Professional Documents
Culture Documents
Microwave Heating in Fine Chemical Applications:: Role of Heterogeneity
Microwave Heating in Fine Chemical Applications:: Role of Heterogeneity
role of heterogeneity
Proefschrift
Ditproefschriftisgoedgekeurddoordepromotoren: prof.dr.L.A.Hulshof en prof.dr.J.Meuldijk Copromotor: dr.J.A.J.M.Vekemans TheresearchdescribedinthisthesiswasfinanciallysupportedbySenterNovem. Project:FCSMART Druk:Universiteitsdrukkerij,TechnischeUniversiteitEindhoven CoverDesign:DymphDressen AcataloguerecordisavailablefromtheEindhovenUniversityofTechnologyLibrary. ISBN:9789038618210 Trefwoorden:microwaveassistedchemistry,microwaveeffect,heterogeneoussystems, multimode,flowchemistry,continuousflowprocessing,fumehoodconcept,(trans) esterification,biocatalysis,racemization,Nacetylaminoacids,ureidopyrimidinone, iron(III)nitrateoxidation,scalingup.
Glossary
BINAP C CALB CFR CH CLEA cp CSTR DKR DMA dp EDTA ee fMW HDI HI MTBE MW NMP PEEK PPL R (rMW)t=0 rt RTD T tan Teflon THF Upy Weflon solnH
2,2BIs(diphenylphosphino)1,1biNAPhthyl Cooler CandidaantarticalipaseB Continuousflowreactor(FlowSYNTH) Conventionalheating Crosslinkedenzymaticaggregate Heatcapacity Continuousstirredtankreactor Dynamickineticresolution N,NDimethylacetamide Penetrationdepth Ethylenediaminetetraaceticacid Enantiomericexcess Microwaveeffect Hexamethylenediisocyanate Hexylisocyanate Methyltert.butylether Microwave(heating) NMethylpyrrolidinone Polyetheretherketone PorcinePancreaslipase Reactor Initialreactionrateformicrowaveheating Roomtemperature Residencetimedistribution Temperature(sensor) Losstangent Poly(tetrafluoroethylene)(BrandnameofDuPont) Tetrahydrofuran Ureidopyrimidinone Teflonincorporatedwithcarbon(25%wt) Heat(enthalpy)ofsolution Dielectricconstant Dielectricloss
Table of content
Table of contents
Chapter 1
1.1 1.2 1.3 1.4 1.5 1.6 1.7
Introduction
Changingchemistry icrowaveassistedchemistry M icrowaveirradiation M icrowaveequipment M imofthethesis A Outlineofthethesis Referencesandnotes
page
3 3 6 10 11 12 13
Chapter 2
Chapter 3
3.1 3.2 3.3 3.4 3.5 3.6
Chapter 4
4.1 4.2
69 70
Table of content
77 80 82 83 85
Chapter 5
5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9
Chapter 6
6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9
Chapter 1
Introduction
Chapter 1
Abstract
Microwave or dielectric heating is an accepted new technique to run chemical conversions on a lab scale as an alternative heating principle. Microwave heating is believed to be more depending on the molecular properties and the reaction conditions than conventional heating. At first sight microwave-assisted chemistry looks very promising given the overwhelming number of publications during the last two decades. To predict its potential for industrial purposes, it is necessary to make a proper comparison between both heating techniques. The underlying theory, Green Chemistry and the microwave effect are relevant topics to be studied in detail in nowadays organic chemistry. According to literature microwaveassisted chemistry seems to perform well and sometimes it even works better than expected. Magic or hidden logic?
Chapter 1
Chapter 1
widespread application of microwave heating throughout numerous types of chemical transformations. It is nearly a Herculean effort to give a complete overview where microwave heating has been applied to heat up chemical reactions.216Evenbelowroomtemperaturemicrowaveirradiationisusedtoruna reactionincombinationwithsimultaneouscooling.17,18 One of the first reactions described is the hydrolysis of benzamide in aqueous sulfuric acid (see Scheme 1.1).19 The use of polar components facilitates the absorptionofmicrowaveenergyandconsequentlytheconversionintoheat.With the use of a sealed vessel it was possible to increase the temperature and eventuallythepressureandthereforereactionrates.Thereductionofreactiontime and the growth in chemical yield were seen as the advantages of microwave heating.
O NH2
20% H2SO4
O OH
Scheme1.1 Hydrolysisofbenzamideinaqueoussulfuricacid. Workinginaqueousmediaorapplyingsolidphasechemistry(drymedia)were possible solutions to prevent explosion in the domestic machines. This way of workingwaslaterrecognizedasonewayofGreenChemistry. GreenChemistry.Basedonareport(OurCommonFuture,1987)fromUnitedNations World Commission on Environment and Development (WCED)20 sustainability was acknowledged as an important development. Green Chemistry is a sciencebased, non regulatory, economically driven approach toward sustainable development.2123 This advance has led to a program supported by U.S. Environmental Protection Agency, including a Presidential Green Chemistry Challenge Award since 1996. Later the twelve principleswereaddedtobringtheoryintoamorepracticalperspective,seeTable1.1. Table1.1 The twelve principles of Green Chemistry by U.S. Environmental ProtectionAgency.
PrinciplesofGreenChemistry
No 1 2 3 4 5 6 Preventwaste Designsaferchemicalsandproducts Designlesshazardouschemicalsyntheses Userenewablefeedstock Usecatalysts,notstoichiometricreagents Avoidchemicalderivatives No 7 8 9 10 11 12 Maximizeatomeconomy Usesafersolventsandreactionconditions Increaseenergyefficiency Designchemicalsandproductstodegradeafteruse Analyzeinrealtimetopreventpollution Minimizethepotentialforaccidents
Chapter 1
However, the chemical reaction itself is only a small part of the entire process to obtain products pure. In many cases workup is more laborious and waste producing than the chemical conversion itself (e.g. extractions or pHshifts). The overall greenish color is, therefore, questionable. The energy consumption is another parameter to determine whether microwave heating is environmentally friendly. In1986severalDielsAlderreactions(e.g.seeScheme1.2)wereinvestigatedinthe microwaveoven.24Thepaperdemonstratesnicelythestruggletomakemicrowave irradiationsuitablefororganicchemistry,involvingsmartsolutionstomeasurethe temperature inside the oven and even including an example of an exploding reactionvessel.
O O O O
p-xylene
O O
Scheme1.2 ADielsAlderreactionofanthracenewithmaleicanhydride.24 Twoadditionalattractiveareastostudymicrowaveheatinghavebeenbiocatalysis and the application of ionic liquids. Both, classified under Green Chemistry, were expected to show synergy with microwave heating.25,26 Important factors to workwithionicliquidsaretheirhighboilingpointsandlowvolatility.Extremely fast heating rate curves (up to 10 C per second) are observed for ionic liquids under microwave irradiation, due to the presence of ions which strongly interact with the electromagnetic waves. Scheme 1.3 shows a Heck reaction performed in anionicliquidasanexample.27
O X O
+
R
X = Br, I
PdCl2
OBu R
OBu
Scheme1.3 Heckreactioninanionicliquidheatedbymicrowaveirradiation. Alsothesynthesisofionicliquidsisverysuccessfulwithmicrowavetechnology.A recentexamplefromsupramolecularchemistryapplicationsisdepictedinScheme 1.4.Thisisatypicalpublicationwhichdemonstratesthepossibilityofasuccessful switch to an alternative heating source, albeit with lack of accurate reaction/temperature control by using a domestic oven.28 A critical view is necessarytomakeafaircomparisonwithconventionalheating.Thispointneedsto
Chapter 1
beaddressedinwhichreactionconditionsarereallycomparable.Independenton thereactionconditions(e.g.heterogeneousvs.homogeneousmixturesoropenvs. closedsystems)acorrecttemperaturemeasurementofthereactionmixturehasto be ensured. The absence of temperature control or the usage of domestic ovens (multimode machine, see chapter 1.4) till 2001 characterized the majority of published microwaveassisted chemistry. Nonstirrable reaction mixtures (e.g. solventfreeconditionswithahighpercentageofsolids)inamicrowaveovengive irreproducible results upon duplication or lead to misinterpretation when comparedwiththeconventionallyheatedexperiments.
NC CN
O(CH2)nCH2Br
N
m
NMP, MW NC CN
+
Br-
N
m
Scheme1.4
Formationofanionicliquidbymicrowaveirradiation.28
Chapter 1
wavelength (m)
10
10
10
10
10
-1
10
-2
10
-3
10
-4
10
-5
10
-6
10
-7
10
-8
10
-9
10
-10
10
-11
10
-12
Infrared
10
10
10
10
10
10
10
10
13
10
14
10
10
16
10
17
10
18
10
19
10
20
10
-8
10
-7
10
-6
10
-5
10
-4
10
-3
10
-2
10
-1
10
10
10
10
10
10
10
H CI
12.2 cm 2.45 GHz -5 ~10 eV
Figure1.1 Table1.2
Energyofvariousintermolecularbondsincomparisontotheenergyofa microwavephoton. Microwave photon Hydrogen Covalent Ionic (2.45GHz) bond(H)* bond(C)* bond(I)* Energy(eV) ~105 ~0.04to0.44 ~3.82(CC) ~7.6
* intermolecularbondsarealsodepictedinFigure1.1withthecorrespondinglettersH/C/I.
Heatingreliesontheabilityofanelectricfieldtopolarizechargesandtheinability of these charges to follow rapid reversals of an electric field. (Dielectric) polarizationisthesumofmultipletypesofpolarizations,seeEquation1.1. t=e+a+d+i (1.1)
t=totalpolarization e=electronicpolarization a=atomicpolarization d=dipolarpolarization i=interfacialpolarization
Electronic polarization arises from the alignment of electrons around specific nuclei.Atomicpolarizationresultsfromtherelativedisplacementofnucleidueto the unequal distribution of charges within the molecule. Dipolar polarization results from the orientation of permanent dipoles by the electric field while
Chapter 1
interfacialpolarization(MaxwellWagnereffect)occurswhenthereisabuildupof chargesatinterfaces.Thefirsttwo(electronicandatomic)polarizationeffectsare much faster than the microwave frequencies and, therefore, these effects do not contributetothedielectricheatingeffect. Successively to electronic and atomic polarization, dipolar polarization (in the range of microwave frequencies) results in polarization of dipoles which lags behindthechangesoftheelectricfield.Thelagindicatesthatthesampleabsorbs energy from the field and is heated. Two parameters define the dielectric properties of materials; dielectric constant () and dielectric loss (). The first parameterindicatestheabilityofamoleculetobepolarized,thesecondonegives insight in the efficiency of converting radiation into heat. The ratio of these quantities(/)isthelosstangent(=tan).Thelosstangentdefinestheabilityof amaterialtoconvertelectromagneticenergyintoheatenergyatagivenfrequency andtemperature(seeTable1.3). Table1.3 Losstangentofseveralliquids/materials solvent solid/liquid losstangenta losstangent (temperature) ethyleneglycol 1.350 ethanol 0.941 ice 0.00027(0C)d DMSO 0.825 water 0.207(0C)b methanol 0.659 water 0.123(20C)a NMP 0.275 water 0.097(25C)b aceticacid 0.174 water 0.056(50C)b water 0.123 material chloroform 0.091 losstangentc tetrahydofuran 0.047 dichloromethane toluene hexane 0.042 0.040 0.020 Teflon fusedquartz 0.00028 0.00006
a:20C,2.45GHz[reference30];b:2.0GHz[reference31];c:20C,3GHz[reference32]; d:2.45GHz[reference33]
Forexamplewithininterfacialpolarization,asuspensionofconductingparticlesin anonconductingmediumisaheterogeneousmixturewhosedielectricconstantis frequency dependent. The loss relates to the buildup of charges between the interfaces. Its importance in the microwave region has not been welldefined. Absorption centered around a frequency of 107 Hz may tail into this frequency range.Forhighlyelectricalconductiveliquidsandsolidsapointisreachedwhere
Chapter 1
the conductive loss effects are larger than the dipolar relaxation effects (large amountofsalts,hightemperature). Thefinaltemperatureriseisdeterminedbythelosstangent,specificheatcapacity, emissivityofthesampleandthestrengthoftheappliedfield.Alltheseproperties are temperature dependent. As the composition of the reaction mixture develops withtime,allinteractionswithmicrowaveirradiationwillalsochange.Thismakes the complete system very dynamic. The loss tangent is also determining the penetrationdepthofmicrowavesintothereactionmixture.Forwaterasamedium microwaveabsorberthepenetrationisonlyafewcentimeters.Thislimitationisa determiningfactortoassessthescalingupofbatchprocesses. Microwaveeffect By studying microwave irradiation it is impossible to leave the discussion about microwave effects untouched. In the scientific world the microwave effect is a collective noun for all the positive contributions to reaction rate enhancement whichmicrowaveirradiationhasbroughtup.Inmostcasesthereisalackoffair comparisonbetweenorgoodcontrolofreactionconditions.Attemptstoreproduce reactions claiming beneficial microwave effect have mostly led to a reasonable explanation, namely higher temperatures.34,35 The overview of all claimed microwave effects can simply be divided into two basic parts: thermal and non thermal.3638 Good temperature control during the reaction and a known heating profile will preclude the existence of most nonthermal effects on a macroscopic scale (see Figure 1.2),16 although the heating rate in some cases is irreproducible during conventionalheating.
O NH2 O N H
O H H
Figure1.2
AsymmetricMannichreaction.
*[reference39],reactionatconstantMWpower **[reference16],reactionatconstanttemperature
Thiscoulderroneouslybeinterpretedasanonthermaleffect.Overheatingcanalso be observed under special conditions. Due to the absence of nucleation, boiling
Chapter 1
points of pure solvents can increase up to thirty degrees.40 This increase in temperature can influence reflux conditions depending on the heating mode. In otherwordsrefluxconditionsundermicrowaveheatingwithahighpowerinput induce a higher conversion rate as compared to a conventionally heated experiment under reflux conditions. Likewise proper temperature control will revealthepresenceofathermaleffect. Anyunexplainableeffectsarealmostautomaticallytransferredtothenonthermal part. This is also the part which embraces many speculations and theories. The debate is currently ongoing and the field can be divided into two camps: those who advocate the existence of nonthermal effects and those who accept only purelythermaleffectsasexplanation.
10
Chapter 1
The limitation of the monomode machine is the lack for proper scalingup potential.Thesizesofthecavityremainrathersmallindimensionsandtechnical solutions within this area of interest are not available. This brings us almost automaticallytotheothertype,namelythemultimodemicrowaveoven. In principle both types have a magnetron which produces the electromagnetic waves. These waves are emitted by an antenna indirectly into the cavity which holdsroomtoperformsomechemistryasintheordinarydomesticmodels.Inthe case ofa multimode the waves are scattered with a mode stirrer throughout the cavity (which is considerably larger than the monomode). Reflection against the wallsandextinctionoramplificationinsidethecavitymayoccur.Thisisnotideal forsmallsamples,butitcreatesroomforlargerperspectivessuchaseithervessels orincorporationofacontinuouslyoperatedsetup.Eithernormalglassware,quartz or PTFE reactors can be inserted into the cavity in combination with stirring and temperature control by an internal fiber optic device, gaspressure sensor or an infrared sensor. The boundaries of performance are 300 C and 100 bar for some setups.Thesuppliersinthismarket(includingmonomode)arelimited,seeFigure 1.4.
Source:EvalueserveAnalysis
Figure1.4
11
Chapter 1
effects demonstrating its added value in process research. It would facilitate the study on the feasibility of a microwave oven integrated fume hood setup for producinglargervolumesoffinechemicals.
12
Chapter 1
13
Chapter 1
27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44
Vallin,K.S.A.;Emilsson,P.;Larhed,M.;Hallberg,A.J.Org.Chem.,2002,67,6243. Pal,S.K.;Kumar,S.TetrahedronLett.,2006,47,8993. Kingston, H. M.; Haswell, S. J.; Ed. Microwaveenhanced chemistry: fundamentals, samplepreparation,andapplications;ACS:WashingtonD.C.,1997. Gabriel,C.;Gabriel,S.;Grant,E.H.;Halstaed,B.S.J.;Mingos,D.M.G.Chem.Soc. Rev.1998,27,213. Kaatze,U.J.Chem.Eng.Data,1989,34,371. http://www.rfcafe.com/references/electrical/dielectricconstantsstrength.htm Luque de Castro, M. D.; Luque Garcia, J. L. Acceleration and automation of solid sampletreatment;Elsevier,2002. Herrero,M.A.;Kremsner,J.M.;Kappe,C.O.J.Org.Chem.2008,73,36. Westaway, K.C.; Gedeye, R.N. J. Microwave Power and Electromagn. Energy, 1995, 30,219. Jacob,J;Chia,L.H.L.;Boey,F.Y.C.J.Mater.Sci.,1995,30,5321. Stuerga.D.A.C.;Gaillard,P.J.MicrowavePowerandElectromagn.Energy,1996,31, 87. Stuerga.D.A.C.;Gaillard,P.J.MicrowavePowerandElectromagn.Energy,1996,31, 101. Rodrguez,B.;Bolm,C.J.Org.Chem.,2006,71,2888. Gedeye,R.S.;Smith,F.E.;Westaway,K.C.Can.J.Chem.,1998,66,17. http://www.cem.com/synthesis/mwbscs.asp http://www.milestonesci.com http://www.antonpaar.com http://www.biotage.com
14
Chapter 2
vs.
Chapter 2
Abstract
Several heterogeneous reactions, for which microwave effects are claimed, have been re-evaluated. Biocatalyzed (trans)esterifications were performed under standardized conditions and this showed similar time - conversion histories and selectivities for both, microwave or conventional heating. In the piperidine/piperazine formation from 1,5-dibromopentane or bis(2-chloroethyl)amine.HCl the observed rate enhancement with microwave heating was completely counterbalanced by proper temperature control. The oxidation of benzyl alcohol with iron(III)-nitrate to benzaldehyde was studied in detail. This led to the observation of identical reaction rates when conventional heating was substituted by microwave heating. Also the heterogeneous Williamson synthesis of methyl gallate with dodecyl bromide afforded identical reaction rates for both heating techniques. In the only industrially applied, scaled-up process - the Laurydone synthesis, transforming pyroglutamic acid into its decyl ester - vigorous stirring guaranteed a homogeneous temperature distribution in the reactor. With this approach the claimed microwave effect vanished. In the aforementioned process types the claims regarding rate enhancement by microwave heating have been rationalized by differences in reaction temperature. However, the results also suggest that heterogeneity in itself may have profound influence on the rate of microwave-heated experiments.
16
Chapter 2
2.1 Introduction
Reactionsystemsareeitherheterogeneousorhomogeneous.Inviewofspacetime volume efficiencies, high concentrations are preferred over dilute reaction conditions. Chemical reactions tend to run faster when all components are dissolved tothe same extent. Liquidshandlingis generallylesscomplicated than solidshandling,butthelatterisnotaninsurmountableobstacle.Safetyissuesmay, however, requiredilute reaction mixtures. Unfortunately, there is no general rule toperformchemicalreactions.Statistically,heterogeneousreactionsareinvolvedin the majority (> 75 %)1 of fine chemical processes. Associated with the chemical reaction is the workup of the reaction mixture. The more undesired components areintroduced,themoreeffortisneededtoremovethesecomponents,e.g.alsothe removalofadditionalsolventbydistillation.Alltogetherourattentionwillmainly befocusedonheterogeneousreactionsystems.
Figure2.1 Table2.1
entry 1 2 3
Heterogeneoussystems;solidliquid,liquidliquidandgasliquid. Subdivisionofheterogeneousreactionsdiscussedinthischapter.
reactiontype Biocatalysis(2.3) Williamsonsynthesis(2.6) Piperidineformation(2.4) Piperazineformation(2.4) Laurydoneprocess(2.7) Oxidationofbenzylalcohol(2.5)
*Thissubdivisionisbasedoninitialreactionconditions
Each of these three combinations needs appropriate mixing to ensure efficient contactbetweenthereactants.Masstransferlimitationmaystronglyinfluencethe
17
Chapter 2
reaction rates. If the interfacial area is enlarged to decrease the resistance against masstransfer,thereactiontimeisreducedinmostcases. Theultimatelydesiredcombinationistooperateaconcentratedmixturewithhigh reaction rates under safe conditions. How good is the fit of the involved parameters (e.g. temperature, time or concentration) with the setup of a continuousflow reactor heated by microwave irradiation compared to conventional methods? Before answering this question our research was initiated with a number of heterogeneous chemical reactions to validate the occurrence of anybeneficial(microwave)effectswhichhadbeenclaimedinliteraturereports.
2.3 Biocatalysis
Our life should not exist without biocatalysis (i.e. catalysis by enzymes). In an organism enzymes catalyze numerous processes. Enzymes, complexes of amino acidsunitedinapolypeptidechain,canbeseenasspecialistswhichdemonstrate highchemo,regioorstereoselectivity.Anumberofthesecatalystsareappliedin commercialchemicalprocessesorareanessentialpartofaperformanceproduct, e.g.lipases(afamilywithintheenzymes)inwashingpowder. More modified or dedicated enzymes are nowadays accessible to the organic chemists. Novozym 435 is a widespread lipase (Candida antarctica lipase B) immobilized on an acrylic resin. Novozym 435 catalyzes esterifications, transesterificationsandreactionsbasedonaminolysis(Scheme2.1).
Asp187 O H N O Ser105 O O R' O 1. O R" O R 4. 2. R' OH O Ser105 N N O R R Asp187 O H O Ser105 O R O R
Oxyanion hole
N H
N + N H O
His224
His224
Asp187 O
3. O Ser105 H O R O R"
Oxyanion hole
Asp187 OH O
R"
N + N H O
His224
His224
Scheme2.1
18
Chapter 2
Theimplementationofabiocatalystinasyntheticrouteimprovestheaccessibility ofchiralintermediatesorpharmaceuticallyactivecompounds. Synergistic effects of microwave heating and biocatalysis were claimed in literature.3,4Rateenhancementsuptoafactorof140werereportedinseveralcases and in general the combination of microwave heating and biocatalysis seems to work better than conventional heating. A review from Besson et al. gives a good pictureofthecurrentclaims.5Reproductionofsomespecificexampleswasthefirst step involving biocatalysis as a mediator for kinetic resolution. A followup with dynamickineticresolution(DKR)shouldhavebeenalogicalnextapproach.6 The DKR concept is a smart tool for realizing 100 % yield and 100 % enantiomeric purity in the production of fine chemicals and this concept will be elaborated in section3.1. Combinations of alcohols and esters as well as combinations of alcohols and carboxylicacidshavebeenmadetoperform(enantioselective)transesterifications and esterifications with various types of enzymes. First the (enantioselective) transesterificationisdiscussedinthenextparagraph. Transesterification In the classical transesterification three substitution locations at the substrates (R, RandR)areavailabletointroduceawidediversityof(by)products(Scheme2.2 (top)).Thisdiversitymakesthetransesterificationanattractivereactioninamulti step synthesis. A common substrate for the study of kinetic resolutions is 1 phenylethanolfromwhichesterscanbemade(seeScheme2.2(bottom),Table2.2 or Figures 2.2 2.10). Especially the formed alcohol enables the design of an irreversibleprocess.
R O O R O O R' R'
+ +
OH OH (R) OH (S)
R''
O O
R'
+ +
R OH
R''
O O
R'
R OH R'' OH (S)
Scheme2.2
AcombinationofestersandsubstratealcoholshasbeentestedusingNovozym435 asbiocatalyst(entries1to6inTable2.2).Thecombinationsinentries1and2have been reported by Loupy and Petit.7 However, the procedure of Loupy and Petit7 proved to be irreproducible and thermally uncontrollable. In their setup the biocatalyst was used as a solid carrier for the substrates by impregnation via an etherealsolution.Theamountofcatalysttheyusedwasreconsideredandamore
19
Chapter 2
practicalandscalablequantitywasappliedinourwork.Theseauthorshavealso reported the use of isopropenyl acetate as an ester (entry 5), but instead of biocatalystNovozym435theyusedanotherlipase. Thecombinationinentry10hasbeenreportedbyLinandLinusinganotherlipase typethanNovozym435.8,9Porcinepancreaslipase(PPL)iscommerciallyavailable, albeitwithalowactivityandnotstabilized.Thereactivitywasalsotestedwith1 phenylethanol,seeentry9. From the group of Park it was known that subtilisin has an opposite enantio selectivitycomparedtoNovozym435andPPL.10Asmallamountofsubtilisinwas obtained from CLEA Technologies.11 Crosslinkage of enzymes by aggregation (CLEA) also stabilizes the enzymes with preservation of most of the enzymatic activity.AlthoughParkscoworkersdidnotperformthereactionwithmicrowave heating, microwave as well as conventional heating were applied in our work (entry 7). The same combination of esters and alcohols has been tested with subtilisin(entry8)andwithNovozym435(entry4). Table2.2 Biocatalytic enantioselective transesterification of (R,S)1phenylethanol (A) and (R,S)1,2,3,4tetrahydro1naphthol (B) with various esters by microwaveheating.
entry 1 2 3 4 5 6 7 8 typeof biocatalyst* Novozym 435 alcohol (R,S)
A A A A A B A A
conv. (%)
42 38** 47 18 48 43 39 4.4
Subtilisin
A 9 PPL vinylacetate 6.4 1376 B 10 vinylacetate 10 375 * Biocatalyst: Novozym435=CandidaAntarticalipaseB,immobilizedonresin; Subtilisin=NovozymeAlcalase,acrosslinkedaggregate9; PPL=PorcinePancreaslipase,freeenzyme. ** Conventionalheating *** 2,2,2Trifluoroethylbutyrate
20
Chapter 2
50
50
25
25
Time (min)
Time (min)
Figure2.2
Transesterification of vinyl acetate with (R,S)1phenylethanol by Novozym 435 in toluene at 70 C (left) in oil bath or (right) by microwaveheating;(R)1phenylethylacetate();(S)1phenylethyl acetate (); (R)1phenylethanol () and (S)1phenylethanol (),seeTable2.2,entry3.
50
50
25
25
0 0 15 30 45 60
0 0 15 30 45 60
Time (min)
Time (min)
Figure2.3
Transesterification of isopropenyl acetate with (R,S)1phenylethanol by Novozym 435 in toluene at 70 C in (left) oil bath or (right) by microwaveheating;(R)1phenylethylacetate();(S)1phenylethyl acetate (); (R)1phenylethanol () and (S)1phenylethanol (),seeTable2.2,entry5.
21
Chapter 2
50 50
25
25
0 0 15 30 45 60
Time (min)
Time (min)
Figure2.4
Transesterification of isopropenyl acetate with (R,S)1,2,3,4tetrahydro 1naphthol by Novozym 435 in toluene at 70 C in (left) oil bath or (right) by microwave heating; (R)1,2,3,4tetrahydro1naphthyl acetate (); (S)1,2,3,4tetrahydro1naphthyl acetate (); (R)1,2,3,4 tetrahydro1naphthol()and(S)1,2,3,4tetrahydronaphthol( ),seeTable2.2,entry6.
50 50
25
25
Time (min)
Time (min)
Figure2.5
Transesterification of 2,2,2trifluoroethyl butyrate with (R,S)1 phenylethanolbySubtilisininTHFat40Cin(left)oilbathor(right) by microwave heating; (R)1phenylethyl butyrate (); (S)1 phenylethyl butyrate (); (R)1phenylethanol () and (S)1 phenylethanol(),seeTable2.2,entry7.
22
Chapter 2
50 50
25
25
Time (min)
Time (min)
Figure2.6
Transesterification of vinyl acetate with (R,S)1phenylethanol by Subtilisin in THF at 40 C in (left) oil bath or (right) by microwave heating; (R)1phenylethyl acetate (); (S)1phenylethyl acetate (); (R)1phenylethanol () and (S)1phenylethanol (), seeTable2.2,entry8.
50
50
25
25
Time (min)
Time (min)
Figure2.7
Transesterificationofvinylacetatewith(R,S)1phenylethanolbyPPLin toluene at40 C in (a) oil bath or (right) by microwaveheating;(R)1 phenylethyl acetate (); (S)1phenylethyl acetate (); (R)1 phenylethanol () and (S)1phenylethanol (), see Table 2.2, entry9.
23
Chapter 2
50 50
25
25
Time (min)
Time (min)
Figure2.8
Transesterification of vinyl acetate with (R,S)1,2,3,4tetrahydro1 naphthol by PPL in benzene at 40 C in (left) oil bath or (right) by microwave heating; (R)1,2,3,4tetrahydro1naphthyl acetate (); (S)1,2,3,4tetrahydro1naphthylacetate();(R)1,2,3,4tetrahydro 1naphthol () and (S)1,2,3,4tetrahydro1naphthol (), see Table2.2,entry10.
Esterification Besides enantioselective transesterifications also biocatalytic esterifications have beenclaimedtoundergoapositivemicrowaveeffect.Consequently,toverifythese microwave effects the work of Yadav and Lathi has been reproduced in two cases,3,4 namely noctanol and nbutanol in combination with adipic acid (see Scheme 2.3, Table 2.3 and Figure 2.9). nOctanol should have an initial rate enhancement factor of over 2.2. The presence of a double carboxylic acid functionality allowed to study selectivity for mono or doubleesterification for both heating methods. However, only monoesterification was observed. Other studiesonesterificationsusinglipozymeRMIM(LRI)claimingacceleratingeffects formicrowaveheatinghavebeenreportedbyYadavetal.12
O HO O R-OH = n-butanol or n-octanol OH
R OH
O RO O OH
Scheme2.3
MonoesterificationofadipicacidwiththeaidofNovozym435.
24
Chapter 2
Table2.3
entry
Biocatalyticesterificationofadipicacidat60C(conventionalheating).
alcohol acid conversion(%) time(min) 180 180
typeof biocatalyst*
100 100
75
75
Conversion (%)
50
Conversion (%)
0 60 120 180
50
25
25
0 0 60 120 180
Time (min)
Time (min)
Figure2.9
Esterification of adipic acid with noctanol by Novozym 435 in 1,4 dioxaneat60C(left)byconventionalheatingor(right)bymicrowave heating;noctylester(),seeTable2.3,entry1.
Besides enhanced reaction rates, an improvement in enzyme stability by microwaveirradiationwasclaimedbythegroupofBesson.13 Theenzymestability has been elaborated only briefly. The enzyme activity has been quantified by the degreeofconversionafterapresetperiodof3h(seeTable2.4),buthasalsobeen monitoredduringthisperiod(Figure2.10). Table2.4 Effect of catalyst reuse (4 times) for transesterification of vinyl acetate with (R,S)1phenylethanol catalyzed by Novozym 435 at 70 C in toluene;(a)conventionaland(b)microwaveheating.
entry 1 2 3 4 run 1 2 3 4 conversion(%) (a)(b) 5149 4950 4647 4141 totaltime(min) 180 360 540 720
25
Chapter 2
50
1 3
st
th
2 4
nd
50
st
nd
3
40
th
th
40
th
Conversion (%)
30
Conversion (%)
30
20
R-ester
20
R-ester
10
10
0 0 60 120 180
0 0 60 120 180
Time (min)
Time (min)
Figure2.10
Activity of Novozym 435 after 4 cycles of reusage in the trans esterificationofvinylacetatewith(R,S)1phenylethanolintolueneat70 C (left) by conventional heating and(right) bymicrowave heating;the conversions into (R)1phenylethyl acetate have been plotted against time.
Conclusionsonthebiocatalyzed(trans)esterifications Surprisingly, in contrast to all literature referring to rate enhancements of enzymatic reactions no significant beneficial effects have been observed for each (trans)esterification carried out with microwave heating in our laboratory. All reactions show under consistent conditions a close or identical timeconversion history for conventional and microwave heating. Even a switch to a mono or singlemode microwave oven did not change the timeconversion history as compared to multimode heating conditions. Possible explanations for the observedeffectsinliteraturearerelatedtotheabsenceofpropercomparisonorto incorrect temperature measurements. For example, higher heating rates can be achievedbymicrowaveheatingresultinginanapparentlyhigherconversionrates. Properstirringandaccurateinternaltemperaturemeasurementsformthebasisof reliably comparing different heating methods. This statement is elaborated in chapter4. Thepurportofthisconclusionislikepeddlingupstreamonafaststreamingriver ofallreportedpositiveclaimswithmicrowaveheatinginthefieldofbiocatalysis.
26
Chapter 2
anilinederivates in general has been reported to be enhanced by microwave heating, see the papers of the groups of Varma14 and Collins.15 The workup procedureisassimpleastheseparationoftwoliquidphasesasdemonstratedby Varma.14 Both starting materials (aniline and 1,5dibromopentane) are miscible liquidsintheorganiclayer.Allreactantsabsorbmicrowaveenergyduetosimilar loss tangents, see Table 2.5. The aqueous layer contains a salt which strongly improvestheabsorptionofmicrowaves. Table2.5 Losstangentofthecomponentsinpiperidineformation.
watera aniline
b
1,5dibromopentane
(dielectricloss) 7.5(298K) 0.56 1.5/2.11c d (dielectricconstant) 78(298K) 7.1(293K) 9.1(303K)d losstangent(/) 0.100.14 ~0.1 ~0.2 a:[reference16];b:dielectriclossof2,4dimethylanilineat3GHz,293K[reference17];c:dielectric lossof1,5dibromopentaneat3.87GHz,293Kanddielectriclossof1,4dibromopentaneinbenzene at10GHz,293K,respectively[reference18];d:[reference19]
If selective heating occurs under the studied conditions, it is unlikely that the reactivitybetweenthestartingmaterialsisthermallyimproved.Regardlessofthe methodofheating,propermixingisdemandedandhighmasstransferrateswill alsostimulateheattransport.Theheatexchangeisthenefficientandconsequently nolocalheatingeffectsofanysignificantinfluenceareexpected.
NH2
Br
Br
Scheme2.4 Cyclizationofanilinewith1,5dibromopentane. The reaction in a pure aqueous medium gave irreproducible results when test tubes undermagneticstirring wereused. Witha pitchedblade impellera proper aniline/1,5dibromopentane dispersion in water was obtained resulting in a smooth increase of the amount of product. However, separation of the aqueous and the organic phase was much more difficult for the mechanically stirred dispersionsthanforthemagneticallystirredmixtures.Theadditionofanorganic solvent led to better phase separation without drastic changes in the overall reaction time. The timeconversion histories of the starting materials, aniline and 1,5dibromopentane, depicted for conventional and microwave heating in Figure 2.11.
27
Chapter 2
100 80
100 80
Conversion (%)
Conversion (%)
Time (min)
Time (min)
Figure2.11
Besides aniline also cyclohexylamine and 3bromoaniline have been tested under reflux conditions. Cyclohexylamine and 3bromoaniline were heated with 1,5 dibromopentaneinanaqueoussolutiontocompareconventionalandmicrowave heating.Timeconversionhistorieswerenotsignificantlydifferentforconventional and microwave heating. Table 2.6 shows the product yields after 20 minutes reactiontimebasedonaliphaticoraromaticamineconsumption. Table2.6 Cyclization of cyclohexylamine and 4bromoaniline with 1,5dibromo pentane
cyclohexylamine 4bromoaniline * yieldat20min. yield(%)* conventionalheating microwaveheating 67 38 66 31
Cyclizationof3chloroanilinewithbis(2chloroethyl)amine.HCl The cyclization of 3chloroaniline with bis(2chloroethyl)amine.HCl to a piperazine is a key step in the synthesis of two types of antidepressant (Etoperidone20andNefazodone21),seescheme2.5.
NH2 .HCl N Cl Cl 1,2-dichlorobenzene N NH.HCl
+
Cl
Cl
+ 2 HCl
Scheme2.5
Cyclizationof3chloroanilinewithbis(2chloroethyl)amine.HCl.
28
Chapter 2
At first glance this cyclization looks similar to the cyclization of aniline with 1,5 dibromopentane. However, the mechanism is different through the anchimeric assistance of nitrogen giving a threemembered ring (aziridine) as intermediate species. Also the driving force of this reaction is different due to the thermal removalofhydrogenchloridegasunderrefluxconditionsatelevatedtemperatures making this a (heterogeneous) gasliquid system. Highly boiling, inert 1,2 dichlorobenzene was selected as solvent. The reactions are monitored by measuringtheamountofhydrogenchloride(bytitrationataconstantpH)evolved fromthereactionmixtureatvarioustimeintervals,seeFigure2.12.
100 100
80
182 C
Conversion (%)
80
188 C/ 400 W
182 C/ 170 W
Conversion (%)
60
60
40
40
20
20
Time (min)
Time (min)
Figure2.12
Cyclization of 3chloroaniline with bis(2chloroethyl)amine.HCl in 1,2 dichlorobenzene under reflux (left) by conventional heating and (right) bymicrowaveheating.
Theabsenceofamicrowaveeffectisascertainedbyafaircomparisonbetweenthe energyinputsofconventionalandmicrowaveheating.At400Wpowerinputthe reflux temperature was 188 C, 6 C higher than for 170 W power input. As a consequence the conversion rate is somewhat higher at 400 W than at 170 W power input, resulting from a higher solubility of the reactants and a higher reactiontemperature.Againforthiscasepropertemperaturecontrolisdecisivein findingthecauseofanyacceleration. Conclusionsofthepiperidineandpiperazineformation The presented reactions serve as examples for liquidliquid and gasliquid combinations.Althoughthesereactionscannotbestandardizedforthewholefield of microwavemediated organic synthesis, specific microwave effects were found tobeabsent.
170 W was the average power to maintain the temperature at 182 C. 400 W was the constant powerresultinginatemperatureof188C.
29
Chapter 2
Under these reliable conditions the reactions are not accelerated e.g. by selective heatingofthemicrowaveenergy.Comparisonofthepiperazineformationunder bothheatingconditionsisnotprecedentedinliterature. AnearlierreportbyVarma18describesaseriesofefficientandsimplemicrowave assisted syntheses of substituted piperidines in heterogeneous liquidliquid systems.Onlyoneexample,thereactionbetweenethyl4aminobenzoateand1,4 dibromobutane, was investigated under conventional and microwave heating conditions to demonstrate a specific microwave effect. Since the microwave assisted experiment was performed in a CEM monomode device with a builtin infrared temperature sensor it is highly questionable whether the claimed microwaveeffectwillhold,duetotheabsenceofcorrecttemperatureregistration.
Scheme2.6 Oxidationofbenzylalcoholwithiron(III)nitratenonahydrate. The paper prompted us to gain more insight in the claimed microwave effect in this particular case. We investigated in detail whether microwave heating really outperforms conventional heating procedures in a batchwisely heated setup with accurate temperature control aimed at assessing the process scalability of this oxidationstep.
30
Chapter 2
Resultsofconventionalheating In contrast to the reactions in a flowreactor described by Jachuck et al32, the oxidationofbenzylalcoholwithFe(NO3)3.9H2Ohasbeenperformedunderaerobic conditions in an open vessel at various elevated temperatures reached by conventionalheating. Overviewofoperationalsequenceofreactions. Our study quickly revealed that the oxidation of benzyl alcohol with Fe(NO3)3.9H2Oincludesaseriesofredoxreactions.Thesereactionsmayoccurwith the oxidant and benzyl alcohol in the presence of water liberated from the nonahydrate.Thiswaterisreleasedupondissolutionoftheiron(III)saltinbenzyl alcohol. The rather complex reaction network may give rise to uncertainty in the exact stoichiometry. In general this oxidation is based on the hydrolysis of iron(III)nitrate,33seeScheme2.7.
2 Fe(NO3)3 + 3 H2O Fe2O3 + 6 HNO3
Scheme2.7
Theoxidationisbasedonthehydrolysisofiron(III)nitrate.
However, during the reaction also brown fumes were observed referring to the presence of NO2. This phenomenon was not observed when the reaction was performedinanargonornitrogenatmosphere,indicatingadirectoxidationofNO to NO2 under aerobic conditions. The absence or presence of oxygen has a profoundeffectonthestoichiometryofthereaction. Itwasexpectedthatiron(III)eitheractsasacatalystfortheoxidationunderaerobic conditionsorisconsumedbytheoxidation.Therefore,theamountofiron(III)was determined by a titration with EDTA before and after the reaction.34 After the reactionwascompletedtheamountofiron(III)inthereactionmixturewasfound tobeunchangedwithinexperimentalerror.However,insituautodecompositionof iron(II)nitrate prevents the identification of iron(II)asa result of a redox reaction (seeScheme2.8).35
18 Fe(NO3)2 + 15 H2O 4 Fe2O3 + 10 Fe(NO3)3 + 6 NO + 15 H2O
Scheme2.8 (Auto)decompositionofiron(II)nitratereproducesiron(III). The role of iron(III) in the oxidation of benzyl alcohol is purely catalytic by coordinating the oxidant to the alcohol. Even though nitrate and iron(III) are combinedinonesalt,amolarratioofbenzylalcoholandFe(NO3)3.9H2Oof100:1at
31
Chapter 2
50Cisinsufficienttobringthereactiontocompletion.Theconversioncompletely flattensoutafterafewhours,seeFigure2.13. Althoughoveralliron(III)remainsinitsoriginaloxidativestate,thecounterionin its original salt form has changed into the oxide thus lowering the iron(III) concentrationinsolutionbyprecipitationduringthereaction. Formationof(by)products. Throughoutthereactionatvarioustemperaturestheformationoftwobyproducts has been observed, namely benzyl nitrite36 and benzyl nitrate.37 The proposed sequence of reactions during the oxidation of benzyl alcohol to benzaldehyde is depictedinSchemes2.9and2.10. In Scheme 2.9 the postulated reactions lead to a 2:1 molar ratio of benzyl alcohol and Fe(NO3)3.9H2O. To rationalize the observation that a 4:1 molar ratio nearly leads to completion, the reactions as proposed in Scheme 2.10 should also be integrated.WaterandoxygenactasoxidantsinthisprocessbyreactivatingNO.
100,0 99,5
18
24
Time (h)
Figure2.13
Time molar composition plot of the oxidation of benzyl alcohol at a molarratioofbenzylalcoholandFe(NO3)3.9H2Oof100:1at50Cunder conventionalheatingandaerobicconditions.
32
Chapter 2
OH + 2 Fe(NO3)3
O H + 2 Fe(NO3)2 + 2 HNO3
OH 8 + 8 HNO3 OH 2 + 2 HONO 2 8
O H + 8 HONO + 8 H2O
6 HONO ONO 2
2 HNO3
+ 4 NO + 2 H2O OH + O H + 2 NO
OH 9 + 6 HNO3 OH 3 + 6 Fe(NO3)3 3 9
OH 12 + 6 Fe(NO3)3 12
O H + 6 NO + 12 H2O + 6 Fe(NO3)2
Scheme2.9
Reactionsrationalizinga2:1ratioinstoichiometryofbenzylalcoholand Fe(NO3)3.9H2O.Decompositionofiron(II)nitratereproducesiron(III).
12 NO + 6 O2 12 NO2 24 HONO 8 HNO3 + 16 NO + 8 H2O 8 HNO3
Scheme2.10
Reactionsrationalizingahigherthan2:1ratioinstoichiometryofbenzyl alcoholandFe(NO3)3.9H2O.
33
Chapter 2
75
2:1 4:1
50
25
16:1
0 0 4 8 12 16 20 24
Time (h)
Figure2.14
At a molar ratio of 16:1 (benzyl alcohol : Fe(NO3)3.9H2O) at room temperature benzyl alcohol was found to be saturated with Fe(NO3)3.9H2O affording a homogeneous solution. Larger amounts of Fe(NO3)3.9H2O resulted in two liquid phases.Inthosecasesthebottomlayerprimarilyconsistsofinorganicsalt,melting point47C,whilethetoplayerisrichinbenzylalcohol.Theheterogeneousnature of such reaction systems hampers a simple kinetic evaluation of the oxidation process.EachreactionmixturewithvaryingamountsofFe(NO3)3.9H2Oeventually forms a substantial amount of redbrown sediment on the reactor wall. An XPS analysis of the sediment points to a mixture of Fe(NOx)y. The formation of sediment is explained in Schemes 2.7 and 2.8. Nearly full conversion within one hour was achieved at 100 C when a 4:1 molar ratio of benzyl alcohol and Fe(NO3)3.9H2Owasappliedgivingaselectivityof75%tobenzaldehyde,seeFigure 2.15. Two important issues are noteworthy. Firstly, overoxidation leading to benzoic acid has been observed in all cases, which implicates the presence of an optimumfortheconversiontobenzaldehyde.Secondly,hightemperaturesshould beavoidedduetoinstabilityofthebyproductsduringthisoxidationstep.38
34
Chapter 2
Resultsofmicrowaveheating Inanextstep,theinfluenceofmicrowaveheatingonthecourseoftheoxidationof benzyl alcohol was assessed at temperatures of 50 C and 100 C. These two temperature levels were primarily used to gain insight into the potential of microwave heating (MW) compared to conventional heating (CH) as claimed by Jachucketal.27Themolarcompositionsofbenzaldehydeandbenzylalcoholinthe reactionmixtureareshowninFigure2.15.
100
100
80
60
40
Benzyl alcohol Benzaldehyde Benzoic acid Benzyl alcohol Benzaldehyde Benzoic acid
Molar composition(%)
80
Benzyl alcohol (MW) Benzaldehyde (MW) Benzyl alcohol (CH) Benzaldehyde (CH)
60
40
20
20
0 0 10 20 30 40 50 60
Time (min)
Time (min)
Figure2.15
(left)Timemolarcompositionplotoftheoxidationofbenzylalcoholata ratio of benzyl alcohol and oxidant of 4:1 at 100 C under comparable conventional (CH) and microwave (MW) heating; aerobic conditions. (right)Timemolarcompositionplotoftheoxidationofbenzylalcoholat aratioofbenzylalcoholandoxidantof16:1at50Cundercomparable conventional(CH)andmicrowave(MW)heating;aerobicconditions.
Experimentswithconventionalandmicrowaveheatingat50Cand100Ccanbe directly compared due to similar reaction vessels, stirring speeds and heating profiles. Surprisingly, significantly higher reaction rates with microwave heating were not observed at 100 C for the selected experimental conditions (4:1 molar ratioofbenzylalcoholandFe(NO3)3.9H2O).Subsequently,nobeneficialeffectwas observedeitherfora16:1molarratioofbenzylalcoholandFe(NO3)3.9H2Ounder equalinitialreactionconditionsat50C(Figure2.15). Conclusionsoftheoxidationofbenzylalcohol In conclusion, by direct comparison (i.e. open vessels, same stirrer speeds, same heating rates and internal temperature measurements) of the oxidation of benzyl alcohol with iron(III)nitrate nonahydrate under atmospheric conditions for homogeneous (initially) and heterogeneous reaction mixtures did not reveal any positivemicrowaveeffect.
35
Chapter 2
Based on the results obtained for both conventional and microwave heating it is concluded that a conventional batch heating procedure should be selected as the method of choice for a scalable process. This conclusion is supported by the observation that any microwaveinduced reactionrate enhancement is lacking whentheheatingconditionsaresystematicallycompared.Neitherprocessingina continuous mode nor microprocessing are preferable due to the necessity of aerobic conditions for completion of the reaction with minimal economical amounts of oxidant, the formation of a sediment during the reaction and the evolution of nitric fumes. The tremendously high heating rate as a result of a strong absorption of microwave energy by the reaction mixture leads to the conclusionthatmicrowaveheatingisnotthepreferredmethodforsafeoperations onlargerscales.Thesehighheatingratesleadtoarapidincrementoftemperature (and pressure) when the reaction is performed in a closed system with indirect temperature control as was probably the case in the described experiments.26 In generalaccuratetemperaturecontrolisaprerequisiteforassignmentofmicrowave effectsbasedondifferentrateconstants.
Scheme2.11 Williamsonsynthesiswiththemethylesterofgallicacid. Mainly due to the heterogeneous character caused by an excess of potassium carbonate in this chemical reaction, the synthesis of methyl 3,4,5tri dodecyloxybenzoate was selected for investigation with microwave heating in comparison to conventional heating. The excess of base, present in six molar equivalents with respect to the methyl ester of gallic acid, demonstrated the necessity of mechanical stirring. The optimal conditions as performed in earlier
36
Chapter 2
work39 (and slightly adapted conditions, which are not further elaborated) under microwaveorconventionalheatingresultedinsimilartimeconversionhistories (seeFigure2.16).
100
100
Conversion (%)
75
Conversion (%)
75
50
50
25
25
0 0 30 60 90 120 150
0 0 30 60 90 120 150
Time (min)
Time (min)
Figure2.16
ConclusionsoftheWilliamsonsynthesis With a large amount of a heterogeneous base in the Williamson synthesis, the overallreactionratehasnotbeenaffectedbymicrowaveheating.Clearlytherate determining step, the conversion of the three potassium phenolate intermediates, wasnotsignificantlyinfluenced.Duringthisreaction,sufficientmixingandgood temperaturecontrolwasneededtomakeafaircomparisonbetweenconventional andmicrowaveheating.
O N H OH
+
Scheme2.12
- H2O OH
N H
O O
Esterificationof(S)pyroglutamicacidwithndecanol.
37
Chapter 2
According to early reports this process was described as being catalyzed by p toluene sulfonic acid44 and sulfuric acid45 in an appropriate solvent (e.g. toluene). After optimization addition of a solvent and a catalyst were proven to be superfluous.44 Finally, the esterification is performed under neat conditions at elevated temperature (150 C) by microwave heating. In this way the product is easier to purify (no acidic catalyst contamination) and no solvent removal is needed. The final process design was a cooperation between Bioeurope, SAIREM and De Dietrichcompanies.TheindustrialrealizationwaspartiallyfinancedbytheFrench ADEME (Agency for Environment and Energy Control). With respect to the first reports involving the synthesis of Laurydone, the process was simplified. These simplificationsresultedinenergypluschemicalssavingsandtheprocesstimewas reducedbyafactoroffive.Theclaimofreductionoftheprocesstimewasbasedon comparing a conventionally heated process using toluene and ptoluenesulfonic acid at reflux temperature (110 C) with a microwaveheated process without solvent and catalyst at 150 C. Moreover, in the conventionally heated process a complicated workup procedure was necessary. According to these reports it remained unclear whether a conventionally heated experiment with all the optimized parameters would reveal a microwave effect for this reaction. See chapter5forapictureandadescriptionoftheflowsheet(Figure5.2).
100 MW CH
100 MW CH
Conversion (%)
75
50
Conversion (%)
0 30 60 90 120 150 180
75
50
25
25
Time (min)
Time (min)
Figure2.17
Esterification of pyroglutamic acid with ndecanol; initial experiments with(left)(R,S)pyroglutamicacid(7mmol)at145Cinatesttubewith magneticstirringand(right)(R,S)pyroglutamicacid(97mmol)at145 C in a 50 mL glass vessel with mechanical stirring (pitchedblade impeller). See also section 2.8 [experimental section] Figure 2.19 for picturesofexperimentalsetup.
38
Chapter 2
OurexperimentswereaimedatreproducingtheresultsreportedbyRochas42and atallowinganaccuratecomparisonbetweenconventionalandmicrowaveheating. Initialexperimentswereconductedwithracemicpyroglutamicacid(mp=182185 C, see Figure 2.17), but later a switch was made to (S)pyroglutamic acid (mp = 155160 C, see Figure2.18). In the original recipe also (S)pyroglutamicacid was reported. Melting points of both stereochemical forms (racemic and (S) pyroglutamicacid)differsostronglythattheheterogeneityintheinitialstagesof the reaction is significantly different. Also the solubility of the enantiomerically purepyroglutamicacidprovedtobehigherinndecanol.Bothfactorsgaveriseto ahigherreactionratefor(S)pyroglutamicacidthanforracemicpyroglutamicacid.
100
100
Conversion (%)
75
145C, (S)
Conversion (%)
75
145C, (R,S)
50
(MW) (CH)
50
25
25
Time (min)
60
120
Time (min)
180
240
300
360
Figure2.18
Esterificationof pyroglutamic acid withndecanol;(left)(S) and (R,S) pyroglutamic at 145 C in a glass vessel with overhead stirring by conventionalheatingand(right)(S)pyroglutamicacidat150Cwithan argon flow over the reaction mixture by microwave and conventional heating.
ConclusionsoftheLaurydoneprocess TheFrenchGermancollaborationgaveanintriguingandinterestingexampleofan industrial application of microwave heating on large scale. The handicap of the limited penetration depth of microwaves was overcome by the application of a loopsystem.AscanbeconcludedfromFigure2.17themicrowaveeffectofafactor 3 (also reported by the French authors) in reaction rate difference between conventional and microwave heating conditions vanished when a more efficient stirring method was selected. This surprising result can be attributed to an improvement in masstransport and more importantly to approach a more homogeneousdistributionoftemperatureinthereactionmixtureforbothheating techniques, due to ideal mixing (see Figure 2.17). A precedent in the literature showedsimilarresults.46
39
Chapter 2
TheresultscollectedinFigures2.17and2.18demonstratethatacomparableresult could have been achieved when the esterification would have been performed under conventional heating conditions at 150 C. Although the admirable efforts demonstrate as such that microwave heating in process scalingup is feasible, in our opinion the added value of microwave heating in comparison with conventionalheatingislacking. AlsotheLaurydonesynthesisnicelydemonstratesthatsmallchangesinsolubility mayhaveadirecteffectonthereactionrate.
40
Chapter 2
Figure2.19
A50mLdoublewalledreactorpositionedin(left)themicrowavecavity with mechanical stirrer (pitchedblade impeller) and fiber optic temperature sensor and (right) a similar reaction setup under conventionallyheatedconditions.
41
Chapter 2
Transesterifications(Microwaveoven;averagepower:~17W/max:300W): Ethyl caprylate with (R,S)1phenylethanol [Table 2.1; entries 1,2]; A 25 mL pear shapedflaskwaschargedwithNovozym435(104mg),asolutionofdiethylether(2mL) with(R,S)1phenylethanol(244mg,2mmol)andethylcaprylate(1035mg,6mmol).The (magneticallystirred)reactionmixturewasheatedtill90Cforaperiodof10min.After cooling the mixture was diluted with diethyl ether (10 mL) and filtered over cotton, and washed with diethyl ether. The filtrate was analyzed by GC. During reaction it was impossibletotakealiquotsbecauseofabsorbedliquidbytheexcessofcatalyst. Vinylacetatewith(R,S)1phenylethanol[Table2.1;entry3];A25mLpearshaped flaskwaschargedwithasolutionoftoluene(7mL)with(R,S)1phenylethanol(480mg,4 mmol)andvinylacetate(1722mg,20mmol).The(magneticallystirred)reactionmixture washeatedtill70C.Subsequently,Novozym435(50mg)wasaddedandthetemperature wasmaintainedforaperiodof60min.Aftercoolingthemixturewasdilutedwithdiethyl ether and filtered over cotton, and the residue washed with diethyl ether. The combined filtrateswereanalyzedbyGC.Duringreactionaliquotsweretakenfromtheliquidphase andanalyzedbyGC. 2,2,2Trifluoroethylbutyratewith(R,S)1phenylethanol[Table2.1;entry4];A25 mL pearshaped flask was charged with a solution of (R,S)1phenylethanol (480 mg, 4 mmol) and 2,2,2trifluoroethyl butyrate (1722 mg, 20 mmol) in toluene (7 mL). The (magneticallystirred)reactionmixturewasheatedtill70C.Subsequently,Novozym435 (50 mg) was added and the temperature was maintained for a period of 10 min. After coolingthemixturewasdilutedwithdiethyletherandfilteredovercotton,andtheresidue washed withdiethyl ether. The combined filtrates were analyzed by GC. During reaction aliquotsweretakenfromtheliquidphaseandanalyzedbyGC. Isopropenylacetatewith(R,S)1phenylethanol[Table2.1;entry5];A25mLpear shapedflaskwaschargedwithasolutionof(R,S)1phenylethanol(244mg,2mmol)and isopropenyl acetate (1001 mg, 10 mmol) in toluene (5 mL). The (magnetically stirred) reactionmixturewasheatedtill70C.Subsequently,Novozym435(100mg)wasadded andthetemperaturewasmaintainedforaperiodof60min.Aftercoolingthemixturewas diluted with diethyl ether and filtered over cotton, and the residue washed with diethyl ether. The combined filtrates were analyzed by GC. During reaction aliquots were taken fromtheliquidphaseandanalyzedbyGC.
42
Chapter 2
Isopropenyl acetate with (R,S)1,2,3,4tetrahydro1naphthol [Table 2.1; entry 6]; A 25 mL pearshaped flask was charged with a solution of (R,S)1,2,3,4tetrahydro1 naphthol(297mg,2mmol)andisopropenylacetate(996mg,10mmol)intoluene(5mL. Thereaction(magneticallystirred)mixturewasheatedtill70C.Subsequently,Novozym 435(100mg)wasaddedandthetemperaturewasmaintainedforaperiodof60min.After coolingthemixturewasdilutedwithdiethyletherandfilteredovercotton,andtheresidue washed withdiethyl ether. The combined filtrates were analyzed by GC. During reaction aliquotsweretakenfromtheliquidphaseandanalyzedbyGC. 2,2,2Trifluoroethyl butyrate with (R,S)1phenylethanol [Table 2.1; entry 7]; A 25mLpearshapedflaskwaschargedwithasolutionof(R,S)1phenylethanol(305g,2.5 mmol) and 2,2,2trifluoroethyl butyrate (851 mg, 5.0 mmol) in THF (6.6 g). The (magneticallystirred)reactionmixturewasheatedtill40C.Subsequently,Subtilisin(50 mg)wasaddedandthetemperaturewasmaintainedforaperiodover40h.Aftercooling themixturewasdilutedwithdiethyletherandfilteredovercotton,andtheresiduewashed withdiethylether.ThecombinedfiltrateswereanalyzedbyGC.Duringreactionaliquots weretakenfromtheliquidphaseandanalyzedbyGC. Vinylacetatewith(R,S)1phenylethanol[Table2.1;entry8];A25mLpearshaped flaskwaschargedwithasolutionof(R,S)1phenylethanol(305mg,2.5mmol)andvinyl acetate(430mg,5mmol)inTHF(6.6g).The(magneticallystirred)reactionmixturewas heated till 40 C. Subsequently, Subtilisin (25 mg) was added and the temperature was maintainedforaperiodover40h.Aftercoolingthemixturewasdilutedwithdiethylether andfilteredovercotton,andtheresiduewashedwithdiethylether.Thecombinedfiltrates were analyzed by GC. During reaction aliquots were taken from the liquid phase and analyzedbyGC. Vinylacetatewith(R,S)1phenylethanol[Table2.1;entry9];A25mLpearshaped flaskwaschargedwithasolutionof(R,S)1phenylethanol(415mg,3.4mmol)andvinyl acetate(1.46g,16.9mmol)intoluene(15mL).The(magneticallystirred)reactionmixture was heated till 40 C. Subsequently, PPL (1.0 g) was added and the temperature was maintained for a period over 6 h for microwave heating and over 40 h for conventional heating.Aftercoolingthemixturewasdilutedwithdiethyletherandfilteredovercotton, and the residue washed with diethyl ether. The combined filtrates were analyzed by GC. DuringreactionaliquotsweretakenfromtheliquidphaseandanalyzedbyGC.
43
Chapter 2
Vinylacetatewith(R,S)1,2,3,4tetrahydro1naphthol[Table2.1;entry10];A25 mL pearshaped flask was charged with a solution of benzene (3 mL) with (R,S)1,2,3,4 tetrahydro1naphthol (102 mg, 0.7 mmol) and vinyl acetate (293 g, 3.5 mmol). The (magneticallystirred)reactionmixturewasheatedtill40C.Subsequently,PPL(608mg) wasaddedandthetemperaturewasmaintainedforaperiodover6hformicrowaveheating andover20hforconventionalheating.Aftercoolingthemixturewasdilutedwithdiethyl ether and filtered over cotton, and the residue washed with diethyl ether. The combined filtrates were analyzed by GC. The filtrate was analyzed during reaction aliquots were takenfromtheliquidphaseandanalyzedbyGC. Esterifications: Adipic acid with nbutanol or noctanol [Table 2.2; entries 1,2]; A 25 mL pear shaped flask was charged with adipic acid (1.46 g, 10 mmol) and 1 equivalent of alcohol (0.74gofnbutanolor1.30gnoctanol,10mmol)andthetotalvolumewasadjustedwith 1,4dioxanetill20mL.The(magneticallystirred)reactionmixturewasheatedtill60C. Subsequently,Novozym435(70mg)wasaddedandthetemperaturewasmaintainedfora periodover3h.Aftercoolingthemixturewasdilutedwithdiethyletherandfilteredover cotton,andtheresiduewashedwithdiethylether.Thecombinedfiltrateswereanalyzedby GC.DuringreactionaliquotsweretakenfromtheliquidphaseandanalyzedbyGC. Piperidine/piperazineformation: Cyclization of aniline with 1,5dibromopentane [Scheme 2.4]; A 50 mL reaction vesselequippedwithanoverheadstirrer(265rpm)waschargedwith1,5dibromopentane (1281mg,5mmol),aniline(463mg,5mmol)andasolutionofK2CO3(760mg,5mmol) inH2O/1,4dioxane(10mL/10mL).Thereactionmixturewasheatedtill90C,andthe temperature was maintained for a period over 120 min. After cooling the mixture was allowed to separate phases with addition of ethyl acetate (10 mL). Organic phase (upper) was used for taking aliquots. The conversion of starting materials was analyzed by GC. (Microwaveoven;averagepower:30W/max:100W) Cyclizationof3chloroanilinewithbis(2chloroethyl)amine.HCl[Scheme2.5];A 50mLflaskwaschargedwith3chloroaniline(3.68g,28.8mmol)andasuspensionofbis (2chloroethyl)amine.HCl (5.14 g, 28.8 mmol) in 1,2dichlorobenzene (20 mL). The reactionmixturewasheatedtoreflux(182C)atasettemperatureofanoilbathat200C duringaperiodof120min(magneticallystirred).Theconversionwasmonitoredintime (intervals of 10 min.) by titrating the gaseous HCl after collection in a NaOHsolution withaconstantpHof12,withapHstat.
44
Chapter 2
Oxidationofbenzylalcoholtobenzaldehyde: Typicalprocedure:Benzylalcohol(1.0mL,9.7mmol)and1,3,5tritertbutylbenzene(10 mg, 0.041 mmol) as internal standard were introduced in a 10 mL twoneck round bottomedflask,andheatedinanoilbathormicrowaveoven.Whenthemixturereachedthe selected temperature, Fe(NO3)3.9H2O (0.24 g, 0.6 mmol) was introduced, and the (magneticallystirred)reactionwasstarted.Whenthedesiredreactiontimewasreached,the reactionmixture was elutedover a short silica columnwith a 5 mfilter attachedtothe end,withanexcessofdiethylether.ThediethyletherphasewasdriedwithMgSO4,filtered, and evaporated to yield the product(s). 1HNMR (CDCl3, 200 MHz) typical signals (ppm)4.70(s,2H,benzylalcohol),5.43(s,2H,benzylnitrate),5.71(s,2H,benzylnitrite), 10.03(s,1H,benzaldehyde). Iron(II)determination;Quantativelyasuitableamountofspecimenwastransferredinto 5mlof4MHCl.Thecontentwasheatedtoapproximately60C.Thefilterwaswashed withampledistilledwater.Thevolumetricflaskwasfilledtillitsfinalvolume.Asuitable amountwaspipettedintoabeakerandtheindicator1,10phenanthroline(seepreparationof 1,10phenanthroline solution) was added. A Ce(IV)solution (see preparation of Ce(IV) solution)wastitratedtillcolorchangefromorangetogrey. Preparation of 1,10phenanthroline solution; Analytical grade 1,10phenanthroline (49.60mg)wasaddedtodistilledwateranddilutedto250mL. PreparationofCe(IV)solution(0,1M);Ammoniumcerium(IV)sulphate(6466g)was added into a solution prepared by adding concentrated sulphuric acid (28 mL) to water (500 mL): the mixture was stirred until the solid had dissolved. The content was transferredintoa1Lgraduatedflask,andmadeuptothemarkwithdistilledwater. Iron(III)determination;Avolumeof25mLiron(III)solution(0.1M)waspipettedintoa conical flask and diluted to 100 mL with deionised water. The pH was adjusted to 23; Congoredpaperwasusedtothefirstperceptiblecolorchange.Anumberof5dropsofthe indicatorsolutionwasadded,thecontentsoftheflaskwaswarmedto40C,andtitrated with standard (0.1 M) EDTA solution until the initial blue color of the solution turned greyjustbeforetheendpoint,andwiththefinaldropofreagentchangedtoyellow. Preparation of EDTAsolution (0,1 M); Dry disodium dihydrogenethylenediamine tetraacetate(37.224g)wasdissolvedindistilledwateranddilutedto1000mL. Preparationofvariaminebluesolution(0,1M);Aquantityvariamineblue(1g)was dissolvedindeionisedwater(100mL).
45
Chapter 2
Williamsonsynthesis: Synthesisofmethyl3,4,5tridodecyloxybenzoate:A50mLglassreactorvesselwas charged with methyl gallate (0.74 g, 4 mmol), K2CO3 (3.32 g, 24 mmol), tetrabutylammonium bromide (TBAB, 65 mg, 0.2 mmol), MIBK (10 mL) and 1 bromododecane(3.09g,13mmol).Subsequently,thereactionmixturewasheatedtoreflux (120C)andstirredwithapitchbladeimpellerat800rpmfor2h.Theconversionofthe reaction was monitored by 1HNMR. Upon completion the brown mixture was cooled to below 100 C, and water (10 mL) was added. The aqueous layer was separated, and the organiclayerwaswashedwithwater(10mL),dilutedHClsolution(10mL1.0M),and water(10mL)again. 1HNMR(acetoned6,300MHz):7.27(s,2H,orthoH),4.113.98 (m, 6H, OCH2), 3.85 (s, 3H, OCH3), 1.901.70 (m, 6H, OCH2CH2), 1.551.20 (m, 4H, (CH2)8),0.86(t,9H,CH3).(Microwaveoven;averagepower:50W/max:300W) Laurydoneprocess: Synthesis of Laurydone: A 50 mL glass reactor vessel was charged with (R,S) or (S) pyroglutamic acid (12.5 g, 97 mmol,) and ndecanol (13.75 g, 87 mmol). The reaction mixturewasheatedto150Cwithin10minandstirredwithapitchbladeimpellerat550 rpmfor3h.Initiallythereactionmixturewasheterogeneous,butwithinafewminutesthe mixturewashomogeneous.Duringreactionaliquotsweretakenandanalyzedby1HNMR (CD3OD,300MHz)typicalsignals(ppm)4.14(t,2H,CH2OCO,product),3.53(t,2H, CH2OH,decanol).(Microwaveoven;averagepower:25W/max:400W)
46
Chapter 2
11 12 13 14 15 16 17 18 19 20
CleaTechnologiesB.V.,Delft,TheNetherlands. Yadav,G.D.;Lathi,P.S.J.Mol.Catal.B:Enzymatic2005,32,107. Rjasse,B.;Lamare,S.;Legoy,M.D.;Besson,T.Org.BiomolChem.2004,2,1086. Ju,Y.;Varma,R.S.Org.Lett.2005,7,2409. Barnard,T.M.;Vanier,G.S.;Collins,M.J.Org.ProcessRes.Dev.2006,10,1233. Kaatze,U.J.Chem.Eng.Data,1989,34,371. Fong,F.K.;McTague,J.P.;Garg,S.K.;Smyth,C.P.J.Phys.Chem.1966,70,3567. a)Anderson,J.E.;Smyth,C.P.J.Phys.Chem.1973,77,230;b)Chandra,S.;Prakash, J.J.Phys.Chem.1971,75,2616. Lide,D.R.HandbookofChemistryandPhysics;89thEd.;20082009(onlineed.). Goldstein,I.;SaenzdeTajada,I.;Riblet,L.A.;Theuseofetoperidoneforpreparing a pharmaceutical composition for treating male impotence; Squibb Bristol Myers CO;PatentnumberEP0362847,1990. Madding, G. D.; 4(2phenoxyethyl)1,2,4triazolone process intermediates; Bristol MyersCO;PatentnumberUS4318731,1986. Ju,Y.;Varma,R.S.Org.Lett.2005,7,2409. Nishiguchi,T.;Bougauchi,M.J.Org.Chem.1989,54,3001. Cornlis,A.;Laszlo,P.Synthesis1980,849. Martin,S.E.;Surez,D.F.TetrahedronLett.2002,43,4475. Cornlis,A.;Laszlo,P.Synthesis1985,909. Heravi,M.M.;Ajami,D.;Mojtahedi,M.M.Chem.Commun.1999,833. Lou,JD.;Zhu,LH.;Ma,YC.;Li,L.Synth.Commun.2006,36,3061. Lou,JD.;Cai,XX.;Li,F.;Li,L.;Gao,CL.Synth.React.Inorg.,Met.Org.,NanoMet. Chem.2006,36,599. Varma,R.S.;Dahiya,R.TetrahedronLett.1997,38,2043. Jachuck,R.J.J.;Selvaraj,D.K.;Varma,R.S.GreenChem.2006,8,29. Experimental of reference 31: The reactant was prepared at room temperature by dissolving solid Fe(NO3)3.9H2O in benzyl alcohol. The mixture was stirred thoroughly and filtered using a Fisherbrand1 filter paper. Experiments were performed using the continuous isothermal reactor under the influence of microwave irradiation for a range of residence times (317 s) corresponding to differentflowrates(15mLmin1)anddifferentmicrowaveintensities(0Wto39 W).ThefeedwaspumpedthroughthereactorusinganHPLCpumpandtheheat transferfluid(water)wascirculatedthroughtheheattransferzoneofthereactorat 120mLmin1withaperistalticpump. Shang,Y.;vanWeert,G.Hydrometallurgy1993,33,273. Vogel, A. I.; Vogels textbook of quantitative chemical analysis, 5th Edition, Burnt Mill: LongmanScientific&Technical,1989,326.
21 22 23 24 25 26 27 28 29 30 31 32
33 34
47
Chapter 2
35 36 37 38 39 40 41 42 43 44 45
Shang,Y.;vanWeert,G.Hydrometallurgy1993,33,255. Cornlis,A.;Herz,P.Y.;Laszlo,P.TetrahedronLett.1982,23,5035. Baruah,A.;Kalita,B.;Baruam,N.C.Synlett2000,7,1064. Gray,P.;Williams,A.Chem.Rev.1959,59,239. Hersmis,M.C.;Spiering,A.J.H.;Waterval,R.J.M.;Meuldijk,J.;Vekemans,J.A.J. M.;Hulshof,L.A.J.Org.ProcessRes.Dev.2001,5,54. Custers,J.P.A.;Hersmis,M.C.;Meuldijk,J.;Vekemans,J.A.J.M.;Hulshof,L.A.J. Org.ProcessRes.Dev.2002,6,645. HowarthP.,Lockwood,M.TCE,2004,June,30. Rochas, J. F.; Bernard, J. P.; Lassalle, L. Ann. Microwave Heating Symposium, 2002, conf.37,2023. Jose, A.; Takeru, H.; Pyroglutamic acid esters used as dermal penetration enhancersfordrugs;Merck&CoInc;PatentnumberEP227531,1987. Thomae,K.GmbH.;NeuePyrrolidoncarbonsurealkylesterundVerfahrenzuihrer Herstellung.PatentnumberDE2102173,1972. Lassalle, L.; Procd de prparation dun ester par chauffage, au moyen dun champ de microondes dun mlange heterogne dacide et dalcool. Patent numberFR2833260,2003. Moseley, J. D.; Lenden, P.; Thomson, A. D.; Gilday J.P. Tetrahedron Lett. 2007, 48, 6084. Herrero,M.A.;Kremsner,J.M.;Kappe,C.O.J.Org.Chem.2008,73,36. Westaway, K.C.; Gedeye, R.N. J. Microwave Power and Electromagn. Energy, 1995, 30,219.
46 47 48
48
Chapter 3
Chapter 3
Abstract
Racemization as part of a dynamic process or classical resolution process can make the difference between 50 % and full conversion of the substrate. The racemization of N-acetylindoline-2carboxylic acid has been studied by conventional and microwave heating under various conditions. This study revealed beneficial microwave effects. The magnitude of this effect is governed by the degree of heterogeneity of the reaction system. The amount of catalyst, the temperature and amount of co-solvent played a decisive role. The microwave effect vanished when a homogeneous solution was obtained. Additionally, a significant rate enhancement was observed during microwave heating of the heterogeneous racemization mixture of Nacetyl-phenylalanine in p-xylene. For both examples the microwave effects could be rationalized by selective heating of the boundary layer i.e. the interphase of solid and liquid. Locally, the enhanced temperature increases the solubility of the substrates and consequently the rate of racemization.
50
Chapter 3
3.1 Introduction
Enzymes are complex chiral structures and their functionality tends to interact, therefore, differently with two enantiomers or mirror images of one molecule as already mentioned in section 2.3. Pharmaceutical intermediates often incorporate one or more stereocenters. To guarantee the right therapeutic treatment specific interactions between drug and target (e.g. receptor/enzyme) should take place. In generaltheenantiomericselectivityisvalidformanymoleculesthatinteractwith differentorgansinourbody,likethoseresponsiblefortasteorodor. Synthesisbecomesmorecomplexwhenmorestereocentersarepresentinthefinal product. Sometimes nature provides us with enantiomerically pure starting compounds,otherwisethesynthesisortheworkupshouldbeenantioselective.1 One way to incorporate an enantioselective reaction is to make use of kinetic resolution, e.g. via biocatalysis. Alternatively and more beneficially, dynamic kinetic resolution (DKR) may offer a solution by starting with a racemic mixture andendingwithanenantiomericallypureproductinhighyields,seeFigure3.1.
Sub (R) Sub(+) Sub(+) fast reaction Product (R) Product(+) Product(+)
slow reaction
Figure3.1
Overview of dynamic kinetic resolution (DKR). A combination of two processes: racemization of substrate (Sub) and kinetic resolution to productfromaracemicmixture(RorS).
DKR is based on two processes: an enantioselective reaction (kinetic part) and a racemization (dynamic part) of the starting material. With DKR the racemization becomesaninsituprocess. Theracemizationof1phenylethanolwiththeaidofarutheniumcatalysthasbeen intensively studied by Van Buijtenen in our group; see Scheme 3.1.2 The racemizationwasultimatelycombinedwithanenzymatickineticresolution(with thehelpofNovozym435).AdditionalresearchontheconventionallyheatedDKR processhasbeenreportedintheMScthesisofVanNispen.3Acomparisonwitha microwaveheatedexperimentwasalsoincluded,seeFigure3.2.
51
Chapter 3
OH
Scheme3.1
(R)
(R,S)
Conventionalheating
Microwave heating
150
200
Figure3.2
Unfortunately, in the racemization of (S)1phenylethanol no rate enhancement with microwave heating was observed. Note that working under O2free conditions for microwave heating was more difficult than for conventional heating.3 AnotheroptionintheDKRprocedureisseparatingthekineticresolutionstepfrom the racemization, when these two processes impede each other (e.g. inhibition or deactivation). On large scale with multiple batches, implementation of the individualprocessesiseconomicallyfeasibleduetotheabilityofrecyclingwaste streamsasshownintheexamplediscussedinthenextsection.
52
Chapter 3
thesolubility,thusimprovingtotherateoftheracemization.However,removalof the solvent (e.g. through distillation) is inefficient from the time and energy perspectives.FortheracemizationofNacetylindoline2carboxylicacid,pxyleneis selected as a solvent in view of its high boiling point, inertness and poor solubilizingpower.Indirectlythehighboilingpointpartiallycompensatesforthe poor solubility. The workup is simplified by filtration, which is more energy effective than distillation. The actual racemization is catalyzed by acetic anhydride,6,7seeScheme3.2(left).
O OH N O (R,S) 1 H Perindopril O
O NH
OH N O (R) 1 O
H N
O OH O
Scheme3.2
(left)Racemizationof(R)Nacetylindoline2carboxylicacid(1). (right)Perindopril,drugforhypertensiontreatment.
Acetic anhydride reacts with the acid functionality thus forming a mixed anhydride. Ring closure eliminates acetate and gives rise to an oxazolium intermediate. At this stage the chiral information is lost by ketoenol tautomerizationandduetothereversiblecharacterofthisreactionthesubstrate racemizes completely; see Scheme 3.3 and Figure 3.3. Other options, such as Pd catalysistoracemizeaminoacids812andracemizationsingeneralarereviewedby Zwanenburgetal.13 In our laboratory the racemization experiments have been performed with two grades with respect to enantiomeric purity and to stereochemical identity. The initialexperimentswereperformedusingascalemicmixturewithanenantiomeric excess (ee) of 67 % (R). Additional experimentation was performed with the (S) enantiomer (99 % ee). This study surprisingly revealed a complex process that requiredmoredetailedinvestigation,seeFigure3.4fortheresults.Onadditionof thecatalysttheeeinitiallydecreasesasexpected,butafteracertainperiodtheee levels off with conventional heating. The rate of racemization dependents on temperatureaswellasontheamountofaceticanhydride(seealsosection3.3fora mechanisticsuggestionandScheme3.6).
53
Chapter 3
H N Chiral
+
O + O [3] O
OH O
O N
+
[4] N Achiral
+
OH + O O
Scheme3.3
Mechanismoftheracemizationof(R)Nacetylindoline2carboxylicacid: [1] formation of mixed anhydride, [2] ring closure, [3] ketoenol tautomerizationand[4]acidbaseprotonshift. Remark: after tautomerization the oxazolium is achiral; the reverse reaction can form either enantiomer (R or S). Only one enantiomer is depictedtomakethisschemecomprehensible.
70 60 50 40 30 20 10 0 0 20 40 60 80 100 120
ee%
Time (min)
ee%
Time (min)
Figure3.3
Racemization of (R)Nacetylindoline2carboxylic acid in pxylene (1.35 M) with 0.1 equivalent of acetic anhydride at 130 C with conventionalheating:(left)eevs.timeforalongperiodand(right)eevs. timefortheinitialperiod.
54
Chapter 3
100
100
75
ee%
25
ee%
50
75
50
25
Time (min)
Time (min)
Figure3.4
Water may influence the course of the reaction due to hydrolysis of acetic anhydride. After complete hydrolysis of acetic anhydride the racemization stops. For molar ratios of water and acetic anhydride smaller than one some catalytic activityremains.However,thereactionproductofhydrolysisofaceticanhydride, aceticacid, acts asa cosolvent for Nacetylindoline2carboxylicacid. Hereacetic acidpositivelyinfluencestherateofracemization. Ifthesolubilityofoneoftheintermediatesinthecatalyticprocessismuchlower than that of the actual substrate, the racemization may stop (i.e. deposition of an intermediatemakestheracemizationirreversible).Thenracemizationcanonlytake place with a stoichiometric amount of acetic anhydride. The actual amount of acetic anhydride necessary for proper racemization strongly depends on the wetnessofthesubstrate. Since saturation is maintained during the complete racemization process and the reactionratepartlydependsontheconcentrationofthesubstrateinsolution,the reactionrateshouldincreaseuntiltheenantiomericexcesshasreachedavalueof 2030%eerepresentingthesituationwiththehighestsolubility(Figure3.5).14For ee<30%thesolubilityseemsnottobesignificantlydependentoneeandpseudo zeroorder kinetics should be expected. The actual observations are different, becausethereactionratedependsonthepresenceofactivecatalyst.
55
Chapter 3
pure Solvent
16 14 12 10 8 6 0 Figure3.5
solubility g/kg
0.00
1.00
0.025
0.975
R-enantiomer
0.050 0.95
solvent
0.075
0.925
0.10
20 40 60 80 100
pure R
0.00
0.025
0.050
0.075
ee%
S -enantiomer
(left) Solubility (g [substrate] per kg [reaction mixture]) of N acetylindoline2carboxylic acid at different values of ee in pxylene at 130 C. (right) Ternary diagram of Nacetylindoline2carboxylic acid andpxyleneat130C.14
Influenceofthecosolvent The racemization of Nacetylindoline2carboxylic acid using pure acetic acid as solventhasbeenreported.4,5Inourcaseaceticacidhasbeenappliedasacosolvent with pxylene to study the role of heterogeneity. With approximately 20 wt% of acetic acid the reaction mixture becomes homogeneous. In this regime substrate concentrationisatamaximumandtheracemizationprocessproceedsmuchfaster (seeFigure3.6).
100
75
100
75
ee (%)
25
ee%
0 60 120 180
50
50
25
Time (min)
Time (min)
Figure3.6
(left)Racemizationof(R)Nacetylindoline2carboxylicacidinpxylene at 130 C with 0.1 molar equivalent of acetic anhydride with conventional heating; influence of the amount of cosolvent, acetic acid. (right) Racemization of (S)Nacetylindoline2carboxylic acid in p xylene at 130 C with 0.1 molar equivalent of acetic anhydride with conventionalheating;influenceoftheamountofcosolvent,aceticacid.
56
Chapter 3
Influenceofthetemperature To study the influence of temperature on the racemization, the reaction was also performed at temperatures below 130 C. This resulted in incomplete racemizations which were already depicted in Figure 3.4 (right). As mentioned before, acetic anhydride is consumed in an intermediate which seems to be irreversibly formed (e.g. by precipitation) at lower temperatures in the apolar solvent pxylene. Increasing the bulk temperature or improving solubility in generalenablesthepossibilitytocompletetheracemization. In addition, the effect of the temperature on the racemization of Nacetyl phenylalanine, which is analogous to Nacetylindoline2carboxylic acid, is also illustratedinthenextsection.Thisreactionappearstohaveabroadertemperature rangeinwhichracemizationoccurs.Nevertheless,herealsoinhibitionisobserved inthecaseoftemperaturesbelow70C. Influenceofthesubstrates Tounderstandthepeculiarmechanisticphenomenaobservedintheracemization of Nacetylindoline2carboxylic acid and to gain more insight into the related microwave effect (see section 3.3), other chiral Nacetyl amino acids were synthesized to investigate their racemization behavior. Two substrates in which formally the CH2Ph and NHPh bands were cleaved have been selected, i.e. N acetylphenylalanine(2)andNacetylNphenylalanine(3),seeScheme3.4.
O H2N OH Ac2O base O HN O O N H OH Ac2O base N O OH 2 O OH
O H2N OH
Scheme3.4
NAcetylphenylalanine and NacetylNphenylalanine could in principle be synthesized from optically pure natural amino acids as starting materials. Acetylationof(S)phenylalanineisastraightforwardproceduretoobtainthefirst
57
Chapter 3
substrate 2. Unfortunately, it was not possible to isolate (S)NacetylNphenyl alanine(3)insufficientamountsbyanUllmanntypeofcoupling. Under benchmark conditions (i.e. 130 C) the racemization of (S)Nacetyl phenylalanine (2) was complete within 5 minutes. For practical reasons the temperaturewasloweredtoavalueof85C(Figure3.7). Theheatofsolution(solnH)ofNacetylphenylalaninefortheracemicmixtureand fortheenantiomericallypurecompoundwerebothcalculated(seeequation3.1,Sol stands for solubility; [mol.L1]) as approximately 85 kJ/mol for both cases in p xylene.15 Thelarge value of solnH indicates that the influence of temperatureon the rate of racemization is considerably governed by the solubility of N acetylphenylalanine.16 In comparison, the heat of dissolution (solnH) for N acetylindoline2carboxylic acid was calculated to be 75 kJ/mol in pxylene representingalsoahighvalue.15
ln Sol = constant
soln H RT
(3.1)
Racemizationappearstobefastandhencemeasurementsintheliquidphaseshow onlythepresenceofaracemicmixtureforthetwosubstrates.
OH HN O O Ac2O p-xylene, 85 C HN O OH O
Scheme3.5
Racemizationof(S)Nacetylphenylalanine(2)withtheaidof0.1molar equivalentofaceticanhydride.
100
75
70 C 85 C 100 C 130 C
ee%
50
25
Time (min)
Figure3.7
58
Chapter 3
rCH t =0
The results in Figure 3.4 (left) demonstrate that the racemization seems to stop when using smaller amounts of acetic anhydride. Duplicates illustrate the same behavior. However to our surprise when the same reaction mixture was transferred from an oil bath to the microwave oven racemization starts up again (Figure3.8).
100
90
ee%
80
Time (min)
Figure3.8
Racemization of (S)Nacetylindoline2carboxylic acid with 0.1 molar equivalent of acetic anhydride at 130 C with conventional heating; duplicate( )andanextradriedbatch( ).Afterthreehoursthe reactionvesselwastransferredtocontinuethereactioninthemicrowave oven( ).
Initial period is after a possible induction period and is measured over at least a conversion differenceof10%.
59
Chapter 3
This observation suggests that the molecular species that stops the racemization under conventional heating is actually an intermediate in the process. The previously mentioned intermediate in section 3.2, which has temporarily consumed all of the acetic anhydride, is (partly) dissolved or reactivated by microwave heating. Given the relationship between enantiomeric excess at the plateau and the amount of acetic anhydride (Figure 3.4 (left)) the intermediate presumably originates from a (potentially racemized) substrate and two acetic anhydrides.Oneanhydrideisusedtoformtheoxazoliumintermediateinthefirst part and another anhydride is used to form an eventually insoluble or inactive structurewhichinhibitstheoverallprocessofracemization(Scheme3.6).Onlyan increaseintemperatureoralargeramountofcosolventcanshifttheequilibrium backtocompletetheracemization.Thesolubilitydoesnotplayaroleatlower%ee duetoahighersolubilityoftheracemate(seeFigure3.5).
O O O H N O O Chiral [1] N
+
O O
O O
[2] N Achiral
+
OH O
[6]
[3]
O O O
[7] N O O
O O O
Scheme3.6
Inhibitionofracemization:[1]formationofoxazoliumintermediatewith acetic anhydride, [2] ketoenol tautomerization, [3] formation of oxazoliumacetatewithaceticanhydride,[4]precipitationorinactivation ofintermediate,[5]dissolutionofintermediatebye.g.microwaveheating and [6] addition of water leading to regeneration of the substrate, [7] acetatebondedtointermediate(covalentorionic). Remark: after tautomerization the oxazolium is achiral; the reverse reaction can give either enantiomer (R or S). Only one enantiomer is depictedtomakethisschemecomprehensible.
60
Chapter 3
Shortly after addition of acetic anhydride the solubility of Nacetylindoline2 carboxylic acid temporarily increases. Thereafter, two processes of precipitation andsimultaneousracemizationallowthe%eetoslightlyincreasewithrespectto the plateau level. The (S)enantiomer selectively precipitates in favor of the (R) enantiomer, due to the presence of a majority of (S)enantiomer crystals. After workup with water the insoluble structure is dissociated into (R) or (S)N acetylindoline2carboxylicacid.ByapplyingsolidstateNMRtheintermediatehas beencomparedwiththeenantiomericallypuresubstrateandtheracemicproduct. The spectrum shows another set of peaks, but no conclusive evidence can be producedfromthisanalysis. The increase of % ee in time could, however, not be explained by the method of sampling,whichwascheckedbycomparingrepresentativesampleswiththedirect workup of a whole batch. An increase of % ee in a dynamic system during a longer period of days has also been demonstrated by Blackmond et al.17 The increaseineeisexplainedbyaslightenantioimbalancewhichdirectsthetrendtoa singlechiralsolidstate. Influenceofacosolvent The addition of acetic acid as a cosolvent diminishes the heterogeneity of the reaction mixture. During conventional heating this induced a drastic increase of thereactionrate(seeFigure3.6).Withmicrowaveheatingreactionrateswerenot significantlydifferentforallsolventcompositions,seeFigure3.9.
70 60 50 40
CH
ee %
20 10 0 0 60 120 180
f MW
30
0 0 5 10 15 20
Time (min)
Co-solvent (%)
Figure3.9
Racemization of (R)Nacetylindoline2carboxylic acid with 0.1 molar equivalentofaceticanhydrideinvariouspxylene/aceticacidmixturesat 130 C for (left) conventional (CH) and microwave heating (MW) conditionsand(right)influenceofsolventcompositiononthemicrowave effect(fMW).
61
Chapter 3
As soon as the reaction mixture becomes homogeneous (20 wt% cosolvent) no differences were observed between conventional and microwave heating. The microwaveeffectisstrongintheabsenceofaceticacidascosolvent(fMW=8),while it strongly reduces upon addition of acetic acid. Hence the effect vanishes when homogeneityisreached(fMW=1),seeFigure3.9(right). Influenceofthesubstrates (S)Nacetylphenylalanine (Scheme 3.5) also showed an enhancement of the reaction rate at 70 and 85 C with microwave heating, see Figure 3.10 (left). The appearance of a microwave effect in the racemization of Nacetyl amino acids appears not to be a onecaseshow. At 85 C in pure pxylene racemization is observed for both heating techniques while observing a strong microwave effect (fMW = 5.8). However, at 70 C the racemization only continues for microwave heating.Atthistemperature(70C)themicrowaveeffectistheoreticallyveryhigh, seeFigure3.10(right).
100
CH
MW 70 C 85 C
6
>>
75
Conversion (%)
50
25
f MW
0 60 120 180 240
0 65 70 75 80 85 90 Temperature (C)
Time (min)
Figure3.10
62
Chapter 3
irradiation. However, for heterogeneous reactions differences were observed between both heating techniques. During the racemization of Nacetylindoline2 carboxylic acid and Nacetylphenylalanine real microwave effects have indeed been observed under various conditions. Changing the reaction conditions (e.g. temperature or amount of cosolvent) had, however, a direct influence on the magnitudeofthemicrowaveeffects.Itseemsthatthereisacorrelationbetweenthe heterogeneityofthesystemandtheobservedmicrowaveeffects.Theracemization of Nacetyl amino acids in pxylene occurs at or near the surface of the solid particles. In solution only a racemic mixture could be measured throughout the racemization. Neitherthe(pure)solvent(pxylene;losstangent0.0011)18northesolid(Nacetyl amino acid) absorb the microwaves efficiently.18 Although the existence of hot spotsisunlikelytooccurinthissituation,selectiveheatingattheinterfacebetween solid and liquid would be more plausible (see also chapter 4). Mechanistically nothing changes. However, solubilities may increase as a result of selective heating.Smallchangesintemperaturealsohavealargeinfluenceonthesolubility of the Nacetyl amino acid, according to a high heat of solution (solnH = 7585 kJ/mol).Forexample,anincreaseof2Cenhancesthesolubilitywith15%andan increaseof10CdoublesthesolubilityofNacetylphenylalanineinpxylene.
63
Chapter 3
theanhydride.Thesamplesweredissolvedinamixtureofformicacid(3mL),isopropanol (25 mL) and nhexane (72 mL) and analyzed by HPLC, chiral column (Daicel, Chiracel OD)withaneluentcomposedofformicacid(1v%),isopropanol(10v%)andnhexane(89 v%).Theyieldafterworkupwas1.35g(76%).Thesameprocedurewasfollowedforthe conventionally heated experiment by substituting microwave irradiation by oil bath heating.Yieldafterworkup:1.62g(91%). Racemizationof(S)Nacetylindoline2carboxylicacid(MW)of99%ee.Around bottomed flask was charged with dry (S)Nacetylindoline2carboxylic acid (1.78 g, 8.7 mmol,99%eeS)andpxylene(5.5g).Themixturewasheatedto130CintheMilestone microwaveoven.(AWeflonbarwasaddedduringmicrowaveheatingat0wt%cosolvent). Thenaceticanhydride(100L,1.1mmol)wasadded.After2hwater(100L)wasadded, and pxylene/water/acetic acid was distilled from the reaction mixture. After cooling the solidwasfilteredandwashedtwicewithpxylene(2mL)anddriedunderreducedpressure at50C.Thereactionwasmonitoredbyquenchingasample(~50mg)withwater(25L) toneutralizetheanhydride.Thesamplesweredissolvedinamixtureofformicacid(3mL), isopropanol(25mL)andnhexane(72mL)andanalyzedbyHPLC,chiralcolumn(Daicel, ChiracelOD)withaneluentcomposedofformicacid(1v%),isopropanol(10v%)andn hexane (89 v%). The yield after workup was 1.35 g (76 %). The same procedure was followedfortheconventionallyheatedexperimentsubstitutingmicrowaveirradiationwith oilbathheating. NAcetyl(S)phenylalanine (2):10 A roundbottomed flask was charged with (S) phenylalanine(19.8g,0.12mol)andanaqueoussolution(30mL)ofNaOH(4,8g,0.12 mol).Tothismixtureaceticanhydride(12.24g,0.12mol)wasaddeddropwisewiththe simultaneous addition of an aqueous solution (15 mL) of NaOH (4,8 g, 0.12 mol) under vigorousstirringandatroomtemperature.Thetemperaturewaskeptbelow30CandpH was maintained around 10. After 30 min. another amount of acetic anhydride (12.24 g, 0.12mol)wasaddeddropwisewithsimultaneousadditionofanaqueoussolution(15mL) ofNaOH(4,8g,0.12mol)undervigorousstirringandatroomtemperature.Thismixture wasextractedwithethylacetate(100mL),pHwassetat6with6MHClundervigorous stirringwithcoolinginanicebath.TheextractionprocedurewasrepeatedforpHvalues setat5,4and3.Theorganiclayerswerecombinedandsubsequentlywashedwithbrine (100mL)anddemiwater(50mL).Organicsolventwasremovedunderreducedpressure. Theyieldafterworkupwas21.0g(84%). 1HNMR(CD3OD,400MHz)typicalsignals (ppm) 7.51 (m, 5H, ArH), 4.92 (m, 1H, CHCH2), 3.45 (dd, 1H, CHCH2), 3.20 (dd, 1H, CHCH2), 2.15 (s, 3H, COCH3); []23D +29.7 (c = 10, MeOH), literature19 []23D +33(c=10,MeOH).
64
Chapter 3
NPhenyl(S)alanine:20Aroundbottomedflaskwaschargedwith(S)alanine(28.51g, 0.32mol),CuI(6.09g,32mmol),K3PO4(135.6g,0.64mol),bromobenzene(39.25g,0.25 mol), diethanolamine (60 mL) and water (200 mL). The reaction mixture was heated at 90Cfor40hundernitrogenatmosphere.Thereafter,themixturewasquenchedon1kgice and was extracted with ethyl acetate (500 mL), pH was set at 6 with 6 M HCl under vigorousstirringwithcoolinginanicebath.TheextractionprocedurewasrepeatedforpH valuessetat5,4and3.Theorganiclayerswerecombinedandsubsequentlywashedwith brine(500mL)anddemiwater(500mL).Theorganicsolventwasremovedunderreduced pressure. The yield after workup was 26.2 g (63 %). 1HNMR (CD3OD, 400 MHz) typical signals (ppm) 7.37 + 6.90 (m, 5H, ArH), 4.30 (q, 1H, CHCH3), 1.73 (d, 3H, CHCH3).HPLCChiracelOD(forconditionsseefirstprocedure);50%ee.
NAcetylNphenylalanine(3):10,21AroundbottomedflaskwaschargedwithNphenyl alanine (23.28 g, 0.14 mol) and triethylamine (85.0 g, 0.84 mol). At room temperature aceticanhydride(14.3g,0.14mol)wasaddeddropwise.Thereactionmixturewaskeptat room temperature for 3 h. Thereafter, the mixture was quenched with water (10 g) and extracted with ethyl acetate (250 mL), pH was set at 5 with 6 M HCl under vigorous stirringwithcoolinginanicebath.TheextractionprocedurewasrepeatedforpHvalues setat4and3.Theorganiclayerswerecombinedandsubsequentlywashedwithbrine(100 mL)anddemiwater(100mL).Yield:5.7g(20%). 1HNMR(CD3OD,400MHz)typical signals(ppm)7.38(m,5H,ArH),4.93(q,1H,CHCH3),1.86(s,3H,COCH3),1.28(d, 3H,CHCH3).HPLCChiracelOD(forconditionsseefirstprocedure);12%ee.
65
Chapter 3
7 8 9 10 11 12 13 14
Chen,F.M.F.;Benoiton,N.L.Int.J.PeptideProteinRes.1988,31,396. Bergmann,M.;Zervas,L.Eur.J.Biochem.1928,03,280. Knorr,R.;Huisgen,R.Chem.Ber.1970,103,2598. Chenault,H.K.;Dahmer,J.;Whitesides,G.M.J.Am.Chem.Soc.1989,111,6354. Parvulescu,A.;VanEycken,E.;Jacobs,P.A.;DeVos,D.E.J.Catal.2008,255,206. Hateley,M.J.;Schichl,D.A.;Fischer,C.;Beller,M.Synlett,2001,1,25. Ebbers, E. J.; Ariaans, G. J. A.; Houbiers, J. P. M.; Bruggink, A.; Zwanenburg, B. Tetrahedron1997,53,9417. Solubility is determined by weight difference (before and after drying) of supernatantofNacetylaminoacidinpxylene(at130C)aftervigorousmixingof asaturatedsolutionduring5minand16h. Solubilityinpxyleneof(S)Nacetylphenylalanineis3.2,4.1and27mmol/lat85, 100and115C;of(R,S)Nacetylphenylalanineis2.3,5.4and21mmol/lat85,100 and 115 C and of (R,S)Nacetylindoline2carboxylic acid is 4.1, 8.9 and 24.4 mmol/lat102,116and132C,respectively. Landolt,H;Brnstein,R.ZahlenwerteundFunktionenausPhysik,Chemie,Astronomie, GeophysikundTechnik;IV/2a;Springer;19501980. Noorduin,W.L.;Izumi,T.;Milemaggi,A.;Leeman,M.;Meekes,H.;VanEnckvort, J.P.;Kellogg,R.M.;Kaptein,B.;Vlieg,E.;Blackmond,D.G.J.Am.Chem.Soc.2008, 130,1158. Magee,M.D.;Walker,S.Trans.FaradaySoc.1966,62,1748. Smith,H.E.;Neergaard,J.R.J.Am.Chem.Soc.1996,118,7694. a)Lu,Z.;Twieg,R.J.TetrahedronLetters2005,46,2997b)Ma,D.;Zhang,Y.;Yao,J.; Wu,S.;Tao,F.J.Am.Chem.Soc.1998,120,12459. a)Cardillo,G.;Gentilucci,L.;Tolomelli,A.;Tomasini,C.Tetrahedron1999,55,6239 b)Czernecki,S.;Ayadi,E.;Randriamandimby,D.Org.Chem.1994,59,8256.
15
16 17
18 19 20 21
66
Chapter 4
Chapter 4
Abstract
The combination of isocytosines and isocyanates leads to building blocks for supramolecular polymers. Nucleophilic additions of various C6substituted isocytosines (methyl, ethyl, isopropyl and phenyl) to (di)isocyanates have been investigated. The heterogeneous reaction mixtures again showed reaction rate enhancements with microwave heating, comparable to the examples described in chapter 3. Variation of substituents, temperature and amount of a co-solvent essentially influenced the magnitude of the microwave effects. Based on the solubility and reactivity of various C6-substituted homologues of isocytosine the magnitude of the microwave effect could be predicted. The diffusion layer near the solid surface is the area where selective heating by microwaves is believed to occur. As a result locally higher temperatures increase solubility and reaction rates while, in contrast, a relatively lower bulk temperature is measured. In brief, the observed microwave effects have a thermal rationale based on direct, fast and selective heating. This effect is not reproducible by conventional heating. .
68
Chapter 4
4.1 Introduction
In the previous chapter the racemization of Nacetyl amino acids revealed the importanceofaheterogeneouscharactertoobserveamicrowaveeffect.Giventhe heterogeneity of the nucleophilic addition of isocytosines to isocyanates (Scheme 4.1)startingmaterialisocytosineandendproductarepracticallyinsolublethis processseemedtobeavaluablecandidateforasuccessivestudy.
N H2N N H O Neat NCO OCN H N O H N H N N O
+
OCN
Scheme4.1
The addition of an amine to an isocyanate leads to a urea functionality, which is capable of undergoing intermolecular hydrogen bonding. In case of isocytosine this leads to a selfcomplementary ureidopyrimidinone (UPy) moiety that is characterizedbyquadruplehydrogenbonding.AmonofunctionalUPydimerizes and a bifunctional UPy leads to a supramolecular polymer. The monofunctional Upy with a reactive terminal isocyanate can be used for postpolymerizations (e.g.polymericalcohols).Theintroductionoftwoureidopyrimidinonemoietiesin one molecule makes the latter suitable to act as a monomer for supramolecular polymerization(Figure4.1).
NH O H N N O HN O HN NH N N H O
Figure4.1
Incorporation of two ureidopyrimidinone (UPy) moieties in a block co polymer leads to a macromonomer for supramolecular polymerization (rightpartofFigurefromreference1).
This work has been initiated and studied in detail by the groups of Meijer and Sijbesma,culminatinginacommercialproductlinesuppliedbySupraPolixB.V.2,3, SupraPolixB.V.,Eindhoven,TheNetherlands.
69
Chapter 4
Potentialapplicationsareinthefieldofadhesives4,printing5,cosmetics6,personal care7andcoatings8. With the knowledge gained in our group (Applied Organic Chemistry) from a previous study,9 it would be of interest to investigate the synthesis of the ureidopyrimidinonemoiety,intermsofrateandmonoselectivitywithmicrowave heatingincomparisonwithconventionalheating.Mostimportantly,variationsin the reaction parameters may enable to find a rationale for potential differences betweenbothheatingmethods.
Masstransfer control
Kinetic control
Figure4.2
Temperature effect on the overall rate coefficient adopting pseudo first orderkinetics[reference9]
EquipmentfromBelatecAG,Wintersingen,Switzerland.
70
Chapter 4
Below 100 C monoadduct is selectively formed. Monoaddition, essential for further downstream chemistry, is rationalized by the limited solubility of 6 methylisocytosine in HDI and by the presence of an excess of HDI. After mono additiontheproducthasanevenlowersolubilitypreventingtwofoldaddition. Influenceoftemperature Timeconversion experiments were performed at three temperatures to find ideal conditions for reaching completion within a time span of one to three hours, see Figure4.3.Thisdurationwasselectedfromtheprospectoffurtherprocessscaling up.
100 100 C 85 C 74 C
75
Conversion (%)
50
25
Time (min)
Figure4.3
Inthefirstperiodafterthestartofthereaction,uptoconversionsofapproximately 75 %, the reaction nearly obeys (pseudo) zeroorder kinetics, see Figure 4.3. The substrate is sparingly soluble and hence is present in solution at a constant, very lowconcentration.HDI,onthecontrary,isassolventandreactantinitiallypresent in7.5molarexcessandis,therefore,alsoconstantthroughoutthereaction. An accurate value of the solubility of 6methylisocytosine in the reaction mixture couldnotbedeterminedandthereforeabsolutevaluesfortheratecoefficientcould notbeproducedfromtheArrheniusequation(seeEquation4.1a). Ea dx Ea = A e RT r = C0 (4.1b) (4.1a) dt k = A e RT dx A Ea , apparent ln = ln C RT dt 0 t =0
Ea,apparent=apparentactivationenergy.
71
Chapter 4
InsteadofintroducingthekvalueastherateconstantinEquation4.1a,(rCH)t=0an initial reaction rate was applied to estimate the apparent activation energy,10 see Equation 4.1b. The initial conversion rate [d(x)/d(t)]t=0 is expressed in [min1]. An overview of the initial conversion rates is given in Table 4.1 depending on three temperaturesandsomeweightfractionsofcosolventNMP. Table4.1 Overview of initial conversion rates (rCH ) of nucleophilic addition of 6 methylisocytosine to hexamethylene diisocyanate at different temperatures and some weight fractions of cosolvent (NMP) under conventionalheatingconditions.
entry 1 2 3 temperature 74C 85C 100C withoutcosolvent* 1.2 2.8 12.5 conversionrate 1wt%NMP* 3.3 4.8 13.7 5wt%NMP* 9.5 14.9 29.1
*[103.min1] Influenceofacosolvent Inanadditionalseriesofexperimentsitwasfoundthattheheterogeneitycouldbe influencedbyadditionofacosolvent(seeTable4.1andFigure4.4).Reducingthe heterogeneity increased the reaction rate as expected. NMethylpyrrolidinone (NMP)wasselectedasacosolventbasedonearlierwork.9
100
75
Conversion (%)
50
CH
25
60
120
180
240
Time (min)
Figure4.4
Nucleophilic addition of 6methylisocytosine to hexamethylene diisocyanate in the presence of cosolvent (NMP) at 85 C with conventionalheating(CH).
72
Chapter 4
Scopewithvarioussubstratesandreagents To broaden the scope of the addition of 6methylisocytosine to hexamethylene diisocyanate, other substrates and isocyanates were selected. Various C6 substituted isocytosines were synthesized (Scheme 4.2).11 Additionally, HDI was replaced in the reaction with 6methylisocytosine by nhexylisocyanate (HI, Scheme4.3).
R N H2N N H O Neat OCN H N O H N H N N O R = C2H5, i-C3H7, Ph
OCN
Scheme4.2
Scheme4.3 Nucleophilicadditionof6methylisocytosinetonhexylisocyanate. Theuseofmonofunctionalizedisocyanates(Scheme4.3)automaticallyovercomes the possibility of twofold addition. With respect to the reaction with the diisocyanate, only a small change in solubility of isocytosine and, therefore, in conversionratecanbeexpected.Duringpreliminaryexperimentationtheaddition of 6methylisocytosine to HI revealed the significant influence of an impurity, leading to irreproducible timeconversion histories. A striking difference in reactionratewas observed between an aged, openbottle of HIand anunopened one.
NCO + H2O - CO2 NH2 + OCN
H N O
H N
Scheme4.4
Reactionofnhexylisocyanate(HI)withwater(<1eq.).
73
Chapter 4
When HI from a new bottle was taken, the reaction needed more time for completion than in the case of the opened bottle. A solid precipitate in the aged bottlewasidentifiedasN,Ndihexylurea.Thiscompoundresultsfromareactionof hexylisocyanatewithwater,seeScheme4.4.Theinfluenceofwateronthereaction as depicted in Scheme 4.4 was not detectable nor was the occurrence of an analogousimpurityinHDI. At room temperature N,Ndihexylurea is a solid, but at the reactor temperatures used, it has melted and this liquid phase acted as a cosolvent. When N,N dihexylurea was subsequently deliberately added to the reaction of 6 methylisocytosine with HI, reaction rates significantly enhanced, see Figure 4.5. Similar rate enhancements were observed with the addition of N methylpyrrolidinoneasacosolvent.
100
75
Conversion (%)
50
25
HDI plus N,N-dihexylurea 1st & 2nd experiment from: old bottle (HI) new bottle (HI) 0 60 120 180 240 300 360
Time (min)
Figure4.5
Variationofthesubstituentatthe6positionofisocytosinewasexpectedtolargely influenceitssolubility12andasaconsequencetheoverallreactionrateasexplained earlier (Influence of a cosolvent). The results of the relationship between various substituentsatthe6positionandthesolubilityoroverallreactionratearelistedin Table4.2. AlthoughthesolubilitycouldnotbemeasureddirectlyinHDI,thesolubilitywas visuallycategorizedandmeasuredintwoothersolvents,N,Ndimethylacetamide (DMA)and1,4dioxane.
74
Chapter 4
Table4.2
entry
OutcomeofvaryingsubstitutionsattheC6positionofisocytosinewith respecttosolubilityandoverallreactionratewiththeadditiontoHDI.
sol.inDMA [mol/l]* 168 555 1232 1079 sol.in1,4dioxane [mol/l]* <0.1 10.1 27.9 40.5 overallreactionrate low medium medium higher
substituent
The polarity of these solvents was chosen above (DMA) and below (1,4dioxane) that of HDI. Ranking the various alkylated isocytosines with respect to solubility explains an increase in reaction rate. However, the selectivity and reactivity of 6 phenylisocytosine differ significantly from those of 6isopropylisocytosine with a comparable solubility. A plausible rationale for this behavior may relate either to theproductsolubility,ortotheketoenoltautomerizationoftheisocytosines(see Scheme 4.6) that may influence the intrinsic reaction rate of the amine functionality,respectively. Thesolubilitiesofthemethyl,ethyl,isopropylandphenylsingleadditionproducts in 1,4dioxane are < 0.10; < 0.10; 67.2 and 31.5 mol/l at 85 C, respectively. The higherthesolubilityofthesingleadditionproductthelowertheselectivity,dueto potentiallytwofoldaddition(Scheme4.5).
Neat N H2N N H O OCN H N O H N H N N O
+
OCN NCO H2N H N H
N N H O
N O N
O N H
H N O
H N H N
Scheme4.5
75
Chapter 4
Sointhealiphaticseries,thereactionratereflectsthesolubilityoftheisocytosines andtheselectivitydecreasesfrommethyl,ethyltoisopropyl.6Phenylisocytosine is, however, exceptional since the solubility of the starting material is relatively high,whilethemonoadductislesssolublethanthestartingmaterial. In general, isocytosines may be represented by three interconverting tautomers (Scheme 4.6). Calculations on the charge separation for each tautomer might explainthehigherreactionrateofthe6phenylisocytosine.
Keto' steric hindrance Keto Enol
H N O N H NH2 O N H N
H N NH2 HO N
NH2
long dipole O N
H N H NH2 N H N O
Scheme4.6 Ketoenoltautomerizationof6phenylisocytosine. Theaberrantprofileof6phenylisocytosinemayrelatetoitspreferencefortheenol formandtheconcomitantenhancednucleophilicityoftheaminefunction. However,extendedHckelcalculationswiththeMM214andMMFF9414modelsdid not unequivocally rationalize the difference in reactivity between 6 phenylisocytiosineand6isopropylisocytosine. Anextended(IR)studyontheketoenoltautomerizationforureidopyrimidinones demonstratedafavortowardstheenoltautomerwithelectronegativesubstituents ontheC6position.13 Fortunately, IRmeasurements showed indeed a higher absorption wavenumber for the NH2 band (blueshift / more nucleophilic) when 6phenylisocytosine was
76
Chapter 4
comparedwith6isopropylisocytosine(Table4.3).Thisdifferenceinthelocationof theNH2bandwasnotobservedamongthethreealkylanalogues. Table4.3 Some typical values of absorption bands in the IRspectra of 6 isopropylisocytosineand6phenylisocytiosine.
entry 1 2 3 bandtype OHstretch NH/NH2stretch C=Ostretch 6isopropylisocytosine 3333cm1 3072cm1 1657cm1 6phenylisocytosine 3374cm1 3134cm1 1668cm1
Presumably,theaminefunctionof6phenylisocytosineismorenucleophilicdueto the electrondonating character of the phenyl moiety in comparison with the aliphaticC6substitutedisocytosines.
77
Chapter 4
100 CH 75 MW 100 C 85 C 74 C
7 6 5
Conversion (%)
50
f MW
4 3 2 1
25
0 70 80 90 100
Time (min)
Temperature (C)
Figure4.6
(left) Nucleophilic monoaddition of 6methylisocytosine to hexa methylenediisocyanate at 74 C, 85 C and 100 C for conventional heating (CH) and microwave heating (MW). (right) Relation between temperatureandmicrowaveeffect(fMW).
Plottingtheinitialreactionrateagainstthereciprocaltemperatureforexperiments with conventional and microwave heating (see Figure 4.7 and Equation 4.1b) showsthattheintercept[lnA/C0]inthegraphdiffersmorethantheslope[Ea/R] of both lines.15 The preexponential factor and the apparent activation energy comprises multiple physical parameters, the difference between the pre exponential factors (i.e. ln A/C0) points out a plausible explanation for the observedrateenhancements.
MW CH
MW CH
ln rini
ln rini 1/T (K )
-1
1/T (K )
0 2,70 -1 2,80 2,90 3,00
-1
-1
17 C shift
-2
-2
Figure4.7
(left) Correlation between temperature and initial reaction rate for nucleophilic monoaddition of 6methylisocytosine and hexamethylene diisocyanate using conventional and microwaveheating.(right) Results of conventional heating shifted with 17 C to project on microwave heatingresults.
78
Chapter 4
The apparent activation energy, consisting of the sum of actualactivation energy andheatofsolution(chapter3.2),remainsthesameunderbothheatingtechniques. The preexponential factor, although, seems to differ. Thus, this leads to the assumptionthatlimitationsintheratedeterminingsteparelesspronouncedunder microwaveirradiation. Influenceofthecosolvent Asdiscussedinsection4.2theconventionallyheatedexperimentsshowedthatthe use of a cosolvent (e.g. NMP) resulted in higher reaction rates (Figure 4.3). The resultsinFigure4.8(right)demonstratethatthemicrowaveeffect(fMW)decreases with increasing weight fraction of NMP. Further addition of NMP to reach a homogeneoussolutionlikeintheracemizationofNacetylaminoacids(chapter3) was unrealistic. Replacing 6.5 equivalents of HDI by NMP did not result in a homogeneous reaction mixture. It can be concluded that the addition of a co solvent leads to a partial dissolution of 6methylisocytosine and a concomitant vanishingmicrowaveeffect.
100
5
75
Conversion (%)
f MW
CH MW 5 wt% NMP 1 wt% NMP 0 wt% NMP 0 60 120 180 240
50
25
0 0 1 2 3 4 5
Time (min)
Co-solvent (%)
Figure4.8
(left) Nucleophilic addition of 6methylisocytosine to hexamethylene diisocyanate at 85 C for conventional heating (CH) and microwave heating (MW): influence of the weight fraction of cosolvent NMP. (right)MicrowaveeffectasafunctionofweightfractionofNMP.
Scopewithvarioussubstratesandreagents Heterogeneitycanbealtered,likeintheprevioussectionbycosolventaddition,as well as with variation of the reactants. Not only would this expand the scope of microwaveassisted addition of 6methylisocytosine to hexamethylenediiso cyanate,butalsoitmightgivemoreinsightintothemicrowaveeffect. The differences between both heating techniques are depicted in Figure 4.9 (left) forthemonoadditionsof6methylisocytosinetoHDIandHIat85C,showinga fMWforHIof7.4andafMWforHDIof4.9.Apossibleexplanationforthelowerrates
79
Chapter 4
with HI as compared to HDI is the expected poorer solubility of 6 methylisocytosineinHIduetoitslowerpolarity. Variation of the substituents at the 6position in the substrate gives also large differences between conventionalandmicrowave heatingat 70 C, see Figure 4.9 (right).Thecorrelationbetweensolubilityandreactivitywasingeneralpredictable as explained in the previous section of this chapter. However, the correlation between solubility and microwave effect is not completely in line with our expectations (Table 4.4). A better correlation between heterogeneity and microwave effect (fMW) for all four substrates can be made when the overall heterogeneityofthereactionmixtureisconsideredforbothsubstrateandproduct solubility during the conversion (mainly initially). This might explain the high microwaveeffectfor6phenylisocytosineincontrasttotheabsenceofamicrowave effectfor6isopropylisocytosinewithasimilarsubstratesolubility. Thereactivitiesof6ethyland6isopropylisocytosinearenotsignificantlydifferent for conventional heating (see Table 4.2). However, with microwave heating only the rate of 6ethylisocytosine is enhanced. During the reaction of 6 isopropylisocytosine the solubility increases to such a level that no microwave effectcouldbegeneratedanymore.
100
100
MW CH HDI HI
80
75
Conversion (%)
Conversion (%)
60
MW CH Phenyl Ethyl Methyl Isopropyl
50
40
Duplicate
25
20
Time (min)
Time (min)
Figure4.9
(left) Nucleophilic addition of 6methylisocytosine to hexylisocyanate (HI) and hexamethylenediisocyanate (HDI) at 85 C for conventional heating (CH) and microwave heating (MW). (right) Nucleophilic addition of several 6substituted isocytosines to hexamethylene diisocyanateat70C(andfor6methylisocytosineat74C).
80
Chapter 4
Table4.4
entry
substituent
1 methyl low low 2 ethyl medium 3 isopropyl low higher 4 phenyl higher *fMWat70C(andfor6methylisocytosineat74C)
81
Chapter 4
6methylisocytosine for the conventionally heated experiment at 100 C almost similar to the microwaveheated reaction at 85 C. If the reaction predominantly takes place in the diffusion layer then this layer conclusively has an average temperature of 100 C, but the bulk temperature is measured to be 15 C lower. Again,avaluecloseto17Cderivedfrompreexponentialfactorconsiderationsin section4.3. Improvement of stirring has an effect on the thickness of the diffusion layer, neverthelessdidnotleadtoavanishingmicrowaveeffectinourcases.Incontrast to the racemization process as described in chapter 3 a homogeneous solution could not be obtained and, therefore, a diffusion layer always remained present. Consequently,asimilarmicrowaveeffectisstillobserved.
4.5 Remark
Shifting point of view: microwave effect. Electromagnetic acoustic transducers (EMAT)aredeviceswhichgenerateelectromagneticpulses.Ateachboundaryordefectan acousticwaveisformed.Byreflectionthespecimenrevealsitsinnerstructure.Thismethod is generally used to test metal objects. The creation of an electromagnetic pulse allows formationofanacousticoranultrasonicwaveatboundarylayers. This phenomenon is also the reason why humans or animals exposed to electromagnetic waves hear clicks. A thermal elastic wave is produced and it travels to the inner ear, whereasoundisheard.Thethermalexpansioninthiscaseissmall,butveryfast. Two specialists (Hosten (Bordeaux, France)19 and Lin (Illinois, Chicago, USA))20 in this field have not excluded the existence of the transformation of electromagnetic waves to ultrasonic or acoustic waves in our heterogeneous reaction systems. Although EMAT technologyisusedatanotherscaleandenergiesarehigherinpower,thegenerationoflocal ultrasonic waves may have a positive influence on the reaction rate of solidliquid reactions.21Cavitationattheinterfaceofsolid/liquidcanlocallystimulatemasstransport or temperatures. The application of ultrasonic waves is a generally accepted alternative energysourcetoactivateaspecificreaction(e.g.Grignardreagentformation),alsoknown assonochemistry.
82
Chapter 4
83
Chapter 4
N(6Isocyanatohexyl)N(6isopropyl4oxo1,4dihydropyrimidin2yl)urea: A 20mL reaction tube was charged with 6isopropylisocytosine11 (0.92g, 6.0 mmol) and hexane1,6diisocyanate (7.5g,44.6mmol). The tube was flushedwithargon and closed. Temperaturewasmeasuredviaaninsert.Thereactionmixturewasheatedat85Cwith stirringfor2hinthemicrowaveoven(averagepower:24W/max:70W).Thereafter,the mixturewasdilutedwithMTBE(15mL),andaftercoolingthesuspensionwasfilteredand washedtwicewithMTBE(1.5mL).Theremainingsolidwasdriedunderreducedpressure at 40 C and analyzed. 1HNMR (400 MHz) in CDCl3/dTFA (1/1 vol %) 6.4 (s, 1H, CH=CCH3),3.3(t,2H,NHCH2,3.2(t,2H,CH2NCO),3.0(m,1H,C2H6CHC=CH), 1.8 (m, 2H, NHCH2CH2), 1.6 (m, 2H, CH2CH2NCO), 1.4 (m, 4H, (CH2)2(CH2)2 NCO,1.3(d,6H,C2H6CHC=CH).Yield:63%. N(6isocyanatohexyl)N(6phenyl4oxo1,4dihydropyrimidin2yl)urea: A 20 mL reaction tube was charged with 6phenylisocytosine11 (1.12g, 6.0 mmol) and hexane 1,6diisocyanate (7.5 g, 44.6 mmol). The tube was flushed with argon and closed. Temperaturewasmeasuredviaaninsert.Thereactionmixturewasheatedat85Cwith stirringfor2hinthemicrowaveoven(averagepower:22W/max:120W).Thereafter,the mixturewasdilutedwithMTBE(15mL),andaftercoolingthesuspensionwasfilteredand washedtwicewithMTBE(1.5mL).Theremainingsolidwasdriedunderreducedpressure at40 C andanalyzed. 1HNMR(400MHz) in CDCl3/dTFA (1/1 vol %) 7.67.8 (m, 5H, C6H5C=CH), 6.8 (s, 1H, CH=CCH3), 3.4(t,2H, NHCH2, 3.2 (t, 2H, CH2NCO), 1.8 (m, 2H, NHCH2CH2), 1.7 (m, 2H, CH2CH2NCO), 1.5 (m, 4H, (CH2)2(CH2)2 NCO).Yield:83%. NhexylN(6methyl4oxo1,4dihydropyrimidin2yl)urea (5). A 20mL reaction tubewaschargedwith6methylisocytosine(0.75g,6.0mmol)andhexylisocyanate(5.67g, 44.6mmol).Thetubewasflushedwithargonandclosed.Temperaturewasmeasuredvia an insert. The reaction mixture washeated at85 C with stirring for4 h in aMilestone microwaveoven(averagepower:25W/max:200W).Thereafter,themixturewasdiluted withMTBE(15mL),andaftercoolingthesuspensionwasfilteredandwashedwithMTBE (1.5mL).Theremainingsolidwasdriedunderreducedpressureat40Candanalyzedby 1HNMR(400MHz)inCDCl3/dTFA(1/1vol%)6.3(s,1H,CH=CCH3),3.3(t,2H, NHCH2,2.5(s,3H,CH3C=CH),1.8(m,2H,NHCH2CH2),1.5(m,6H,(CH2)4(CH2)2 NH),1.4(t,3H,CH2CH3).Yield:73%.
84
Chapter 4
Typicalprocedureforalloilbathheatedreactions: Thesameprocedurewasexecutedfortheoilbathexperimentsubstitutingmicrowavewith oilbathheating,exceptthedurationofthereactionmightbeprolonged. N,Ndihexylurea (6). A 20mL reaction tube was charged with nhexylamine (234 mg, 2.31mmol),toluene(5mL)andhexylisocyanate(280mg,2.2mmol).Thereactionmixture wasrefluxedfor2h.Thereafter,thesolventwasremovedat50Cunderreducedpressure. 1HNMR(300MHz)inCDCl34.3(s,1H,NHCO),3.2(t,2H,CH2NH),1.5(m,2H, CH2CH2NH),1.3(m,6H,(CH2)3CH3),0.9(t,3H,CH3).GCMS(m/z):228(FW228).
11 12
13
85
Chapter 4
14 15
MM2 and MMFF94 are Molecular Mechanism and Merck Molecular Force Fields models;Halgren,T.A.J.Comput.Chem.1996,17,490. Eyring equation (statistical thermodynamics) defines ln A as ln (kB/h) + S/R if ln(rMW)t=0isdividedbyT;(kB =Boltzmannsconstant,h=Plancksconstant,S= entropyofactivation),Eyring,H.,J.Chem.Phys.1935,3,107. Conner,W.C.;Tompsett,G.A.J.Phys.Chem.B,2008,112,2110. Chemat, F.; Esveld, D. C.; Poux, M.; DiMartino, J. L. J. Microwave Power Electromagn.Energy1998,33,88. Chemat,F.;Esveld,E.Chem.Eng.Technol.2001,24,735. (a) Hosten, B.; Bernard, P. A. J. Acoustical Soc. Am. 1998, 104, 860 (b) Hosten, B.; Bacon,C.;Guilliorit,E.Ultrasonics2002,40,419. Lin,J.C.;Wang,Z.W.HealthPhys.2007,92,621. Ullmannsencyclopediaofindustrialchemistry,Sonochemistry,chapter3.1.Solid fluidreactions,WileyVCHVerlag,Weinheim,2001.
16 17 18 19 20 21
86
Chapter 5
Chapter 5
Abstract
The best opportunities to scaling-up microwave-assisted chemistry lie in continuous operation or in flow processing. An effective transformation of organic reactions from batch to continuous operation is only suited for fast reactions. The introduction of a loop expands the scope of the continuousflow reactor. Microwave heating is more expensive in the acquisition and in its energy consumption. However, efficiency grows as the loss tangent or the scale increases. Nonetheless on an industrial scale, steam heating stays more beneficial when other criteria, such as safety, are less decisive. The development of a fume hood concept allows having a large degree of flexibility over process research, with the option to produce a considerable amount of product. Further scaling out of the fume hood concept is expected to be fast and straightforward with parallel units, if the costs are acceptable within the process budget. Although the technology is available to scaling-up processes under microwaveheating conditions, the sum of all process criteria shows whether it is a preferred method of operation.
88
Chapter 5
5.1 Introduction
Inrecentyearsmuchefforthasbeeninvestedinanincreasinglycompetitiveworld to meet the requirements of the fine chemical and pharmaceutical industry with respect to firsttimeright performance, a shorttimetomarket and avoiding surprises during process scaleup. A demand for larger quantities is not just rewritingarecipebyreplacingmgwithkgduringchemicalprocessresearch. Process intensification based on e.g. microreactor technology and microwave heating, as relatively novel tools in the field of organic chemistry, is actively pursued to achieve a better position in the industrial scene. Introduction of these novel techniques in the field of organic chemistry expands the toolbox of todays chemists. Increase of reaction rates and improved selectivity, combined with the automationofrepetitiveprocedures,demonstratetheadvantageousapplicationof these enabling techniques and are considered to be suitable to improve this position in the industrial scene. One of the major weaknesses of microwave assisted chemistry is scalingup. Since the penetration depth of microwaves is limited,thebestchancesforincreasedproductionvolumeslieinrelativelysmall scalecontinuousoperations,somewhatuncommoninthefinechemicalindustry.13 Inthefirsttwosectionsmicrowaveheatingisdiscussedintermsofapplicationin continuousoperationsandinaloopreactoroperatedbatchwise.Thenmicrowave irradiationisconsideredasapotentiallyefficientheatingsource.Subsequently,the tubular reactor part of our Milestone FlowSynth equipment is discussed to characterize and to model the reactor with respect to standardized equipment. Ultimately, the performance of the FlowSynth is illustrated in the fume hood concept.
89
Chapter 5
problemtoovercomewhenmicroreactorsareused.Withsolid/liquidsystemsthere isariskofplugging.Forcatalyticreactorsthepluggingbycatalystparticlescanbe overcomebycatalystcontainingcoatingsorafixedbed.Increasingthediameterof thetubingabove1mmopensupawindowforheterogeneousreactions,especially solid/liquidsystems. As mentioned before the penetration depth (p) of microwaves is limited in the rangeofcentimetersdependingonthecompositionandphysicalpropertiesofthe reaction mixture. The penetration depth is expressed in the distance where the energy is reduced to 37 % (see Equation 5.1).5 Therefore, the reactor diameter for microwaveheatingequipmentrangesbetween5to100mm.6 Basedonthespecificationofthereactionconditions,selectionofreactortypeand peripheral equipment (for example type of pump) has to be made. So far in our caseonlyonetypeofmicrowaveequipmentiscommerciallyavailableinvolvinga combination of a membrane pump, a tubular reactor and a microwave oven; all controlled by one computer. This setup is known as FlowSynth of Milestone srl, Italy(Figure5.1(left)).7,8TheprecommercialversionoftheFlowSynthisdescribed bythegroupofOndruschka.9Thisnoveltyhasbeenpatented.10,11
p =
(5.1)
Figure5.1
(left)FlowSynthofMilestonesrl,Italy.(right)Continuousflowreactor fromthegroupofStrauss.20
90
Chapter 5
Flowchemistry Most studies with continuousflow reactors are reported in the field of organic chemistry and describe the application of conventional heating; socalled flow chemistry.1114 Todays flow chemistry is based on the principle of Merrifields solidphase peptide synthesis.1416 Solidphase chemistry itself is specifically used under flow conditions11andthistypeofworkhasbeenextensivelystudiedandpromotedby, forexample,LeyandBaxendale.18,19 InitialresearchinflowchemistrybymicrowaveheatingwasdonebyStraussand coworkers.Figure5.1(right)showstheequipmentused.Thefirstpublicationfrom the group of Strauss appeared in 1992 with positive expectations of this novel reactordesignforthefuture.2023Thecombinationofmicrowaveheatingandflow chemistryhasalsobeeninvestigatedbygroupswhichhadtheirexpertiseprimarily ineithermicrowavetechnologyorflowchemistry.1,9,2429Attemptstoscaleupflow chemistrymostlyinvolveeitherhomogeneousreactionsorheterogeneousreactions by means of immobilized reagents / catalysts.3033 There is, to the best of the our knowledge, only one smallscale example of an inorganic solidliquid flow system.34 The use of continuously operated reactors demands for relatively fast reactions, otherwiseflowrateshavetobecomeextremelyloworreactorsbecometoolarge.35 Residence time is the key parameter determining the degree of conversion. One waytocombineahighflowratewithcompleteconversionistocompletelyrecycle theoutlet,resultinginabatchloopreactor.Inthisway,theadvantagesofatubular reactorandabatchreactorarecombined.
91
Chapter 5
Figure5.2
Picture of the batchloop reactor in the Laurydone process. On the left part of this picture the tank is situated with a recycle loop including pump. In the middle section a safety screen is positioned and the right part depicts on the bottom a MW control unit on top of that the MW generator and adjacent the MW applicator. The recycle loop is partly situatedintheMWapplicator[reference36].
Tank
Pump
Our process design is predominantly based on the FlowSynth, which was commerciallyprovidedbyMilestonesrl(Figure5.1(left)andFigure5.4). The reactants are preferably premixed in a vessel. In another small vessel pure solventisstoredonbehalfofstartupandshuttingdownprocedures.Withaseries ofballvalvesitispossibletoswitchbetweenseveralflowprofilesincludingoneto clean the reactor coil in a reverse flow. All the liquids or slurries are transported either by a membrane pump or by a gear pump. The pump forces the reaction mixture into the microwaveheated reactor. Inside this tubular reactor mixing is possible. On top of the reactor a cooling unit is placed. The cooler is manually controlled in temperature, but it can be switched on/off by the computer. At the outletinitiallyabackpressurevalvehasbeeninstalledtocontrolthepressuretill30 bars.Twosensorsmonitoronlinethetemperatureintheupperpartofthereactor coil and at the final outlet. The pressure is monitored at the outlet of the pump. Besidestheseparametersthestirrerspeed,pumpspeedandthemicrowavepower areregisteredbythesoftwareprogram. The tubing has been thermally insulated to prevent drastic temperature drops, which largely influence the solubilityof the reactants. Plugging in these channels orinthevalvesisoneofthekeyissuesinthissetup.Inprincipletheloopreactor canbeoperatedasaclosedloopsystemorinacontinuousmode.Thecontinuous
92
Chapter 5
modewillbesimilartooneoraseriesofcontinuousstirredtankreactors(CSTRs). In the batchloop as well as in the continuous mode a feed point can be incorporatedintheloop.Feedsofgasesandliquidsarethenrapidlydispersedin the circulating liquid. Consequently, the reaction rates are improved provided reaction rates are controlled by mass transfer between the phases. A typical exampleistheBussreactorappliedforhydrogenations,seeFigure5.3(right).37
T2
Pump
Figure5.3
(left)Batchloopreactorandcomponents.(right)Bussloopreactorfrom BussChemTechAG,Switzerland.
Based on the results and the experience gained in the processes discussed in chapter 6 (e.g. biocatalysis or racemization), the FlowSynth reactor has been adapted at various places to enable processing of heterogeneous reactions. Our modificationsaremarkedwithcirclesinFigure5.4.
Removal of back-pressure valve (potential blockage of solids). Removal/widening of stirrer bottomadapter (potential blockage of solids). Widening of inlet (potential bottleneck). Change in ball-valves of pump (potential blockage of solids) and change of pump (potential blockage of solids). Change in tubing (practical reasons).
Figure5.4
VariousmodificationsoftheoriginalFlowSynthapparatus.
93
Chapter 5
50%
44%
Electricity
5065%
Steam Microwave energy
85%
Reaction mixture
8090%
Product:43%
Product:18 26%
Figure5.5
94
Chapter 5
Different energy sources (gas or electricity) include different prices, based on the technologynecessarytogeneratethispower.4144Inshort,microwaveheatingloses atleastafactorof2(43%over22%,Figure5.4)timesmoreenergyinitsforetrack andattheendtheenergybillisminimalafactorof8(0.088over0.011)44higher comparedtosteamheatingforindustrialapplications.Altogether,theenergycosts willbehigherformicrowaveheating. Note that, highly dedicated equipment tuned to a specific process can reach the highest efficiencies compared to a single process in a multipurpose plant. Nevertheless,ifmicrowaveheatingcanbeappliedtoprocesseswithhighdegrees ofselectiveheating(drying),thistechniquemightevenhaveabetterheatefficiency ascomparedtoconventionalheating. The efficiency values given in Figure 5.5 are based on different literature sources and experimental results.4144 The efficiency of the transition from microwave energy to the reaction mixturesis themost fluctuating value depending onscale, equipmenttypeandthelosstangent.Hoogenboometalreportedthatthelosseson smallscalearemuchhighercomparedtolargerscales.45Predominantlythisisdue toheatlossestotheenvironment,basedonahighsurfacetovolumeratioforsmall equipment.Thermalinsulationofthereactionvesselwithmicrowavetransparent material is a simple solution to increase the energy efficiency. Fortunately, a low heat conductivity is also a good material property of polymer materials such as TeflonandPEEK(polyetheretherketone)whicharecommonlyusedtoconstruct microwaveheated reactors with a high chemical resistance or high mechanical strength(e.g.forpressure),respectively. A more detailed overview of the efficiencies of microwave energy into heat for threesolventsunderdifferentconditionsisdepictedinTables5.4(forpxylene),5.5 (for demiwater) and 5.6 (for ethylene glycol) in section 5.8. The efficiency was calculated by the ratio of the thermal energy required to heat up the solvent (see Equation 5.2) and the amount of electromagnetic energy delivered by the microwaveoven. (5.2) c T
( )
observed
microwavepowerused
1 1 1
m = massflowrate[kg s ]
c p = heatcapacity[ J kg K ]
T = difference in temperature,outletinlet[ K ]
= efficiency[ % ]
95
Chapter 5
Experiments were performed with the FlowSynth equipped with a typical stirrer showninFigure5.6.
Figure5.6
Stirring shaft with three Weflon blades for the tubular reactor of the FlowSynth.
Theexperimentalresultsdemonstratethattheabsorptionofmicrowavesincreases in the order of pxylene < demiwater < ethylene glycol. This could be expected from the following literature data; loss tangents for ethylene glycol, demiwater and pxylene at 20 C are 1.35; 0.127 and 0.0011, respectively. Operating in a turbulent flow regime (see section on Flow and mixing) improves the efficiency. Eachsolventappearstohaveanaveragemaximum,whichisnotinfluencedbythe operationalconditions.Theerrormarginofthesemeasuredvaluesisintherange of510%. Besidesenergyefficiency,safetyissuescanbedecisiveinselectingthemethodof operation.Asanexample,acasestudyonhydrogencyanideproductionmadeby DuPont illustrates the advantages of microwave heating in the processing of hazardous chemicals.46 Here hydrogen cyanide is synthesized using a platinium catalyst at temperatures over 1000 C. The advantage of microwave energy is selective heating and the possibility of an immediate interruption of heating. However,tenyearsagoalackofavailabilityoftechnologyandthehighoperating costsmadeDuPonttodecidetogofortheconventionalway.Nevertheless,DuPont patentedtheadvancedtechnology.47
physical constants are obtained from Handbook of Chemistry and Physics; 89th Edition; 20082009(onlineedition,Lide,D.R.)
96
Chapter 5
N0
C(t).dt
0
Benzylalcohol,astracerinastreamoftoluene,hasbeenmeasuredbyGCFID.In this way there is a linear relation between the measured GCpeak area and the actual concentration (C(t)). The exitage function E(t) is used in a oneparameter modelfornonidealreactorsimplyingthenumberofequallysizedCSTRsinseries whichgivesthesameE(t)asforthereactorconsidered.Theonlyparameteristhe numberoftanks,n,inthisseriesanditisexpressedasgiveninEquations5.4. The number of tanks (n) is calculated from the dimensionless variance. The varianceitselfisthesquareofthestandarddeviation()andthemeanresidence timeis<>.48
97
Chapter 5
= t.E(t )dt
0
(5.4a)
E(t ) =
n t
n
n 1
nt
(5.4b)
( n1) !
2 = (t
0
.E(t )dt
n=
TheRTDresultsareinfluencedwhenthereactorsetupisoperatedunderdifferent conditions. The relationship between the number of equally sized tanks in series andthevariationofflowandstirrerspeedaresummarizedinTable5.7,section5.8. On average the FlowSynth is equivalent with approximately three equally sized CSTRsinseries.Iftheflowinsidethereactorisinsynchronizationwiththestirrer speed then an optimum can be observed. This indicates that the stirred tubular reactor tends to perform more like a plugflow reactor. Any disturbance leads to moreaxialmixing,whichhasthetendencytobemoreequivalenttoacontinuous stirredtankreactor(CSTR).Thepresenceofashoulderorsplitpeakisasignfor twoormultiplerouteswhichleadtodifferencesintheresidencetime(seeFigure 5.9 (left) in section 5.8). Figure 5.9 (right) in section 5.8 illustrates the theoretical distributionforanumberofCSTRsinseries. Flowandmixing The Reynolds number, as a dimensionless number, represents the ratio of inertia forcesandviscousforces.TheReynoldsnumbercharacterizesinthiswaytheflow regimeoftheexperimentalconditions. Mixingcanbecrucial,notonlytobringthereactantstogether(masstransfer),but also to remove/supply the heat for temperature control during the reaction (heat transfer).Formassaswellasforheattransferthereactionrateisinfluencedbythe degreeofmixing.Table5.1givesanoverviewofthelengthscalesofconcentration levelingandtimeconstantsofreaction.51 Table5.1 Mixingregimesrelatedtoreactiontimeconstants51
entry timeconstant(treaction) typeofmixing 1 10minhrs independentofmixing(intrinsickinetics) 2 secmin macromixing*(i.e.circulation) 3 ms micromixing*(i.e.diffusion) *Thesenamesreflectthedifferentlengthandtimescalesofthedominatingmechanism.
98
Chapter 5
Thetubingandthereactorhavedifferentdiameters.Dependingontheflowratea certain reaction mixture or solvent has a laminar or turbulent pattern. If the Reynolds number (Re) is over 3500 the flow will be turbulent in a tube, see Equation 5.5a. In case of a turbulent flow the heat exchange from the Weflon blades of the semi coilshaped stirrer inside the reactor column towards the reaction mixture is much more efficient. Table 5.2 gives an overview of the conditions to convert a laminar flow in the reactor (without stirring) into a turbulentregime.TheinfluenceofstirringontheReynoldsnumberissummarized inTable5.3usingequation5.5b.
Re tube =
dt u
(5.5a)
Re imp =
2 N dimp
(5.5b)
= density[kg m ]
3
= density[kg m ]
3
= dynamicviscosity[Pa s]
= dynamicviscosity[Pa s]
Table5.2
temperature (C)
RelationshipbetweenflowrateandReynoldsnumber*inthereactorcoil atfourtemperatures.
53 115 flowrate(mL.min1) 167 100 128 158 191 214 128 165 204 245 250 149 192 238 286
Table5.3
temperature (C)
RelationshipofstirringspeedandReynoldsnumber*inthereactorcoilat fourtemperatures.
60 195 stirrerspeed(rpm) 322 2191 2817 3488 4193 380 2546 3273 4054 4873 425 2854 3669 4544 5463
25 402 1306 50 517 1680 75 640 2080 100 769 2501 *ReusingEquation5.5b;dimp=1.68102m.
99
Chapter 5
Heattransfer The Nusselt number (Nu) is the ratio of the convective heat transfer and the conductive heat transfer. Equations 5.6 and 5.7 give the definition of the Nusselt number and an empirical dimensionless correlation from heat transfer in stirred tanks,respectively.52TheheatflowratefromtheWeflonblades(q)tothefluidin thereactorisestimatedbyEquation5.8.Solventswithalowlosstangentcanonly absorb microwave energy indirectly via the Weflon blades located at the stirrer shaftinthetubularreactor.Efficientheattransportpreventshightemperaturesof the heating surface (A = 95 cm2) of the Weflon blades (Equation 5.8). With the assumptionofdirectandfastabsorptionofmicrowavesbytheWeflonbladesthe temperatureisataconstantvalue.
In case of pxylene ( = 0.124 W.m1.K1) and a relatively low stirrer constant (cr = 0.36; blade stirrer, no baffles)52 the Nusselt number will range from 100 to 200 dependingontemperature(50100C)andstirrerspeed(200400rpm).Theheat transfer coefficient (h), consequently, is calculated to range from 367 till 725 W.m2.K1. Typically the surface of the blades can reach a temperature of 160 C whenanincomingstreamofpxyleneatroomtemperatureisheatedupto120C foranefficientmicrowaveinducedpowerinputof277Wintothesolvent.
Nu = cr Re 2 / 3 Pr1/ 3 Vi 0.14
Nu = Nusseltnumber c = stirrerconstant r Re = Reynoldsnumber Pr = Prandtlnumber Vi = viscosityconstant( 1)
(5.6)
Nu =
h L
(5.7)
q = h A ( T )ln
h =overallheattransfer
(5.8)
q =heatsuppliedtothesystem[ J s1 ]
coefficient[J s 1 m 2 K 1 ] A=surface[m 2 ] ( T )
( T )ln =
(T
out
=logarithmicaverage ln
temperaturedifference[K]
Physical constants are obtained from Handbook of Chemistry and Physics; 89th Ed.; 20082009 (onlineed.,Lide,D.R.);ViisratiooftheviscositiesatTbulkandTbl,whichareequal.
100
Chapter 5
101
Chapter 5
C S T R 1 pump CFR
C S T R 1
pump pump pump
CFR C S T R 2 C C S T R 2 C
F ilter
D ryer
D is tillation
Figure5.7
Basicsetup Figure 5.7 (left) shows a basic process flow diagram of the fume hood where the microwaveheated continuousflow reactor (CFR) is the central unit. The CFR is continuously operated with relatively short reaction times (< 30 min) or when optimalproductionratesforslowerreactionscanbeobtainedwithrelativelyshort reactiontimes.TheCFRisfedviaCSTR1,apremixerforreactantA(solid/liquid) andB(liquid).Gasescanbeintroduceddownstreamwithrespecttothepumpin combinationwithastaticmixer. Modification (1): Processes demanding longer reaction times can make use of a recycle loop (Figure 5.7 (right)). The outlet of CFR is recycled to CSTR1. The system is now in a batchloop operation mode. After completion a connection of CSTR1 to CSTR2 is made. Each CSTR can be replaced with inline static mixers. AfterreactionintheCFRanotherstream(C)canbeintroduced.Beyondthisstage the product/intermediate should be in a physical state to enable separation from theremainingreactionmixture,e.g.bycentrifugation,filtrationorextraction. CSTR2incombinationwithanantisolventstreamofC(andifnecessarycooling ofthevesselwall)simplifiesthissetupforprecipitationorasapossiblecontinuous crystallizer.
102
Chapter 5
Modification(2):ThereplacementofCSTR2bytwoparallelcolumns(containing functionalized resin) allows this setup to separate the product from the reaction mixture in one column while simultaneously regenerating the other column. A similarapproachcanbeadaptedforultrafiltrationorother(selective)membrane filtrations. The product can be isolated or continuously fed as a stream to e.g. a dryer. The otherstreamsmightbeconsideredforrecycling. Extension: More flexibility and production volume are introduced when parallel unitsareinterconnectedwhichupgradesthepossibilitiesofthebasicconcept. Two reactions demonstrate the feasibility of such a fume hood concept by fitting theprocessesasanexercise.OneprocessinvolvestheOacetylationofsalicylicacid which is the final step in the synthesis of aspirin, see Scheme 5.1; the other representstheLaurydoneprocess,seealsochapter2.7.
O OH COOH Ac2O O COOH
AcOH
Scheme5.1 SynthesisofAspirin;Oacetylationofsalicylicacid. Oacetylationofsalicylicacid TheprocessflowdiagraminFigure5.7(left)illustratesabasisforthecontinuous production of aspirin (Scheme 5.2). The aspirin production is an example for a largescalesynthesis.54,55SalicylicacidiscontinuouslychargedassolidA.StreamB representsan amount ofacetic anhydride (dissolved inaceticacid) per timeunit, equimolar (or more equivalents) with respect to stream A. The reaction can be catalyzedbyacidsorbases;thefirsttypeofcatalysisispreferredinviewofatom efficiency (otherwise twofold deprotonation is required). The catalyst can be homogeneous or heterogeneous i.e. an acidfunctionalized macroporous resin incorporatedintheCFR.Selfcatalysisbysalicylicacidisalsoapossibilityinthis case. After reaction, at a certain temperature and residence time (< 30 min), the mixture can be diluted with an antisolvent (water) or directly cooled. Aspirin crystallizesandtheslurrycanbeconcentratedorfilteredandadditionallywashed inacentrifuge(orbeltfilter).Subsequently,theproductisdried(e.g.bymicrowave heating)toremovetheremainingsolvents.Finally,watercanberemovedfromthe recyclestreamviapervaporation.56
103
Chapter 5
Anactualrunningprocess(35kt/a)fortheacetylationofsalicylicacidisshownin Figure5.8.57
N H
OH
- H2O OH
N H
O O
Scheme5.2 Reactionof(S)pyroglutamicacidwithndecanol. For the Laurydone process the introduction of a loop (Figure 5.7 (right)) is necessary due to a long reaction time (3 h). A direct start with an equimolar mixture(Scheme5.2)of(S)pyroglutamicacidandndecanolleadstopluggingat ambienttemperatures.Preferably,preheatingofndecanolwithagradualaddition of(S)pyroglutamicacidovercomesthisproblem.Thereafter,anoverflowtoCSTR 2isopenedtocollecttheenrichedstreamwithproduct.Byapplyingvacuumthe remaining water and ndecanol can be separated from the reaction mixture. The product crystallizes by cooling or by feedingantisolvent (water).Cooling canbe done in the filter unit to prevent clogging in CSTR2 which arises due to a high weightratioofproduct.Theremaining(S)pyroglutamicacidisremovedfromthe desiredesterproductbywashingwithwater.
104
Chapter 5
105
Chapter 5
stirrerspeed (rpm)
(a) (b)
flowrate (103kg.s1)
powerinput (W)
(a)
T (K)
(b)
efficiency (%)
(a) 34 47 39 66 53 44 55 58 50 46 (b) 52 73 69 71 67 65 60 75 74 63
1 0 195 0.75 200 53 81 2 0 195 0.75 400/600/800 3 0 195 1.61 200 34 53 4 0 195 1.61 400 57 101 5 0 195 1.61 600/800 6 0 195 2.26 200 34 37 7 0 195 2.26 400 55 70 8 0 195 2.26 600 68 101 9 0 195 2.26 800 10 0 195 2.82 200 23 25 11 0 195 2.82 400 48 62 12 0 195 2.82 600 62 92 13 0 195 2.82 800 77 105 cp(pxylene)=1709J.kg1.K1 Steadystatenotreached(below130C,whichwassetasoperationalmaximum)
106
Chapter 5
Table5.5
entry
RelationshipbetweenenergyefficiencyandoperationalconditionsofCFR fordemiwaterwith(a)Weflonand(b)Teflonstirrer.
flowrate (103kg.s1)
stirrerspeed (rpm)
(a) (b)
powerinput (W)
(a)
T (K)
(b)
efficiency (%)
(a) 85 87 72 84 88 81 70 84 85 81 70 84 88 89 (b) 61 67 92 84 91 84 81 87 89 91 84 91 95 92
1 195 195 1.04 200 39 28 2 195 195 1.04 400 80 62 3 195 195 1.04 600/800 4 195 195 1.92 200 18 23 5 195 195 1.92 400 42 42 6 195 195 1.92 600 66 68 7 195 195 1.92 800 81 84 8 195 195 2.78 200 12 14 9 195 195 2.78 400 29 30 10 195 195 2.78 600 44 46 11 195 195 2.78 800 56 63 12 195 195 3.33 200 10 12 13 195 195 3.33 400 24 26 14 195 195 3.33 600 38 41 15 195 195 3.33 800 51 53 cp(water)=4180J.kg1.K1 Steadystatenotreached(below110C,whichwassetasoperationalmaximum)
Table5.6
entry
RelationshipbetweenenergyefficiencyandoperationalconditionsofCFR forethyleneglycol(onlywithWeflonstirrer).
flowrate (103kg.s1) powerinput (W) T (K) efficiency* (%)
63 65 91 92 90 84 110 106 104 99 97 103 105 101
stirrerspeed (rpm)
1 195 1.08 200 49 2 195 1.08 400 100 3 195 1.08 600/800 4 195 2.44 200 31 5 195 2.44 400 63 6 195 2.44 600 92 7 195 2.44 800 115 8 195 2.73 200 30 9 195 2.73 400 58 10 195 2.73 600 85 11 195 2.73 800 108 12 195 3.28 200 22 13 195 3.28 400 47 14 195 3.28 600 72 15 195 3.28 800 92 cp(ethyleneglycol)=2394J.kg1.K1; Steadystatenotreached(below140C,whichwassetasoperationalmaximum) *Theerrormarginofthesemeasuredvaluesisintherangeof510%
107
Chapter 5
Table5.7
entry 1 2 3 4 5 6 7 8 9 10 11 12 13
Relationship of flow rate and stirrer speed with the number of equally sizedtanksinseries.
flowrate (mL.min1) 191 191 191 191 191 191 214 214 214 214 232 232 232 stirrerspeed (rpm) 0 140 215 285 327 380 0 140 215 327 0 140 215
2.0
reactorsinseries (number)
3.0 4.9 5.3 2.4 1.8 2.3 3.7 3.6 2.5 4.2 3.5 3.9 2.9
1,0 0,8 0,6
20
1.5
E(t)
1.0
1 2
a.u.
(a)
0.5
(b)
0.0 0 1 2
t/
1,0 1,5 2,0
Figure5.9
(left) Exitage function E(t) for the FlowSynth with Weflon impeller blades(seeFigure5.6):(a)typicalplotwithsmalldisturbance[Table5.7 entry 2]. (b) large disturbance in tubular reactor leading to enhanced axial mixing [Table 5.7 entry 8]; normalized at the intensity (a.u. arbitrary units). (right) Theoretical residence time distribution with an increasingnumberoftanks(n=1,2,3,5and20).
108
Chapter 5
7 8 9 10 11 12 13 14 15
16 17 18 19 20 21
109
Chapter 5
22 23 24 25 26 27 28 29 30 31 32 33 34
Roberts,B.A.;Strauss,C.R.Acc.Chem.Res.2005,38,653. Cablewski,T.;Faux,A.F.;Strauss,C.R.J.Org.Chem.1994,59,3408. Nchter, M.; Ondruschka, B.; Weiss, D.; Beckert, R.; Bonrath, W.; Gum, A. Chem. Eng.Technol.2005,28,871. Nchter, M.; Mller, U.; Ondruschka, B.; Tied, A.; Lautenschlger, W. Chem. Ing. Tech.2002,74,910. Will,H.;Scholz,P.;Ondruschka,B.;Burckhardt,W.Chem.Ing.Tech.2002,74,1254. Baxendale, I. R.; Hayward, J. J.; Ley, S. V. Comb. Chem. High Throughput Screen. 2007,10,802. Saaby,S.;Baxendale,I.R.;Ley,S.V.Org.Biomol.Chem.2005,3,3365. Baxendale,I.R.;GriffithsJones,M.;Ley,S.V.;Tranmer,G.KChem.Eur.J.2006,12, 4407. Khadilkar,B.M.;Madyar,V.R.Org.ProcessRes.Dev.2001,5,452. Kabza,K.G.;Capados,B.R.;Gestwicki,J.E.;McGrath,J.L. J.Org.Chem. 2000, 65, 1210. Shieh,W.C.;Dell,S.;Repi,O.TetrahedronLett.2002,43,5607. Baxendale, I. R.; Hayward, J. J.; Ley, S. V. Comb. Chem. High Throughput Screen. 2007,10,802. Technically,slowreactionscanbehandledinacontinuoussetup.However,incase ofaslowreactioninamultistepsynthesis,batchwiseoperationispreferred.With the combination of microwave heating and continuous operation relatively fast reactionsarepreferred,duetohighoperationalcostsoftheequipment. Bonaccorsi,L.;Proverbio,E.MicroporousMesoporousMater.2008,112,481. PicturesuppliedbymanufacturerSairem,France(J.P.Bernard). Malone,R.J.Chem.Eng.Progress1980,3. Vaucher,S.;Nicula,R.Chem.Today,2008,26,38. Stankiewicz,A.Chem.Eng.Res.Des.2006,84,511. Gronnow, M. J.; White, R. J.; Clark, J. H.; Macquarrie, D. J. Org. Process Res. Dev. 2005,9,516.
EnergyTipsSteam,2006,sheet#15.
35 36 37 38 39 40 41 42 43 44
Graus,W.H.J.;Voogt,M.;Worrell,E.EnergyPolicy,2007,35,3936. Nchter,M.;Ondruschka,B.;Bonrath,W.;Gum,A.GreenChem.2004,6,128. Assumingthatelectricalenergycosts0.15perkWhandnaturalgas0.65perm3 (energycontent33MJperm3)fortheprivatesectorand0.27perm3forindustry. ThecoststogenerateoneMJofsteamonanindustrialscaleis0.011,foraprivate boiler 0.046 and for one MJ of microwave heat with the average efficiency the costsare0.088,includingtheindicatedefficiencies. Hoogenboom,R.;Wilms,T.F.A.;Schubert,U.S.PolymerPrepints,2008,49,930.
45
110
Chapter 5
46 47 48 49 50 51 52 53 54 55 56 57
Koch,T.A.;Krause,K.R.;Mehdizadeh,M.Proc.SafetyProgress,1997,16,23. Wan, J. K. S.; Koch, T. A.; Preparation of hydrogen cyanide; DuPont, Patent number US5393393,1995. Fogler, H. S.; Elements of chemical reaction engineering, 3th Ed., Prentice Hall PTR, NewJersey,1999. Macmullin,R.B.;Weber,Jr.M.Trans.Am.Inst.Chem.Eng.1935,31,409. Danckwerts,P.V.Chem.Eng.Sci.1953,2,1. Atherton, J. H.; Carpenter, K. J.; Process Development: Physicochemical concepts, OxfordUniversityPress,Oxford,1999. Zlokarnik, M.; Stirring; (chapter 7) Intensification of heat transfer by stirring, Wiley VCH,Weinheim,2001. Realff,M.J.;Shah,N.;Pantelides,C.C.Comp.&Chem.Eng.1996,20,869. Drevina,V.M.;Nesterov,V.M.Khim.Farm.Zh.1976,10,1546. Mcketta,J.J.;Encyclopediaofchemicalprocessinganddesign,Dekker,Basel,1976,4,28. Fontalvo, J.; PhD Thesis, Eindhoven University of Technology, The Netherlands, 2006. Moulijn, J. A.; Makkee, M.; Diepen Van, A.; Chemical Process Technology, Wiley, Chichester,2001,pag.314.
111
Chapter 6
Chapter 6
Abstract
Heterogeneous reactions on the one hand represent the majority of fine chemical processes and on the other may give rise to potentially significant microwave effects. However, exactly this reaction type requires the most effort and adaptation under continuous operation. Nevertheless, a series of four reactions has been studied in an adapted, commercially available continuous-flow reactor. The (homogeneous) aspirin synthesis and a (heterogeneous) biocatalyzed, enantioselective transesterification successfully demonstrated a close correlation in reaction rate between batch-performed and continuous processing. The Laurydone process (47 wt % substrate) ran in a more diluted (10 wt % substrate) condition with a similar reaction rate. However, the removal of the formed water remained a limiting factor in the final stage of the conversion. Although the (heterogeneous) racemization of N-acetylindoline-2-carboxylic acid retained its microwave effect under continuous operation, plugging has to be dealt with. Technological solutions are demanded to improve the current setup, while an increase of the dimensions thereof may also contribute to an even better feasibility of the fume hood concept. In general, the results described in this chapter demonstrate that microwaveassisted, heterogeneous batch reactions are easily translated to microwave-assisted flow processing.
.
114
Chapter 6
6.1 Introduction
In literature the application of the FlowSynth developed by Milestone to accommodate organic processes has been reported. Typical example are the esterificationofcarboxylicacidswithdimethylcarbonate13ortheesterificationof acetic acid with butanol.4 More examples in organic chemistry can be found in a reviewofGlasnovandKappe.5InparticulartheWeflonbladescombinedwiththe helicalcoilenableefficientheattransferandmasstransportinthetubularreactor that is positioned in the microwave oven, see Figure 5.6 in the previous chapter. Alsohandlingofheterogeneousreactionsisfeasiblewiththissetup. From the results described in the chapters 2, 3 and 4 it was concluded that beneficial microwave effects were only observed when heterogeneity plays a crucial role in the ratedetermining step. All of the heterogeneous reactions reported in the chapters 2, 3 and 4 were originally performed batchwise on a relativelysmallscale(<20mL)intheMilestoneMicroSynth.Onlyafewexamples were selected from our study to assess the scalability of these heterogeneous reactionsintheFlowSynth. Inchapter5thecharacteristicsofthecontinuousflowreactorFlowSynthhavebeen described. The initial experimental results discussed in chapter 5 looked very promising.However,bringingthecontinuousflowreactortoanextlevelcloserto commercial scale e.g. larger temperature differences, higher solid content and lower solubilities, have put forward new challenges (e.g. solid and slurry handling). Consequently, the proof of the pudding is in the eating and each studiedreactionwillbeseparatelydiscussedinthenextsections.
115
Chapter 6
O +
HO + H O
Scheme6.1
Esterificationof(R,S)1phenylethanolwithvinylacetateinanenantio selectivereactionusingNovozym435.
The continuousflow experiment (0.4 mol) has been scaled up by a factor of 100 with respect to the batch mode reaction (4 mmol). The reaction mixture without biocatalystwaspremixedina1Lroundbottomedflaskandmagneticallystirred. Afterashortperiodthishomogeneousmixturewaspumpedthroughtheloop.The reaction temperature in the top of the tubular reactor (T1) was set at 70 C. The microwavepowerwasautomaticallyadjustedtomaintainthetemperatureatT1 of thissetpoint,seeFigure6.1.
T2 C R2 T1 R1
Figure6.1
P Batchloop reactor setup for the esterification of (R,S)1phenylethanol with vinyl acetate: R1 = microwaveheated tubular reactor; R2 = round bottomedflaskwithmagneticstirrer;P=(membrane)pump;C=cooler withArchimedesscrew;T1=temperaturesensorinsidetheupperpartof the tubular reactor and T2 = temperature sensor just downstream the cooler.
Theesterificationwasperformedtwice(experimentsAandB)intheFlowSynth.In experiment A the reaction mixture should have been distributed evenly over the batchloop setup. In experiment B the catalyst is retained on purpose in the microwaveheatedtubularreactor(R1). ExperimentA TheflowratehasbeenmeasuredbeforeandaftertheadditionofNovozym435.In both cases the flow rate was 175 mL.min1. Once the reaction setup reached a steady state (in which the temperature was constant in time) Novozym 435 was chargedtoflaskR2.DuringthereactionaliquotsweretakenjustafterpointT2,but beforevesselR2(seeFigure6.1).ThetimeconversionhistoryisplottedinFigure
116
Chapter 6
6.2.ItisnoteworthytomentionthattheamountofNovozym435appearedtobe higherinsidethereactorthaninvesselR2 oroutsideR1.Theliquidvelocityinthe tubing (tub = 0.15 m.s1) was considerably higher compared to that in the tubular reactor (react = 0.012 m.s1), which might explain settling of Novozym 435 in the columnR1.ThenarrowexitafterT2canalsoleadtoaholduporpluggingofthe spherical particles of Novozym 435. The temperature in vessel R2 has been monitoredmanuallyandneverexceeded30C.
50
(a) Batch
Conversion (%)
(Exp. A) Flow
25
Time (min)
Figure6.2
Esterification of (R,S)1phenylethanol with vinyl acetate in toluene at 70CcatalyzedbyNovozym435in(a)anoilbathheatedbatchreactor (4 mmol scale); (R)1phenylethyl acetate ( ) and (Exp. A) in a microwaveheatedbatchloopreactor(0.4molscale)at=0.175L.min1; (R)1phenylethylacetate( ).
Thetotalreactionvolumewas0.940Landthevolumeofthetubularreactorwas 0.180L.Withaflowrateof0.175L.min1andatotalreactiontimeof173min,the numberofcycleswascalculatedtobe168withanaverageresidencetimeofalittle more than one minute per cycle in column R1. Subsequently, for this batchloop reactor experiment only 19 % of the reaction mixture was exposed to the temperature of 70 C. In the batch experiment, however, 100 % of the reaction mixture was subjected to a temperature of 70 C. The uneven distribution of Novozym 435 over the complete reaction mixture hampers a quantitative comparisonofthereactionratesforthermalandmicrowaveheating. Note that comparison of the batch experiments with respect to both heating methodsasdescribedinchapter2demonstratedtheabsenceofarateenhancement effectbymicrowaveheating.
VR1/Vtot=0.19(fractionalresidencetime)
117
Chapter 6
ExperimentB The procedure applied in Exp. A was repeated for the esterification in the continuousflow reactor in a batchloop mode. Now, Novozym 435 was charged directlyintothetubularreactorandthereactorwasputtogethertooperateunder flowconditionswithcompleteretentionoftheNovozym435insidethemicrowave cavity.Thereactorcontentwasnotstirred. During the reaction aliquots were taken after point T2, but before vessel R2. The resultsareplottedinFigure6.3.
50
(a') Batch
Conversion (%)
(Exp. B) Flow
25
Time (min)
Figure6.3
Esterification of (R,S)1phenylethanol (4 mmol) with vinyl acetate in toluene at 70 C catalyzed by Novozym 435 in (a) an oil bathheated batch reactor; (R)1phenylethyl acetate ( ) and (Exp. B) in a microwaveheated batchloop reactor (0.4 mol) with catalyst charged in thetubularreactorat=0.12L.min1;(R)1phenylethylacetate( ).
The timeconversion history in Figure 6.3 is comparable with that of Figure 6.2. This observation confirms that the majority of the solid particles in the experimental setup for experiment A could have been retained in the tubular reactor.Incontrasttothebatchexperimenttheconcentrationofthebiocatalysthas beenlocallyincreasedwithafactorof5.2.Thetotalmeanresidencetimeperpass hasdecreasedwiththesamefactor.Thesetwochangescancelouteachother. The remaining difference in reaction rate is partially related to either a small temperature gradient in the microwaveheated tubular reactor or to some mass transferlimitationsinsidethereactor(withoutstirring).Withoutstirring(andwith alaminarflow)theregimeof(intraparticle)masstransportlimitationcanchange, asaresultofachangeoftheglobalreactionorder(seeFigure6.4(left)).6Figure6.4 (right) illustrates the effect of transport limitation to or from the particles on the timeconversion.7
118
Chapter 6
1,0
Degree of conversion
Kinetic control
0,8
Kinetic control
CA
Mass transfer control
dp
1 3
0,6
0,4
0,2
0,0 0,0
0,2
0,4
0,6
0,8
1,0
Distance
Dimensionless time
Figure6.4
(left) Concentration profile (CA) around and inside of a catalyst particle (with shell catalyst such as Novozym 435) with strong mass transfer limitation(3)andwithintrinsickinetics(1).(right)Dimensionlesstime conversionprofilesofafirstorderreactionbasedontheconcentrationofa solid(partiallydissolved)andtheratedeterminingregime(13).7
Ingeneral,MichaelisMentenkineticsisapplicabletothissystem,meaningthatfor the applied substrate concentration pseudo firstorder kinetics can be used to describetheconversionwithtime.Inourcasethesubstrateconcentrationappeared nottolimittheactivityoftheenzyme.8 This example of a biocatalytic esterification exemplifies a direct correlation between batch and continuous operation. Approximately the same final conversionsareachievedforcomparablereactiontimes.
Scheme6.2 Racemizationof(S)Nacetylindoline2carboxylicacid. Changes in reaction conditions such as temperature or cosolvent influence the performance of the batchloop reactor (Figure 6.5), including factors such as plugging.
119
Chapter 6
T2 C R2 P T1 R1
Figure6.5
Batchloop reactor setup for the racemization of (S)Nacetylindoline2 carboxylic acid: R1 = microwaveheated tubular reactor; R2 = double walled reactor vessel with mechanical pitchedblade impeller; P = (membrane)pump;C=coolerwithArchimedesscrew;T1 =temperature sensorinsidetheupperpartofthetubularreactorandT2 =temperature sensorjustdownstreamthecooler.
Two racemization experiments under different conditions were performed with (S)Nacetylindoline2carboxylicacidassubstrate,seeTable6.1.Forbothreaction conditions a microwave effect was expected based on already described experimentalresults,seechapter3. Table6.1 Racemization of (S)Nacetylindoline2carboxylic acid performed at different reaction conditions in the continuousflow reactor in a batch loopmode.
entry 1 parameters
(S)Nacetylindoline2carboxylicacid
experimentA
experimentB
178g(S) 126g(S) 0.87mol 0.62mol 2 pxylene 679g 650g 3 aceticacid * 72g 4 aceticanhydride 22.3g 14.8g 0.22mol 0.14mol 5 wt%substrate 21 15 6 wt%cosolvent(aceticacid) * 10 7 flowrate 216mL.min1 222mL.min1 8 totalreactiontime 263min 182min 9 interruptionby: plugging plugging *smallamountofwaterpresent,whichproducesaceticacidfromaceticanhydride(<2wt%).
Although each run was prematurely aborted due to plugging, sampling allowed comparisonwiththebatchexperiments.Themaindifferencebetweenexperiment A and B was the amount of cosolvent. In experiment A the racemization was
120
Chapter 6
performedwithoutcosolvent.InexperimentBaceticacidwasaddedtodecrease the heterogeneity, and therefore to reduce the microwave effect and the risks of plugging. ExperimentA In this experiment the racemization was performed without cosolvent. After additionofthecatalyst(aceticanhydride)afastdecreaseoftheenantiomericexcess was observed in the smallscale batch experiments (Figures 6.6(a) and (b)) in the first period (< 60 min). Subsequently, the racemization during the microwave heated experiment continued, but for conventional heating inhibition was temporarilyobserved.After120minthereactionrateofracemizationincreasedas thesolubilityofthesubstraterose.Forthesmallscalebatchreactionconventional and microwave heating gave rise to different reaction rates. The flow experiment behavedsimilarlytothemicrowaveheatedbatchexperimentasillustratedbythe absence of an inhibition period after the addition of the catalyst. Reaction rates basedonenantiomericexcesschangesfrom90to70%eeareshowninTable6.2for boththebatchexperimentsandthecontinuousflowexperimentA.
100
75
(exp. A) Flow
ee (%)
50
(b) Batch / MW
(a) Batch / CH
25
Time (min)
Figure6.6
RacemizationofNacetylindoline2carboxylicacidat130Cin(exp.A) continuousflow experiment in a batchloop mode, (a) conventionally heated batch experiment and (b) microwaveheated batch experiment. Experimentswereperformedundercomparableconditions,seeTable6.1.
Ideally, experiment A should have gone to completion to demonstrate that complete racemization can take place in the continuousflow reactor. However, pluggingurgedpreliminaryinterruptionofexperimentA.Experiencefromearlier racemizationexperimentsasdescribedinchapter3taughtusthatinhibitionisonly observed during the initial period, after which the reaction goes to completion. This observation combined with the relative reaction rates shows that the
121
Chapter 6
conversions in the continuousflow reactor closely resemble those of the microwaveheated smallscale batch experiment. In this case, in contrast to the biocatalyticreactiondescribedinsection6.2,racemizationcaninprinciplepartially continueoutsidethetubularreactor,beforeatemperatureisreachedatwhichthe rate of racemization is negligible. However, continuation of the racemization outside the microwaveheated tubular reactor R1 is insufficient to distinguish it fromthemeasurementerror. Table6.2 Reactionrates(intervalbetween90and70%ee)oftheracemizationof (S)Nacetylindoline2carboxylic acid under conventional and under microwaveheatingconditions.
entry conditions overallreactionrate [ee.min1] fractionalresidence time* correctedreactionrate [ee.min1]
1 batch(a);CH 0.20 1 0.20 2 batch(b);MW 0.34 1 0.34 3 experimentA 0.08 0.21 0.36 * fractionalresidencetimeinthereactorcomparedtothetotalreactiontime[basedontheratioof thevolumeofthemicrowaveheatedtubularreactor(R1,180mL)pertotalreactionvolume(850
mL); overallrate = VR1 Vtot rate(MW,R1 ) ]
ExperimentB InexperimentBaceticacidwasaddedtoreducetheheterogeneityofthesystem. Afteradditionofthecatalyst(aceticanhydride)sometechnicalproblemsoccurred. Thereactionmixturewastemporarilycollectedinonevesselandatatemperature below 20 C. The system was restarted and two large aliquots were taken to perform two experiments batchwise by oilbath heating at 60 C and 122 C, see Figure 6.7. As a result of these large aliquots, the total reaction volume in the microwaveheated batchloop system was less than in experiment A. As a consequencethe fractional residence time (= real reaction time) in tubular reactor R1(Figure6.5)washigherthaninexperimentA,seeTable6.3[entry4].Experiment Bwasexecutedat127Cinthecontinuousflowreactor.Thetemperatureinvessel R2(Figure6.5)wasbelow60C(manuallymeasured).Thelowreactionrateat60 C(Figure6.7)allowsneglectingracemizationinvesselR2.Theinhibitionperiodas observed for the conventionally heated racemization in experiment A was not observed(seeFigure6.6),duetothedelayofsampling.Reactionratesbasedonan enantiomericexcesschangefrom86to34%ee(areabetweendottedlinesinFigure 6.6)werecomparedbetweenthreebatchexperiments(c,d,e)andthecontinuous flowexperimentB(Table6.3).
122
Chapter 6
Table6.3
entry 1 2 3 4 conditions
* fractionalresidencetimeinthereactorcomparedtothetotalreactiontime[basedontheratioof the volume of the microwaveheated tubular reactor (R1, 180 mL) per total reaction volume*** (785mL); overallrate = VR1 Vtot rate(MW,R1 ) . ** reactionrateobservedundermicrowaveheatingandcomparableconditions. *** Vtot = Vpremixed Vsampling ;785mL=850mL65mL.
80
(c) Batch / CH / 60 C
60
ee (%)
40
20
0 0 30
Time (min)
Figure6.7
Racemization of (S)Nacetylindoline2carboxylic acid in a continuous flow experiment in batchloop mode (Exp. B) compared to (c) conventionallyheatedrunat60C,(d)conventionallyheatedrunat122 Cbatchexperiment,and(e)microwaveheatedrunat129C,allunder comparableconditions.
Experiment B allows us to make a fair comparison between microwave and conventional heating of the racemization efficiency during a longer period (with respect to Exp. A) on the basis of changes of the enantiomeric excess. Again the estimated conversion of the microwaveheated experiment proved to be in line withexpectations. The racemization was effective during a longer period and in a larger volume outsidethemicrowaveheatedtubularreactorR1,duetothedeliberateabsenceof coolingpowerincoolerC(Figure6.5)ascomparedtoexperimentA. Insummary,basedontheexperimentsdiscussedinthissectionitcanbeconcluded that the microwave effect described in chapter 3 has been demonstrated in the
123
Chapter 6
continuousflow reactor with respect to larger scale applications. Thus for the racemization microwave heating is preferable to conventional heating for either batchorcontinuousoperation.
N H
OH
-H2O OH
N H
O O
Scheme6.3 Esterificationof(S)pyroglutamicacidwithndecanol. All our batch reactions of the Laurydone synthesis have been performed with a high concentration of substrate ((S)pyroglutamic acid) in ndecanol. Such a solid weight percentage (47 wt%) appeared to be impractical to handle in the continuousflow reactor. Reduction of the acid weight percentage to 10 wt% guaranteed a smooth flow for the initially heterogeneous reaction mixture, see Scheme 6.3. Again two different conditions have been applied for this type of esterification (Table 6.4). It is important to point out that the membrane pump in experiment A was replaced by a better functioning gear pump in experiment B. Comparedtoamembranepumpagearpumpismoresuitableforhandlingmore viscousmixtures,likendecanol.TheflowschemesofthesetupforexperimentsA andBareschematicallydepictedinFigures6.1and6.5,respectively.
124
Chapter 6
Table6.4
Esterification of (S)pyroglutamic acid with ndecanol under different reaction conditions (Exp. A and Exp. B) in the continuousflow reactor operatedinabatchloopmode.
experimentA
69.1g 0.54mol 622g 3.93mol 10 115mL.min1 membrane 140min medium/bottom plugging
entry 1 2 3 4 5 6 7 8
parameters
(S)pyroglutamicacid ndecanol wt%substrate flowrate pump totalreactiontime mixingspeedR2/suctionlinefromR2 interruption:
experimentB
69.0g 0.53mol 622g 3.93mol 10 343mL.min1 gear 178min low/top manual
ExperimentA AfterchargingvesselR2withreactantsandstirringwithamechanicalimpeller,the resulting solid/liquid mixture was pumped through the tubular reactor (without heating). This procedure enabled to run the flow system for many hours without any complications. Subsequently, heating was applied to the tubular reactor R1 (Figure6.5,150C).Coolingagentof8CwassuppliedtothejacketofcoolerC and cooling of vessel R2 was started by pumping a coolant of 12 C through the jacket.Shortlyafterstartinguptheheatingandcooling,pluggingoccurredinthe cooler C. This prevented any flow. Also stirring of the coil in the tubular continuousflowreactorwasnotpossible.Adepositofsolidmaterialonthewalls ofcoolerCwasnoticed. IntherestartedexperimentcoolingwasonlyappliedtovesselR2,thecoolerCwas notfilledandthejacketwasopentoair.Underthesecoolingconditionsplugging occurred again and this could be attributed to the pump. Due to a larger temperaturedifferenceoverthepumpheadsectionsettlingoccurred.Thissettling hamperedtheflowthroughthepumpandultimatelytheflowstopped.Therefore, a fair comparison between the continuousflow microwave reactor and the conventionallyheatedexperimentscouldnotyetbemade. ExperimentB Experiment A was repeated adopting the previous procedure. Instead of a membranepumpagearpumpwasengagedinexperimentB.Introductionoflarge particles relative to the tubediameter led in this case also to plugging in the suctionline.Duringthefirstperiodthispluggingproblemcouldbecircumvented by slowly stirring the contents of vessel R2. Suction occurred via the top. In this
125
Chapter 6
way the reaction mixture was always saturated in substrate, although not completelyhomogeneouslymixedwithrespecttothetotalcomposition. TheLaurydoneprocesscanbedividedintothreestages(seeFigure6.8):thestart up stage I, stage II of predominant esterification and stage III to remove the reaction product (water). Stage I is not appropriate for making a comparison betweenconventionalandmicrowaveheating.
100
80
Conversion (%)
60
III II I
40
20
(Exp. B) Flow
0 0 60 120 180
Time (min)
Figure6.8
Esterificationof(S)pyroglutamicacid(10wt%)withndecanolin(Exp. B)continuousflowexperimentBinbatchloopmode,(a)conventionally heated batch experiment at 150 C and (b) conventionally heated batch experimentat130C.
Microwave heating is much faster than oilbath heating. For oilbath heating it takesmorethan10mintoreachthedesiredtemperature.Fortheflowexperiments heating is, however, almost instantaneous. In stage II the esterification is predominant at a constant temperature, making this stage suitable for a comparison.StageIIIisdominatedbythepresenceofwaterandthelattersrateof removalwhichcontrolstheequilibriumintheesterification.Alsothedifferencein setup plays a role in the removal of water, which appeared to be important (chapter2.6). Forthebatchexperimentat150Ctheconversionreachedaplateauof60%after 90minutes.Thetimeconversionhistoryresultingfromthemicrowaveheatedrun in the continuousflow reactor operating in a batchloop mode approaches this plateauapproximatelylinearly.Thebatchloopsetupclearlyhasanotherinfluence on the removal of water than the batch operation. In the Laurydone synthesis comparingconversionratesofbatchandflowprocessingismorecomplicatedthan in the cases of biocatalysis and racemization. A much higher temperature is measuredoutsidethetubularreactorR1thanintheothercases.Asaconsequence,
126
Chapter 6
direct correction with the fractional residence time is insufficient due to the continuationoftheesterificationoutsidethetubularreactorR1. Therefore, to makean estimation for the rate inR1,a correctionhas beenapplied withacombinationofthefractionalresidencetimeandthereactionrateoutsidethe tubularreactorR1basedonbatch(b),seeentry2,Table6.5. Table6.5 Conversion rates (interval of 13 and 42 % conversion) of the esterification of (S)pyroglutamic acid under different reaction conditions.
entry 1 2 3 conditions batch(a);150C batch(b);130C experimentB withcorrection overallreactionrate [conv.min1] 1.14 0.51 0.70 fractionalresidence time* 1 1 0.26 relativereactionrate [conv.min1] 1.14 0.51 2.68 1.27**
After correction of the overall reaction rates, Table 6.5 demonstrates almost identicalbehaviorforbatchorbatchloopoperationbymicrowaveheating.Thisis, however,onlyvalidforstageIIintheLaurydoneprocess.ThestagesIandIIIare controlledbytheheatingrateandtheexperimentalsetup,respectively.
AcOH
Scheme6.4 Synthesisofaspirin;Oacetylationofsalicylicacid.
127
Chapter 6
The synthesis of aspirin proceeds by a relatively straightforward procedure and allowsinvestigationofanintrinsicallyfasthomogeneousreactioninapotentially singlepassoperation(Scheme6.4).So,theaspirinsynthesisintheFlowSynthisa realcontinuousprocess. Starting with a recipe based on a high concentration (25 wt%; 2.3 molL1) of salicylicacidinaceticacid,atwhichthesubstratewascompletelydissolvedat50 C, instantaneously led to plugging. Lowering the concentration of salicylic acid improved the handling in a continuousflow mode, although reaction rates were negatively affected. Simultaneously, a switch from a gear pump to a membrane pumpwasnecessarytoensureastableflow(seeFigure6.1).Afterthefirst150mL of the product stream (R1 = 180 mL) operation was at steady state. So residence time and conversion can be correlated at multiple points (Figure 6.9). The mean residence time (390 39 s) was determined by measuring the volume of the product stream during 3 min process time. The measurement was repeated 20 times.Alsothesalicylicacidconversion(574%)wasmeasuredinthesesamples, seeFigure6.9. The actual product stream shows a delay which is in accordance with the mean residencetime(thesumofvolumeoftubingandcoolingunitisapproximatelyalso 180mL).Therelativelylongerresidencetimesoffractions9and13arewitnessed asrelativelyhigherconversionsatfractions11and15,respectively.
100 Conversion Residence time 14
12.3
80
12 8
60
40
20
0 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Fraction
Figure6.9
Conversion profile of the aspirin synthesis in acetic acid (1 M) with 2 molar equivalents of acetic anhydride at 120 C in a continuousflow experiment.Volumesoftheproductstreamduring3minprocesstime.
128
Conversion (%)
Chapter 6
Batchexperimentationusingmicrowaveorconventionalheatingwithanidentical reactiontime(390s)resultedinconversionsof66%and64%,respectively. WiththeassumptionthattheOacetylationofsalicylicacidobeysapproxiametely firstorderkineticsinthefirstpartofthereaction,thereactionconstant(k)canbe estimated (see Equation 6.1). Together with the model, representing the tubular reactorR1asaseriesofCSTRs,theconversion(xA)relatedtoresidencetimecanbe calculatedforthecontinuousflowprocess(Equation6.1).Thetotalvolume(Vtotal) of the tubular reactor R1 is 180 mL and the flow rate () is 0.46 mL.s1. If the number(n)ofequallysizedtanksinseriesis3or2thenthetheoreticalconversion wouldbe60or57%,respectively;whichperfectlyagreeswiththeexperimentally observedconversion(57%). (6.1) 1 n xA = 1
1 + k total n
k=2.7103s1[120C]
The translation from batch to continuous processing appears to be a relatively simple procedure, although more knowledge (e.g. on the internal temperature of tubularreactorR1)isrequiredtoallowareasonableprediction. However, a longer residence time is not possible with the current setup. The reaction rate can be improved with a catalyst (see section 5.6). Complete conversion demands for a long mean residence time. Nevertheless, batchloop operationallowshighconversionsinashorttimewithpropertemperaturecontrol.
6.6 Discussion
Theenergyefficienciesoftheperformedreactionshavebeencalculatedinasimilar wayasreportedinchapter5.4.TheresultsarecollectedinTable6.6andillustrate highyieldsforthetransitionfrommicrowaveenergytoheatingforthefirstthree typesofexperiments. Table6.6 Efficiencyofconversionfrommicrowaveenergytoheat.
entry experiment 1 2 3 4 biocatalysis racemization laurydone aspirin description
exp.A;basedonphysicalconstantsoftoluene exp.A;basedonphysicalconstantsofpxylene exp.B;basedonphysicalconstantsofndecanol basedonphysicalconstantsofaceticacid
efficiency(%)
89 88 92 37
129
Chapter 6
The aspirin synthesis with an extremely low flow rate demonstrates poor energy efficiency.Fortheselowflowratesheatlossestothesurroundingsseemnottobe negligible. The lowest efficiencies in chapter 5 were also found with the lowest flowrates. AllcalculationsinTable6.7arebasedontheheatcapacityofthemajorcomponent (e.g.solvent)presentinthereactionmixture.
6.7 Conclusion
Aspects of scalingup microwaveassisted reactions have been investigated and discussedtobepossible,evenwithoutoptimization.Highproductionratescanbe obtainedforbatchwiseoperationbynumberingupthemicrowaveheatedreactor units.6Thesteptohomogenouscontinuousflowprocessinghasbeeninvestigated also without further process research involving a fast reaction (i.e. aspirin synthesis).4,10 The results described in this chapter show the success of a straightforward translationfromoperationinstirredvesselsonbenchscaletooperationinaloop system with a tubular microwaveheated reactor with respect to chemical performance in terms of conversion and selectivity. Furthermore, preservation of the microwave effect was observed for the heterogeneous racemization of (S)N acetylindoline2carboxylic acid. Conversely, handling and processing of heterogeneous reaction mixtures require more development work for the transfer from smallscale stirred batch reactors to a batchloop or even a continuous operation. Redesigning the present FlowSynth setup allows to cope with issues suchassolidandslurryhandling. Larger scale operation is expected to enable minimization of the obstacles (e.g. plugging)encounteredintheexperimentsreportedinthischapter. From the perspective of safety or selectivity, rapid cooling and heating are easily achievedwiththecurrentFlowSynthequipment. In summary, organic chemistry can be performed in a microwaveheated batch loop system or a continuously operated tubular reactor configuration. Safety aspects and an easy, direct translation to continuous processing are two positive elementsfortheFlowSynthsetup. Inspecificcasesapositivecontribution(i.e.microwaveeffect,seechapters3.3and 4.4)isobservedjustasinsmallscalebatchprocesses.However,largeacceleration of reactions is not expected by microwave heating under controlled reaction conditions.
130
Chapter 6
131
Chapter 6
T1
P-set T-set
T-set
T2
Pump Stirrer
Pressure
T1
T2
Pump
Pressure
Figure6.11
132
Chapter 6
Racemization (R,S)Nacetylindoline2carboxylic acid [exp. A]; A 2L roundbottomed flask was chargedwithdry(S)Nacetylindoline2carboxylicacid(178g,0.87mol,99%eeS)andp xylene(679g).Thereactionmixturewascirculatedthroughthecontinuousflowreactorof MilestoneequippedwithaWeflonstirrerviaamembranepumpwithaflowof216mL.min 1andheatedto130C(T1).Ontopofthereactorcoolingwasapplied.Subsequently,acetic anhydride(22.3g,0.22mol)wasaddedtotheflaskandthesettemperaturewasmaintained for a period of 263 min. The total reaction volume was 850 mL. Average power usage during reaction was calculatedto be407W.Duringthe reaction aliquots(~50 mg) were taken and quenched with water (25 L) to neutralize the anhydride. The samples were dissolvedinamixtureofformicacid(3mL),isopropanol(25mL)andnhexane(72mL) andanalyzedbyHPLC.Chiralcolumn(Daicel,ChiracelOD)withaneluentcomposedof formicacid(1v%),isopropanol(10v%)andnhexane(89v%). (R,S)Nacetylindoline2carboxylic acid [exp. B]; A 2L roundbottomed flask was chargedwithdry(S)Nacetylindoline2carboxylicacid(126g,0.62mol,99%eeS)andp xylene(650g).Thereactionmixturewascirculatedthroughthecontinuousflowreactorof MilestoneequippedwithaWeflonstirrerviaamembranepumpwithaflowof222mL.min 1 and heated to 130 C (T1). On top of the reactor no cooling was applied. Subsequently, acetic anhydride (14.8 g, 0.14 mol) was added to the flask and the set temperature was maintainedforaperiodof182min.Thetotalreactionvolumewas785mL.Averagepower usage during reaction was calculated to be 280 W, see Figure 6.12 for the reaction parameters.Duringthereactionaliquots(~50mg)weretakenandquenchedwithwater(25 L)toneutralizetheanhydride.Thesamplesweredissolvedinamixtureofformicacid(3 mL), isopropanol (25 mL) and nhexane (72 mL) and analyzed by HPLC. Chiral column (Daicel,ChiracelOD)withaneluentcomposedofformicacid(1v%),isopropanol(10v%) andnhexane(89v%).
133
Chapter 6
T-set P-set
T1 T2
T-set
Pressure
Laurydoneprocess [exp. A]. A doublewalled reactor was charged with (S)pyroglutamic acid (69.1 g, 0.54 mol) and ndecanol (622 g, 3.93 mol). The reaction mixture was circulated through the continuousflowreactorofMilestoneequippedwithaWeflonstirrerviaamembranepump withaflowof115mL.min1andheatedto150C(T1).Ontopofthereactornocoolingwas applied. Doublewalled reactor was thermostatic at 60C at the walls. Reaction time was 140min.Thetotalreactionvolumewas700mL.Averagepowerusageduringreactionwas calculatedtobe294W.Duringthereactionaliquots(~50mg)weretaken,quenchedwith dichloromethane(3mL)andmeasuredby 1HNMR(CD3OD,200MHz),typicalsignals (ppm)4.14(t,2H,CH2OCO,product),3.53(t,2H,CH2OH,ndecanol). [exp. B]. A doublewalled reactor was charged with (S)pyroglutamic acid (69.0 g, 0.53 mol) and ndecanol (622 g, 3.93 mol). The reaction mixture was circulated through the continuousflowreactorofMilestoneequippedwithaWeflonstirrerviaagearpumpwitha flow of 343 mL.min1 and heated to 150 C (T1). On top of the reactor no cooling was applied. Doublewalled reactor was not thermostatic at the walls. Reaction time was 178 min. The total reaction volume was 700 mL. Average power usage during reaction was calculated to be 444 W, see Figure 6.13 for the reaction parameters. During the reaction aliquots(~50mg)weretaken,quenchedwithdichloromethane(3mL)andmeasuredby1H NMR(CD3OD,200MHz),typicalsignals(ppm)4.14(t,2H,CH2OCO,product),3.53 (t,2H,CH2OH,ndecanol).
134
Chapter 6
T1
T2
T-set
Pressure
Power
Aspirinsynthesis OAcetylsalicylicacid.A2Lroundbottomedflaskwaschargedwithsalicylicacid(207 g,1.5mol),aceticacid(1228g,1.12L)andaceticanhydride(307g,3.0mol).After15min of mixing (almost complete dissolution) the reaction mixture was pumped through the continuousflowreactorofMilestoneequippedwithaWeflonstirrerviaamembranepump with a flow of 27 mL.min1and heated to120 C (T1).On top of thereactor cooling was applied.Residencetimewas6.5min.Averagepowerusageduringreactionwascalculated tobe253W.Duringtherunthereactionmixturewascollectedinfractions,aliquotswere taken, and measured by 1HNMR (CD3OD, 200 MHz), typical signals (ppm) 8.01 (d, 1H,CHCCOOH,aspirin),7.84(d,1H,CHCCOOH,salicylicacid).
135
Chapter 6
7 8 9 10
136
This thesis comprises many examples in which reaction performance with microwave heating was directly compared with that of conventional heating. This comparative study haslednotonlytoamuchbetterunderstandingofmicrowaveassistedchemistry,butalso to recognizing the importance of proper control of the reaction parameters (e.g. temperature). Microwaveirradiationgivesresultsidenticaltoconventionalheatingformostofthestudied reactions, provided that the temperature and stirring efficiency are controlled. Either the possibility of a high heating rate related to microwave heating or any variation in the temperature distribution, due to poor mixing, give rise to (uncontrolled) deviations. Rate enhancements by microwave heating are observed for certain heterogeneous reactions. Heterogeneity alone, however, is not sufficient to induce microwave effects. In particular, when the solubility of the substrate is very low and the absorbance of microwaves by the reactionmediumispoor,microwaveheatingappearstobesuperior.Onamicroscopicscale absorbanceofmicrowavesindirectlyresultsinahigherreactionrate.Rateenhancementsup toafactorof9havebeenregistered.Themagnitudeofthesemicrowaveeffectsdependson theheterogeneityofthesystem.Ifthesolubilityofthereactantinsuchareactionmixtureis increasedbyadditionofasolventorbyselectingabetterdissolvinganalogueofthesame class of compounds then the magnitude of the microwave effect decreases. When the reaction mixture becomes homogeneous the microwave effect has completely vanished. Therefore it is stated that, this rateaccelerating effect originates at the interphase of solid andliquid,duetoalocalhighertemperaturebyselectivemicrowaveabsorptionleadingtoa localincreaseofsolubilityofthesubstrateandanincreaseoftheglobalratecoefficient. The batchwise experiments were performed in a multimode apparatus which, from an energetic perspective, is generally less effective than a monomode microwave oven. An increasing powerdensity under smallersized experimental conditions apparently leads to higherreactionrates,duetotheinabilityofaccuratetemperaturecontrol(partiallybasedon theuseofexternalorslowsensors). In the field of automation of (parallel)synthesis for smallsized vials (a few mLs), microwaveheatinghasprovenitsvalueinthelaboratory.Basedonourobservationswith microwave heating further growth in highthroughput experimentation can be expected. Also for reactions performed at higher temperatures or pressures leading to large rate enhancements,apositivedevelopmentisexpectedwhichallowsthechemisttorunseveral chemicalreactionsperday.Anotherfieldofinterestforapplyingmicrowaveheatingisthe fastproductionoflargervolumesonapilotplantscale.
137
Importantly, flow chemistry, including microreactors, has aroused more interest in recent studies. Part of this development will be beneficial for the implementation of microwave heating in the production of larger quantities of chemical products. In our view, the best opportunitieslieincontinuousprocessingforscalingupmicrowaveassistedchemistry,asa consequence of the limited penetration depth of microwaves. Once the continuously operated microwaveassisted chemistry has been demonstrated to be successful, the laboratory scale operation simplifies the procedure to produce material on a kgscale by microwaveirradiation. The current FlowSynth as supplied by Milestone is a unique commercial device. For homogeneous, nonviscous reaction mixtures the setup of different components (including e.g.pump,tubing)canbeoperatedondemand.Theintroductionofsolidsormoreviscous solvents requires, however, some modifications to prevent clogging or fluctuating flow rates.Redesigningthetubularreactortowardsacharacteristicplugflowreactorprefersthe implementationofastaticmixer(radialmixing).Tocontroltheperformance,eithermultiple temperaturesensorsortheabsenceofatemperatureprofileoverthelengthofthereactorare preferred.Otherwiseabroadresidencetimedistributionrepresentingthecharacteristicsof acontinuousstirredtankreactor,resultinginauniformtemperatureautomaticallyimplies adifferentgeometrical(spherical)design.Increasingthevolumetosurfaceratio,byscaling upunderlimitedpenetrationdepthconditions,createsabetterperspectivetohandlesolids. Such a design reduces energy losses to the environment, thus preventing premature precipitation. The technical feasibility of microwave heating on an industrial scale in fine chemical applications has been demonstrated by the Laurydone process. The appearance of beneficial microwave effects in the presence of sparingly soluble substrates makes microwave heating even more promising if the added value is confirmed. From a more realisticperspective,theuseofmicrowaveequipmentandoperationalcostsassucharenot competitivecomparedtoconventionalheatingbysteamoroil.Compensationbysufficient rate enhancements, increased selectivity and process safety aspects is needed to overcome this hurdle. However, a growing interest in flow chemistry combined with further development of large magnetrons (i.e. providing microwave power on the spot) might induce successful implementation. Recent developments generating effective microwave heating, even in largescale batch reactors or drying equipment advertised by Pschner GMBH in Germany and an Ukranian institute (SSI NAS), are promising. Best chances for application are reactions requiring fast heating rates or processes which benefit from selective heating (e.g. two phases). High heating rates refer to kgscale productions in the earlystagesofproductdevelopment,wheretimeismoreimportantthancosts.In2009the company Cambrex Inc has already reported successes in producing pilotplant quantities (e.g.aPdcatalyzedcontinuousmicrowaveheatedprocess). Common for the research magnetrons is a frequency of 2.45 GHz. Further investigations have to provide an answer to the question which frequency is the best for largescale fine chemicaloperations.
138
Summary
Summary
Microwave heating in fine chemical applications: role of heterogeneity
Only since 1986, microwaveassisted chemistry has been taken up by the academic world. Sofar,microwavetechnologyismainlyusedtoefficientlysynthesizeleadcompoundsona gramscale.Thepharmaceuticalindustry,however,isresponsibleforthelargestpartofthe developmentofmicrowaveassistedchemistry,byproducingmorecompoundsinashorter time span with the aid of an automated microwaveheated setup (e.g. highthroughput experimentation).Primarily,microwaveirradiationheatsupreactionmixtures.
In chapter 1 an overview of the real and apparent successes of microwave technology is given.Inthesubsequentchaptersthefollowingmaintopicshavebeenaddressed.Thefirst partoftheresearchwasfocusedonthesuperiorvalueofmicrowaveheatingwithrespectto the conventional approach in chemical conversions. Its origin is, to a considerable extent, basedonthepublishedclaimsregardingrateacceleratingmicrowaveeffects.Asecondpart deals with the implementation of the gained knowledge in scalingup particular fine chemical processes. In a third part a conceptual design was elaborated in which the microwaveovenhasbeenincorporatedinafumehoodsetuptoproducelargerquantitiesof productonacommercialscale.
Inchapter2,aimedatreproducingreactionsfromliterature,aconsistentapproachhasbeen selected to compare microwave heating with conventional heating. This has revealed the truenatureoftheobservedmicrowaveeffects.Thebiocatalyzedreactionsweregivenextra attention in view of claimed synergetic effects, where besides reactivity also enantioselectivity and stability were considered. All of these biocatalyzed (trans)esterification studies with three types of enzymes (Novozym 435, Subtilisin, PPL) have demonstrated identical time conversion profiles for both heating methods. The cyclizationtopiperidinesandpiperazinesfromthecorrespondinghalidesneitherrevealed rateaccelerationswithmicrowaveheating.Duringtheseconversionstheformedhydrogen halidewasactivelyremovedbyapplyingthermalenergyorbyneutralizationwithbase.The oxidation of benzyl alcohol to benzaldehyde with iron trinitrate revealed to be a complicated process of redox reactions giving an identical result when subjected to microwave irradiation in contrast to literature claims. The same observation holds for the trialkylation of methyl gallate where the conversion rate was not influenced, despite the heterogeneity due to a large amount of solid reactant (K2CO3). Most illustrative was the microwave effect observed in the Laurydone process (esterification of pyroglutamic acid with ndecanol) revealing the importance of agitation to reproducibly compare heating experiments. Here, the originally observed microwave effect (factor 3) vanished by proper agitation. Although in this case, the use of microwave energy to heat the reaction was superfluous,thefeasibilityofscalinguptheLaurydoneprocesstoanindustriallyrelevant scale was demonstrated. For all mentioned reactions in chapter 2, temperature control provedtobecrucialtojudgewhetheranyfavourablemicrowaveeffectispresent.
139
Summary
In chapter 3 the racemization of two types of Nacetylated amino acids (Nacetyl phenylalanine and Nacetylindoline2carboxylic acid) was investigated. Solubility of the substrateinpxyleneispoorandtheaceticanhydridemediatedreactionisacceleratedupto afactor8bymicrowaveheating. Acomparableapproachhasbeenfollowedinchapter4tostudythebeneficialinfluenceof microwave irradiation on the monoaddition of a series of C6substituted (methyl, ethyl, isopropyl and phenyl) isocytosines to hexamethylenediisocyanate. The heterogeneous reactionmixturegivesalsorisetoamicrowaveenhancedreactionrateuptoafactorof9.A change in the substituent influences not only the reactivity, but also the solubility and by that the magnitude of the microwave effect. This accelerating effect finds it origin at the interphase of solid and liquid where local microwave absorption occurs. Locally higher temperatureinducesalocallyincreasedsolubilityofthesubstrateandsoanincreaseofthe globalratecoefficient.Themicrowaveeffectcompletelyvanisheswhenasolventwasadded (chapter 3) or decreases when the solubility of the substrate improves (chapter 4). In summary, heterogeneous reactions may display an added value in the conversion rate by microwaveheating. In chapter 5 the necessity is described to switch from batchwise to continuous operation whenapplyingmicrowaveheating,duetothelimitedpenetrationdepthofmicrowaves.To broadentheoperationalwindowofaflowreactorwithrespecttomoderatereactionrates,a recycleloopsystemhasbeenintroduced.Althoughtheincorporationofamicrowaveoven implieshighercostsinacquisitionandinoperationcomparedtoconventionalheating,the usageisadvantageousregardingpracticalperformanceandsafety.TheFlowSynthreactor ofMilestonecorrespondstoaseriesofthreecontinuousstirredtankreactorsonaverage. In chapter 6 four examples are discussed to demonstrate the actual performance of the FlowSynth under different operational conditions. An Oacetylation of salicylic acid (aspirinsynthesis) was conducted in a single pass under homogeneous conditions. Three heterogeneous reactions (a biocatalyzed esterification, a neat esterification and a racemization) are performed applying a recycle loop (i.e. batchloop operation). Although plugging appeared to be an important obstacle with heterogeneous mixtures, a relatively direct translation from batchwise operation to continuous processing has been demonstrated. Moreover, the beneficial microwave effect, in case of the racemization as mentionedinchapter3,wasmaintainedintheFlowSynth. In the epilogue our results are put in a wider perspective. Further developments of microwavetechnologyincombinationwithindustrialpossibilitieshavebeendiscussed. Inmyviewmicrowaveheatingisonlyinnovativewhenitsprincipleofselectiveheatingis activelypursued.
140
Samenvatting
Samenvatting
Pas vanaf 1986 werd magnetronchemie in academisch onderzoek toegepast. Magnetrontechnologie is tot nu toe vooral gebruikt om efficinter nieuwe teststoffen op gramschaal te synthetiseren. Voornamelijk op het gebied van automatisering van experimenten binnen de farmaceutische industrie heeft magnetronchemie zich sterk ontwikkeld, omdat meer experimenten in een kortere tijd dan met conventionele verwarming kunnen worden uitgevoerd. Het primaire doel bij de toepassing van magnetronstralingisomreactiemengselsteverwarmen. Destandvanzakenindetoepassingmagnetronchemiewordttoegelichtinhoofdstuk1.In deresterendehoofdstukkenwordendevolgendeonderwerpenbehandeld.Inheteerstedeel wordt het onderzoek beschreven hoe magnetronverwarming chemische omzettingen versnelttenopzichtevanconventioneleverwarming.Hiervoorisdeaanzetvooreengroot deeltevindenindevelepublicatiesmetclaimsbetreffendereactieversnellendemagnetron effecten.Inhettweededeelisdeopgedanekennisvandemagnetronverwarmingtoegepast bij het opschalen van fijnchemische processen. In het laatste deel wordt beschreven hoe magnetronverwarming is ondergebracht in een zuurkastconcept voor (continue) productie vanfijnchemicalinopcommercileschaal.
In hoofdstuk 2 zijn voornamelijk reacties uit de literatuur getracht te reproduceren. Consequent is voor iedere eerder beschreven chemische reactie de vergelijking tussen magnetron en conventionele verwarming uitgezocht. Dit heeft geleid tot een logische verklaring voor de waargenomen magnetroneffecten. De biogekatalyseerde reacties zijn hierbij uitgebreid onderzocht. De enantioselectiviteit en stabiliteit zijn twee aspecten naast de reactiviteit die zijn meegenomen in het vergelijkend onderzoek. In de verestering en omestering met drie types enzymen (Novozym 435, Subtilisin, PPL) zijn identieke reactiesnelheden gevonden wanneer beide verwarmingsmethoden werden vergeleken. Bij de vorming van piperidines en piperazines uit de overeenkomstige halides, waarbij het gevormdeHXactief(viabaseofthermisch)werdverwijderd,wasgeenversnellingtezien bij gebruik van magnetronstraling. De oxidatie van benzyl alcohol tot benzaldehyde met behulpvanijzertrinitraatbleekeencomplexprocesvanredoxreactiestezijnwaarbijgeen magnetroneffecten werden gevonden in tegenstelling tot een claim in de literatuur. Bij de trialkylering van methyl gallaat werd eveneens de conversiesnelheid niet gewijzigd door gebruik van de magnetron, ondanks het heterogene karakter van deze reactie (overmaat K2CO3). Het meest illustratieve voorbeeld van een magnetroneffect bleek het Laurydone proces (verestering van pyroglutaminezuur met ndecanol) te bieden. Ten aanzien van de gemiteerde eerdere claim van een gunstig magnetroneffect bleek hier de menging van de reactanten van cruciaal belang te bieden. Een goede menging deed het oorspronkelijk waargenomen magnetroneffect met een factor 3 volledig verdwijnen. Hoewel in dit geval hetgebruikvandemagnetrongeentoegevoegdewaardeindeconversiebleektehebben,is desondanks aangetoond dat de opschaling van het Laurydone proces met magnetronstralingalszodanigtochvanrelevanteindustrilebetekenisisgeweest.Vooralle
141
Samenvatting
reactiesdieinhoofdstuk2beschrevenzijnblijkteenbetrouwbaretemperatuursmetingvan wezenlijk belang te zijn om te kunnen beoordelen of voordelige magnetroneffecten echt kunnenwordenwaargenomen.
Hoofdstuk 3 behelst de racemisatie van twee geacetyleerde aminozuren (Nacetyl fenylalanineenNacetylindoline2carbonzuur).Hetsubstraatlostinpxyleenbeperktopen de door azijnzuur anhydride bevorderde reactie wordt versneld in de magnetron (tot een factor 8). In hoofdstuk 4 is een soortgelijke benadering beschreven waarbij eveneens magnetronstraling een gunstig effect heeft op de enkelvoudige additie van een serie C6 gesubstitueerde (methyl, ethyl, isopropyl en fenyl) isocytosines aan hexamethyleendi isocyanaat. Ook hier leidt heterogeniteit van de substraten tot conversies met magnetroneffecten(toteenfactor9).Veranderingvansubstituentbenvloedtnietalleende reactiviteit, maar ook de oplosbaarheid en daardoor de grootte van de magnetroneffecten. Lokaleabsorptievanmagnetronstralingindegrenslaagvandevasteenvloeibarefaseleidt indirecttoteenverbeterdeconversiesnelheid.Eenverhoogdelokaletemperatuurleidendtot een toegenomen oplosbaarheid ligt hieraan ten grondslag. Het magnetroneffect verdwijnt helemaalindieneenoplosmiddelwordttoegevoegd(hoofdstuk3)ofvermindertindiende oplosbaarheid van het substraat toeneemt (hoofdstuk 4). Resumerend, heterogene reacties kunnen een positieve bijdrage leveren aan de reactiesnelheid bij gebruik van magnetronstraling.
Inhoofdstuk5wordtdenoodzaaktoegelichtom,bijtoepassingvanmagnetronstraling,van een batchgewijze procesvoering over te schakelen naar een continu proces vanwege de geringepenetratiediepte.Deintroductievaneenrecyclestroommaakthettevensmogelijk om reacties met een lagere conversiesnelheid in een doorstroomreactor uit te kunnen voeren.Hoewelverwarmingmetdemagnetronduurderis(ingebruikeninaanschaf)dan de conventionele verwarmingstechnieken zoals stoom en olie, zijn er toch voordelen te behalen vanuit een praktisch oogpunt en veiligheid. Verdere karakterisering van de gebruikteFlowSynthdoorstroomreactorvanMilestonetoondeaandatdereactorzichlaat bedrijvenalseencascadevangemiddeld3evengroteCSTRs. In hoofdstuk 6 wordt aan de hand van een viertal voorbeelden beschreven hoe de doorstroomreactor onder verschillende omstandigheden presteert. Bij de homogene acetylering(aspirinesynthese)werddeproductstroominportiesopgevangennanenkele passage. Drie heterogenene reacties (een biogekatalyseerde verestering, een autogekata lyseerde verestering en een racemisatie) zijn uitgevoerd met een recyclestroom. Hoewel verstopping bij heterogene reacties een belangrijk obstakel bleek te zijn, kan een relatief eenvoudige vertaalslag vanuit een batchgewijze naar een continue procesvoering worden aangetoond. Tevens is de meerwaarde van magnetronverwarming in geval van de racemisatiezoalsbesprokeninhoofdstuk3ookbevestigdbijgebruikvandeFlowSynth.
De resultaten zijn in een breder perspectief geplaatst in een epiloog, waarin de verdere ontwikkelingenvanmagnetrontechnologieendeindustrileslaagkanszijnaangegeven. Naarmijnmeningismagnetrontechnologiealleeninnovatiefalshetprincipevanselectieve opwarmingingezetwordt.
142
Curriculum vitae
Curriculum vitae
Mark Dressen is geboren op 17 augustus 1979 te Heerlen. In 1997 behaalde hij zijn VWO diplomaaanSGSt.MichielteGeleen.OpeenvolgendstarttehijmetdeHogerLaboratorium OpleidingaandeHogeschoolZuydteHeerlenmetorganischechemiealsafstudeerrichting. De bijbehorende stage en afstudeeropdracht zijn volbracht binnen de Lead Discovery UnitvanOrganonN.V.teOssin2001.Inhetzelfdejaarishijbegonnenalschemicusbinnen de R&D afdeling van DSM Pharmaceutical Products te Venlo. Procesontwikkeling en verbetering waren de kerntaken van de projectmatige werkzaamheden voor meerdere locaties. In 2005 is er door hem bewust de keuze gemaakt om zijn passie voor magnetronchemieomtezettenineenpromotieindeonderzoekseenheidAppliedOrganic ChemistryindegroepMolecularScience&TechnologyvandeTechnischeUniversiteit Eindhoven, onder leiding van eerste promotor prof. dr. L.A. Hulshof, tweede promotor prof.dr.J.Meuldijkencopromotordr.J.A.J.M.Vekemans.Debelangrijksteresultatenvan dit onderzoek staan beschreven in dit proefschrift; Microwave heating in fine chemical applications:roleofheterogeneity.
Mark Dressen was born on August 17th 1979 in Heerlen, the Netherlands. In 1997 he receivedhisdegreeinsecondaryeducationatSGSt.MichielinGeleen,theNetherlands.As afollowuphestartedtheHogerLaboratoriumOpleidingatHogeschoolZuydinHeerlen with organic chemistry (BSc) as specialization. He completed his external traineeship with the Lead Discovery Unit of Organon N.V. in Oss, the Netherlands. In the same year he startedasachemistintheR&DdepartmentofDSMPharmaceuticalProductsinVenlo,the Netherlands. The keytasks of this projectbased assignment, located at multiple locations, wereprocessdevelopmentandimprovement.In2005anintentionaldecisionwasmadeto start his PhD research, based on his passion for microwaveassisted chemistry, within the research unit of Applied Organic Chemistry as part of the group Molecular Science & TechnologyatEindhovenUniversityofTechnology,underthesupervisionofprof.dr.L.A. Hulshof, prof. dr. J. Meuldijk and dr. J. A. J. M. Vekemans. The most important results of thisresearcharepresentedinthisthesis;Microwaveheatinginfinechemicalapplications: roleofheterogeneity.
143
Curriculum vitae
144
Dankwoord
Dankwoord
Terugblikkend op periodes geeft mij altijd het gevoel dat de laatste fase de leukste was. Misschien is het omdat deze periode nog vers in het geheugen ligt of omdat je naar een hoogtepunt toewerkt. Of ben ik me meer op mijn plek gaan voelen? Tenminste zo heb ik ook mijn promotie ervaren. Ruim vier jaar heb ik aan een onderwerp mogen werken, waar ik erg nieuwsgierig naar was. Magnetronchemie heeft me inzicht gegeven en niet alleen puur in zijn eigen mechanisme, maar ook in de academische wereld. Het meest belangrijke is misschien wel om kritisch te zijn: zijn de feiten wel feiten, zijn de waarnemingen anders te interpreteren of zijn sommige conclusies te kort door de bocht. Voor deze ontplooiing wil ik graag mijn promotor, directe begeleiders en collegas bedanken: Bert, Jan, Jef en Bastiaanbedankt. Oh ja, voor diegene die het nog niet wisten, ik zelf ben niet een man van veel woorden. Hierdoor zal het dankwoord kort blijven zonder tekort te doen. Naast de vele collegas binnen de vakgroep die hebben bijgedragen aan een goed onderzoeksklimaat te creren door de jaren heen, hebben ook een aantal mensen een (aanzienlijke) bijdrage geleverd in dit proefschrift of aan mijn colloquia. Enkelen zijn binnen het referentiekader terug te vinden en anderen zou ik hier graag willen noemen. In willekeurige volgorde: Jeroen, Bart, Martijn, Sjoerd, Niels, Jelle, Melanie, Jeroen A., Ronnie. Hierbij horen ook nog mijn kamergenoten en labgenoten met wisselende samenstellingen. Twee weten echter eruit te springen. Enerzijds, alhoewel rustig, Jan v/d B. met een enorme lading aan preparaten. Anderzijds, alhoewel verhuisd, Romero met zijn (tijdelijke) hang naar Herman Finkers en de legendarische uitspraak; Er mag gedanst worden. Ook zij worden bij dezen van harte bedankt. Vanuit externe betrekkingen hebben meerdere bedrijven en personen meegewerkt in het FC SMART project. Hierbij dank aan DSM, Syncom, Milestone, Salm en Kipp, Alldos, TU Delft. Speciale dank voor het hoge servicegehalte geleverd door Pieter Stam (Salm en Kipp). Om de promotie tot een afgerond geheel te krijgen is natuurlijk een proefschrift nodig en daarmee een commissie. Het ene is direct zichtbaar en het andere doet veel werk buiten het gezichtsveld. Bij terugkoppeling is dit werk duidelijker geworden en wordt deze inspanning ook sterk gewaardeerd. Hieruit volgt het resultaat zoals het zwart op wit gedrukt is en in uw handen ligt. Many thanks to the members of my PhD-committee: L.A. Hulshof; J. Meuldijk; J.A.J.M. Vekemans; E.M. Meijer; A.I. Stankiewicz; C.O. Kappe en V. Hessel. Promoveren heeft natuurlijk een grote invloed gehad, maar de wereld draait door en mijn wereldje ontgaat niet aan externe impulsen. Mijn vrienden zijn hierbij belangrijk geweest, voornamelijk voor de momenten van ontspanning. Van mijn (schoon)familie waardeer ik hun begrip en geduldpap, mam, Jos & Nancy
145
Dankwoord
En meest belangrijk, Dymph die met mij dit bootje langs levenservaringen wil delen. Later is Lisa er ook nog bijgekomen, waardoor een prettige onbalans ontstond. Dit vergt inspanningen op nieuw terrein, maar ik krijg er iedere dag meer voor terug. Veel liefs retour aan Dymph & Lisa.
146