to the composition of the sample used

for the first experiment (mr=0.5) (Molnar et al. 2008). At this point this phase diagram is respected

it would lead to CoC and SSCHD only.

Remark: if TARacCHDCoC SSCHD we obtain:

Figure 71 : “No solid solution CoC<racem”

alternative phase diagram with RRCHD,

SSCHD, CoC, RacCHD, and TA.

Appendix 143

The central point corresponds to the composition of the sample used in (Molnar et al.

2008): if this phase diagram is respected it would lead to mixture of TA and RacCHD.

7.5.2. A liquid solution exists between the CHD enantiomers.

7.5.2.1.Gibbs free enthalpy of a binary mixture RacCHD

and RRCHD or SSCHD presenting a partial miscibility

We know from (Leitao et al. 2002) that a solid solution is formed between RacCHD and

one of its enantiomers, RRCHD or SSCHD (see 2.3.3). We can plot the molar entropy of a mixture

of RRCHD and SSCHD according to the composition.

Figure 72 : Gibbs

molar enthalpy of a

binary RRCHD-

SSCHD

In the Figure

72, the fact that a

second polymorph of

RacCHD exists

(Lloyd et al. 2007) is

not presented. It

does not affect much the Gibbs enthalpy of CHD and add a phase for molar fraction close to 0.5.

7.5.2.2.Ternary diagram

The ternary phase diagram is more complicated when applied to solid solution and

different cases can be envisaged. The first is that CoC is formed when TA added to CHD that

contain any fraction of RRCHD or that the reaction of Equation 22 takes place for every X (X

being XSSCHD).

dnCoCRRdnXXSSCHDdndnTARRXXSSCHD ))1(,()1())1(,(

Equation 22

Appendix 144

Figure 73: “Solid solution and

CoC>solCHD” ternary phase

diagram

Figure 73 represents a

theoretical phase diagram of RRCHD,

SSCHD and TA supposing that a solid

solution is formed between RacCHD

and SSCHD or RRCHD. The central

point correspond to the composition of

the sample mr=0.5 (Molnar et al. 2008): if this phase diagram is respected it would lead to CoC and

SSCHD only.

The hypothesis taken in the previous diagram can be written using the chemical potential

and the molar enthalpy plotted in Figure 72.

If in the ternary we have 1 mole of CHD in a mixture of X mole SSCHD and 1-X mole

SSCHD (so CHD contribution to the total free energy of the system is gCHD(X) and add an infinity

decimal quantity of TA dn, dn Coc forms so the quantity of CHD has decreased to 1-dn and its

molar fraction of SSCHD is dn

X1

which is egal to X(1+dn) at the first order (Equation 22). The

variation of free enthalpy dG associated to this transformation is inferior to zero and is egal to (at

the first order):

Xgg SSCHD

CHDCHDTACoC d

dXdndndG )(

00

And we introduce the chemical potential of SSCHD (equal to RRCHD’s) we obtain:

Xg

Xg SSCHDCHDSSCHD

CHDSSCHDTACoC d

dSSCHDdn

dG ))()0( ( 000

Equation 23

Then the hypothesis of the last ternary was that dndG

would be always inferior to zero, this

is to say that the formation of CoC is always favourable for it leads to a decrease in the free energy.

Appendix 145

dndG

is a increasing function of XSSCHD that present a plateau for X=0.8 to 1. Theoretically,

there is a possibility that the sign of dndG

changes between 0.5 and 0.8.

Figure 74 : Variation of free energy of the

system CHD + dn TA when CoC forms.

If we accept that the sign of dndG

changes this means that for a certain

composition of SolCHD, that is named Xlim

(SolCHDlim correspond to its Xlim where

0dndG

) the formation of CoC is as favourable as the formation of SolCHD and free TA. Above

Xlim, CoC does not form and, below Xlim, the CoC forms. So another ternary phase diagram can be

plotted taking into account this new possibility.

Figure 75 : “Solid solution and

CoC=solCHDlim” ternary phase

diagram with solCHD 2

The central point corresponds to

the composition of the sample mr=0.5

(Molnar et al. 2008): if this phase

diagram is respected it would lead to

SolCHDlim, CoC and TA. Let’s see what

happen at this point of the ternary for 2

extreme values: if Xlim=0.5 there are only SolCHDlim (equivalent to RacCHD is this very case) and

if Xlim=0.8 there would be TA, CoC and SolCHDlim . In the last case if 1 mol of RacCHD (this is to

say 1 mol of RRCHD and 1 mol of SSCHD) and 1 mole of TA are mixed we would obtain a

SolCHD that contains 1 mol of SSCHD and 0.25 mol of RRCHD, 0.25 mol TA and 0.75 mol Coc

(0.75 mol TA + 0.75 mol RRCHD).

Appendix 146

Remark: the function dndG

indicates the difficulty of formation of CoC: if 0dndG

CoC

cannot form and, then, the lower the easier the formation of CoC. Deviation to ideality (or

thermodynamical equilibrium) is found for values of XSS close to the saturation of the liquid

solution with SSCHD. This is to say that if the system is out equilibrium, the phase diagram Figure

75 might represent a metastable phase diagram. If amorphous phase is included it have much

chance to have a composition close to the ternary eutectic.

Remark 2: A ternary phase diagram where CoC can accept a fraction of SSCHD can also

be constructed, an example is given in Figure 76. This phase diagram is purely indicative and given

just to illustrate the discussion where incorporation of SSCHD in CoC is envisaged.

Figure 76: Example of ternary phase

diagram where CoC can intake a small

fraction of SSCHD

The dotted line on the CoC+SolCHD

indicates the equilibrium between a

solCHD of a certain composition and the

CoC that contain the corresponding

amount of SSCHD. This scheme is only an

example.

7.6. Results of the experiment for the determination of the phase

diagram.

In this annex the different results of analysis for the determination of the ternary phase

diagram is presented as well as different hypothesis that allow interpretation. The explanation

retained is that the ternary phase diagram is of the type of Figure 73: “Solid solution and

CoC>solCHD” ternary phase diagram that presents the formation of a metastable compound called

“X”, and difficulties of formation of CoC. Amorphicity is also to expect.

Table 18: The different sample prepared for investigation of the ternary system, their

composition and the phase observed by XRD

Appendix 147

Table 19: Composition of the different sample if the ternary does not present solid

solution.

It would obey to the phase diagram Figure 70: “No solid solution CoC>racem” phase

diagram of RRCHD, SSCHD and TA supposing that a racemic compound is formed and

not a solid solution.

Table 20: Which deviation do we observe from the model "no solid solution CoC>racem"?

How can we interpret it? Two explanation are tested : phase diagram of the type of Figure

75 : “Solid solution and CoC=solCHDlim” ternary phase diagram with solCHD 2 or the fact

that the system is out of equilibrium and the formation of CoC does not take in place in the

full extent.

Table 21: Composition of the different sample if the ternary presents a solid solution and

the deviations observed.

It would obey to the phase diagram Figure 73: “Solid solution and CoC>solCHD” ternary

phase diagram.

*CHDG was prepared with SSTA and SSCHD. So this is the enantiomer of CoC. We fit

this result in the table as if it was done of TA and RRCHD.

Appendix 148

analytical composition phases observed by XRD name quality? sschd rrchd ta enantiomer racem TA CoC X ANA 1 no 25 0 75 yes yes ANA 2 no 50 0 50 yes yes ANA 3 no 75 0 25 yes yes ANA 4 no 66 34 0 little yes ANA 5 no 35 65 0 very little yes ANA 6 yes 16 84 0 yes yes ANA 7 yes 0 25 75 yes yes ANA 8 yes 0 50 50 trace trace mostly ANA 9 yes 0 75 25 yes yes ANA 10 yes 4 21 75 little little yes yes ANA 11 yes 8 42 50 little Very little little yes yes ANA 12 yes 12 63 25 little Very little little yes much ANA 13 no 3 3 95 yes yes no Very little ANA 14 no 8 8 85 yes yes little ANA 15 no 25 25 50 little yes yes yes ANA 16 no 43 43 15 little yes little yes ANA 17 no 48 48 5 yes trace ANA 19 yes 73 27 0 little yes ANA 20 yes 25 0 75 yes yes ANA 21 yes 50 0 50 yes yes ANA 22 yes 75 0 25 yes yes ANA 23 yes 18 7 75 little yes much ANA 24 yes 37 13 50 not much yes yes ANA 25 yes 55 20 25 little yes little ANA 26 yes 10 45 45 yes not much a lot ANA 27 yes 50 25 25 yes yes a lot yes yes ANA 28 yes 80 10 10 yes a lot yes little chdg* yes 50 50 yes yes mr=0.5 low qual. 33 33 33 yes yes little yes

Appendix 149

the simpler model: CoC is formed then TA then racCHD then the free enantiomer left for 100 mol of RRCHD + SSCHD + TA in % in %w name CoC TA racCHD RRCHD SSCHD CoC TA racCHD RRCHD SSCHD CoC TA racCHD RRCHD SSCHD ANA 1 0 75 0 0 25 0 75 0 0 25 0 79 0 0 21 ANA 2 0 50 0 0 50 0 50 0 0 50 0 56 0 0 44 ANA 3 0 25 0 0 75 0 25 0 0 75 0 30 0 0 70 ANA 4 0 0 34 0 31 0 0 52 0 48 0 0 69 0 31 ANA 5 0 0 35 30 0 0 0 54 46 0 0 0 70 30 0 ANA 6 0 0 16 67 0 0 0 19 81 0 0 0 33 67 0 ANA 7 25 50 0 0 0 33 67 0 0 0 47 53 0 0 0 ANA 8 50 0 0 0 0 100 0 0 0 0 100 0 0 0 0 ANA 9 25 0 0 50 0 33 0 0 67 0 53 0 0 47 0 ANA 10 21 54 0 0 4 26 68 0 0 5 39 57 0 0 3 ANA 11 42 8 0 0 8 72 14 0 0 14 84 9 0 0 7 ANA 12 25 0 12 26 0 40 0 19 41 0 53 0 23 24 0 ANA 13 3 92 0 0 3 3 95 0 0 3 4 94 0 0 2 ANA 14 8 77 0 0 8 8 84 0 0 8 14 80 0 0 6 ANA 15 25 25 0 0 25 33 33 0 0 33 50 28 0 0 22 ANA 16 15 0 28 0 15 26 0 48 0 26 33 0 53 0 14 ANA 17 5 0 43 0 5 10 0 81 0 10 11 0 84 0 5 ANA 19 0 0 27 0 46 0 0 37 0 63 0 0 54 0 46 ANA 20 0 75 0 0 25 0 75 0 0 25 0 79 0 0 21 ANA 21 0 50 0 0 50 0 50 0 0 50 0 56 0 0 44 ANA 22 0 25 0 0 75 0 25 0 0 75 0 30 0 0 70 ANA 23 7 68 0 0 18 7 73 0 0 20 13 72 0 0 15 ANA 24 13 37 0 0 37 16 42 0 0 42 27 41 0 0 32 ANA 25 20 5 0 0 55 25 6 0 0 69 43 6 0 0 51 ANA 26 45 0 0 0 10 82 0 0 0 18 91 0 0 0 9 ANA 27 25 0 0 0 50 33 0 0 0 67 53 0 0 0 47 ANA 28 10 0 0 0 80 11 0 0 0 89 22 0 0 0 78 chdg* 50 0 0 0 0 100 0 0 0 0 100 0 0 0 0 mr=0.5 33 0 0 0 33 50 0 0 0 50 70 0 0 0 30

Appendix 150

How it can be explained? name ok? comments pb formation coc solid solution ANA 1 yes ANA 2 yes ANA 3 yes ANA 4 no not enough enantiomer yes ANA 5 no not enough enantiomer yes ANA 6 yes ANA 7 yes ANA 8 yes ANA 9 yes ANA 10 no X instead TA no no

ANA 11 yes but with X no no

ANA 12 no X instead racem and RR no ANA 13 no no CoC but racem yes no

ANA 14 no not enough CoC too much racem yes no

ANA 15 no racem but no SSCHD partially partially ANA 16 yes ANA 17 no no CoC but racem yes no ANA 19 no enan missing no yes ANA 20 yes ANA 21 yes ANA 22 yes ANA 23 no no coc but racem yes partially ANA 24 no not enough enantiomer/coX yes yes ANA 25 no no coc but racem yes partially ANA 26 yes ANA 27 no too much racem yes partially ANA 28 no no coc but racem yes partially chdg* no X! no no mr=0.5 no racem and little TA unexpected yes partially

Appendix 151

with solid solution limit 0.2 < Xss < 0.8) in %w name CoC TA racsolsolCHD RRCHD SSCHD ok? explanation? ANA 1 0 79 0 0 21 yes ANA 2 0 56 0 0 44 yes ANA 3 0 30 0 0 70 yes ANA 4 0 0 100 0 0 no but better trace of enantiomer ANA 5 0 0 100 0 0 no but better trace of enantiomer ANA 6 0 0 82 19 0 yes ANA 7 47 53 0 0 0 yes ANA 8 100 0 0 0 0 yes ANA 9 53 0 0 47 0 yes ANA 10 39 57 0 0 3 no X ANA 11 84 9 0 0 7 yes ANA 12 53 0 47 0 0 no X ANA 13 4 94 0 0 2 no no CoC formation ANA 14 14 80 0 0 6 no bad CoC formation ANA 15 50 28 0 0 22 no bad CoC formation ANA 16 33 0 67 0 0 yes ANA 17 11 0 89 0 0 no no CoC formation ANA 19 0 0 100 0 0 no but better trace of enantiomer ANA 20 0 79 0 0 21 yes ANA 21 0 56 0 0 44 yes ANA 22 0 30 0 0 70 yes ANA 23 13 72 0 0 15 no no CoC formation ANA 24 27 41 0 0 32 no no CoC formation ANA 25 43 6 0 0 51 no no CoC formation ANA 26 91 0 0 0 9 yes ANA 27 53 0 0 0 47 no bad CoC formation ANA 28 22 0 0 0 78 no no CoC formation chdg* 100 0 0 0 0 no X mr=0.5 70 0 0 0 30 no bad CoC formation

Bibliography 152

7.7. Diffractogram of samples prepared according to different methods

Table 22: Sample for sample preparation prepared in different condition

Rbf: the evaporation of the ethanol in the second step in done over night in the same

round-bottomed flask (rbf) as for the rotoevaporation. pd: the evaporation of the ethanol in

the second step in done over night in a petri dish (pd) or beaker (bk).

Sam

ple

nam

e

Con

ditio

n of

ev

apor

atio

n of

eth

anol

Tem

pera

ture

of

roto

evap

orat

ion

(°C

)

Furth

er e

vapo

ratio

n

perfi

l

mr

Muc

h m

orph

ous

phas

e by

vis

ual o

bser

vatio

n?

Obs

erva

tion

by X

RD

75°C Rotoevap. 75 rbf no 0.5 no Much CoC and enantiom Little racem

55°C Rotoevap. 55 rbf no 0.5 no TA and SolCHD Traces of CoC

40°C Rotoevap. 40 rbf no 0.5 no Only TA and SolCHD

40°C Rotoevap. 40 rbf yes 0.5 no Only TA and SolCHD

30°C Rotoevap. 30 rbf no 0.5 yes Much CoC and SolCHD little enantiom. no TA

40°C Rotoevap. 40 rbf no 1 no Only TA and SolCHD

40°C Rotoevap. 40 rbf yes 1 no TA and SolCHD little CoC

40°C Rotoevap. 40 pd no 0.5 no Much CoC Little Enantiom

40°C Rotoevap. 40 pd yes 0.5 no CoC and Enantiom No SolCHD

Pat 6 none bk no 0.5 yes Much racem, TA traces of Coc no enantiom

Pat 7 none rbf no 0.5 yes Much CoC, racem, trace of enantiom, no TA

preparation of enantiomers using high- pressure technologies · preparation of enantiomers using...

Documents