Voltammetric behavior of triethylamine trihydrofluoride and anisole in acetonitrile as a first approach of studies for electro-fluorination of some adducts

This is an author-deposited version published in:

http://oatao.univ-toulouse.fr/

Eprints ID: 6063

To link to this article

_:

_{DOI:10.1016/J.ELECTACTA.2012.03.045}

URL:

http://dx.doi.org/10.1016/j.electacta.2012.03.045

To cite this version

_:

_{Ciumag, Mihaela Raluca and Tzedakis, Theodore and}

André-Barrès, Christiane (2012) Voltammetric behavior of triethylamine trihydrofluoride

and anisole in acetonitrile as a first approach of studies for electro-fluorination of

some adducts. Electrochimica Acta, vol. 70. pp. 142-152. ISSN 0013-4686



O

pen

A

rchive

T

oulouse

A

rchive

O

uverte (

OATAO

)

OATAO is an open access repository that collects the work of Toulouse researchers and

makes it freely available over the web where possible.

Any correspondence concerning this service should be sent to the repository

administrator:

staff-oatao@listes.diff.inp-toulouse.fr



Voltammetric

behavior

of

triethylamine

trihydrofluoride

and

anisole

in

acetonitrile

as

a

first

approach

of

studies

for

electro-fluorination

of

some

adducts

M.R.

Ciumag

a

_,

_T.

_Tzedakis

a,∗

_,

_C.

_André

_Barrès

b

a_{LaboratoiredeGénieChimique,UMRCNRS5503,UniversitédeToulouse,UPS,118routedeNarbonne,31062Toulouse,France}

b_{LaboratoiredeSynthèseetPhysicochimiedesMoléculesàintérêtBiologique,UMRCNRS5068,UniversitédeToulouse,UPS,118routedeNarbonne,31062Toulouse,France}

a

b

s

t

r

a

c

t

Thisworkfocusesonkineticstudiesofanisoleandtriethylaminetrihydrofluoride(fluorinatingagent) onplatinumelectrodeandacetonitrileassolvent,inordertogetabetterunderstandingoftheiranodic behavior.Resultsshowthatbothcompoundscanbeoxidizedandsomekineticparametersarecalculated: thediffusioncoefficientwithintheworkingmedia,theanodicelectronictransfercoefficientandthe apparentintrinsicheterogeneouselectronictransferconstant.Anunusualvariationoftheseparameters occurswithinthechosenreactionconditions,particularlybyvaryingthetriethylaminetrihydrofluoride concentration.Preliminaryexperimentsforanodicfluorinationofdimethoxyethane(DME)andanisole werecarriedoutandevenifresultsshowapossibleelectrofluorinationfortheDME(classicallyusedas solvent),thereisnofluorinationofanisolewhenelectrochemicalmicroreactorwasused.

1. Introduction

Fluoride chemistry is importantin various fields: one

agro-biological molecule of two has at least one fluoride atom; for

thepharmaceuticaldomaintheproportionis onetothree.It is

alsoimportanttonotetheuseof 18_{F, fluorideartificialisotope,}

in medicalimagistic forpositron emissiontopography, to

diag-noseneuro-degenerativecerebraldisease(epileptic,Parkinson)or tumors.

Fluoride characteristics,e.g.itsstrongelectro-negativity,can significantlymodifythephysico-chemicalpropertiesofamolecule;

this makes fluorination to be difficult and shows the need for

newsynthesis methodologiesandnewmoreefficient

fluorinat-ing agents. Chemicalfluorination ofvarious molecules involves

multi-stepreactionprocess,usinggaseousfluorine(“Shiemannand Halex”intheindustrialscale,toluenefluorination).Forreducing therisks,severalelectrochemicalstudies[1–12]werecarriedout,

ontheanodic fluorination.Generalconclusionsof theseworks,

expressedintermsoffluorinationyield,selectivity,numberand

positioninthefinalproductoffluorideatoms,shownthedifficulty tofluorinatealkylatedgroupsinalphapositionofahetero-atom,

and inform on the optimal conditions (electrolyte nature,

sol-vents,electrodematerialandelectricenergytoprovide),tochoose.

∗ Correspondingauthor.Tel.:+33561558302;fax:+33561556139. E-mailaddress:tzedakis@chimie.ups-tlse.fr(T.Tzedakis).

Nevertheless,inourknowledge,therearenotanyelectrodekinetic

studiesconsideringmixturesofcompoundstofluorinatein

pres-enceoffluorinationagent.

Suryanarayanan and Noel [1] have shown that electrolytes

suchasEt3NF·4HF,Et3N·3HFinsolventssuchasdimethoxyethane,

dimethylformamide or acetonitrileact like nucleophylic

fluori-natingagents,conditioningtheselectivityofanodicfluorination

process.Selectiveelectrochemicalfluorinationofindanone

com-pounds was also carried out in Et4NF·3HFused as Ionic liquid

solvent.

Fuchigami is one of the authors who most studied the

organiccompoundsfluorination[2–8].AmineslikeEt3N·5HFand

Et4NF·4HFwereusedaselectrolyteandfluorinatingagentbyHou

andFuchigamiforethyla-(2-pyrimidylthio)acetate[2],for

diethy-leneglycoldimethyletherordimethoxyethane[3]aswellasfor

flavonescompoundsfluorination[4];fluorinationyieldcanreach 98%,nevertheless8F/molofsubstratearerequired.

Baba et al. [5] studied thioazolidins fluorination in

dimethoxyethane, in presence of fluorinated salts. The yield

inmono-fluorinatedproductdependsonthesolvent:inthecase

ofacetonitrile,conversionincreaseswiththenumber(n)ofHFin Et3N·nHFmolecule.Suzukietal.[6]studied

4-phenylthyomethyl-1,3-dioxolane-2-one fluorination in presence of Et3N·3HF as

electrolyte and dimethoxyethane, dichloromethane and

ace-tonitrile as solvents; meanwhile with dimethoxyethane only

themono-fluorinatedproductisobtained,theuseofacetonitrile

Nomenclature

candc0 _{concentrationandinitialconcentration(molL}−1₎

CE counterorauxiliaryelectrode

D diffusioncoefficient(m2_s−1₎

E,E0_,E0′

potential,standardpotentialandapparentstandard potential(V)

Et3N·3HF triethylaminetrihydrofluoride

F Faradayconstant(96,480Cmol−1₎

I,Ipeak,Ik current,peakcurrent,and currentfreeof

mass-transfereffects(A)

K triethylamine trihydrofluorideadsorption

equilib-riumconstant(concentrationinmolL−1) k0′

and k0 _{apparentintrinsic, and intrinsicheterogeneous}

electronictransferconstants(ms−1₎

n_gl globalnumberofexchangedelectrons

nls numberofelectronsexchangedwithinthelimiting

step

r potentialscanrate(Vs−1)

R idealgaslawconstant(8.31Jmol−1K−1)

S workingelectrodegeometricalsurface(m2₎

T absolutetemperature(K)

˛ anodicelectronictransfercoefficient

and max interfacialconcentrationandinterfacial

concen-trationatsaturation(molm−2₎

 coverageratio

 kinematicviscosity(assumedtobetheacetonitrile

viscosity:4.25×10−4_m2_s−1₎

ω angularvelocity(RPMorrads−1).

usedasfluorinatingagents,thereisnoavailableinformationabout thefluorinatedtriethylamineselectrochemicalbehavior,orkinetic parameters.

Ishii [8] shows that acidity difference between hydrogen

atoms substituted at the oxygenor at the methyl carbon (e.g.

dimethoxyethane),makesthesubstitutiontooccuratthe

hydro-genatomofthecarbonylgroup,i.e.thealphaposition(themost

expected)oftheC Obound.

Et3N·4HFand fluorohydricacid, evenif verycorrosive,were

also used as fluorinating agents by Ilayaraja et al. [9,10]; for

dimethyl glutarate fluorination, monofluoro dimethyl-glutarate

wereobtainedwitha71%yieldfora6F/molcharge.

Thispreliminaryworkpresentselectrochemicalkinetic

stud-ies,firstexpectingtheunderstandingof anisoleelectrochemical

behaviorinpresenceorabsenceofthetriethylamine trihydrofluo-ride,usedbothasfluorinatingagentandelectrolyte.Ontheother hand,theworkscarriedoutaimtohelptoapproachthefinalgoal ofourstudy,thedirectanodicfluorinationofvaluableadducts(e.g.

anisole)using anelectrochemical microreactor,appropriatedto

thiskindofreactions[13–16].

2. Experimental

Analytical degreeproducts wereprovided by Sigma–Aldrich

(dimethoxyethane and acetonitrile as solvents), and by Acros

Organics(tetrabutylammoniumperchlorateand

tetraethylammo-niumperchlorateaselectrolytes,anisoleandtrifluorinatedamine). AnAutolab30potentiostat-galvanostat(GPESsoftware)anda clas-sicalthermoregulated‘three-electrodes’electrochemicalcellwere

used.Theworkingelectrodewasagoldorplatinumrotatingdisk

(r=1mm),cleanedbeforeeachexperiment.Thecounter-electrode

wasaplatinumwireandthereferencewasaSCE.

3. Results

3.1. Electrochemicalstudyoftriethylaminetrihydrofluoride (Et3N·3HF)ongoldanode

Transientstatecurrent–potentialcurveswereplottedinorder

tostudytheinfluenceofthepotentialscanrateandconcentration onthetriethylaminetrihydrofluorideoxidation.

Fig. 1(a) shows results obtained for a solution

contain-ing0.1molL−1 of triethylaminetrihydrofluoride Et3N·3HF,used

both as supporting electrolyte and fluorinating agent, in

dimethoxyethane(DME)solvent.Aplateauappears(insteadofa

peak)intherange0.2–1VvsSCE,attributedtothetriethylamine trihydrofluorideoxidation.Forpotentialshigherthan1.5VvsSCE, theexponentialshapeofthecurvetranslatesapureactivation lim-itation,and thissignalcouldbeattributedtothesolventorthe goldelectrodeoxidation.Thecurrentmagnitudeincreaseswiththe potentialscanrateandshiftstowardsmoreanodicpotentials.For ahigherEt3N·3HFconcentration(1molL−1),theplateauappears

(Fig.1(b))intherangeof1–1.5VvsSCE,andtendstobecomea peakwhenthepotentialscanrateincreases.

Byexaminingtheelectrodesurfaceattheendofthe

experi-mentsonecanobservecorrosionofthegoldanode.Inabsenceofthe fluorinatedagentnocorrosiontakesplace.Fluorideanionarising fromtriethylaminetrihydrofluoridedissociationdecreasethegold oxidationpotentialandcausetheanodicoxidationofthismaterial (secondsignal(E>1.5V).

Current–potential curves are also plotted using acetonitrile

(CH3CN)assolvent,tetraethylammoniumperchlorate(Et4NClO4)

as electrolyte and triethylamine trihydrofluoride (Et3N·3HF) as

fluorinating agent. A similar behavior was observed for both

curve shapes and gold surface corrosion, as in the case of

dimethoxyethane.

Onecanconcludethat thegold electrodeis not appropriate

forthisstudybecauseofitsinstability;aplatinumelectrodewill beusedasworkingelectrodeforthefurtherstudyoffluorination reactions.

3.2. Electrochemicalstudyoftri-fluorinated(Et3N•3HF)

triethylamineonplatinumanode

3.2.1. Influenceofthetriethylaminetrihydrofluoride concentration

TheelectrochemicalbehaviorofEt3N·3HFwasstudiedin

ace-tonitrilecontainingBu4NClO40.2molL−1,inordertohaveabetter

understandingofitsanodicoxidation process.Potential current

curves(Fig.2)obtainedforvariousEt3N·3HFconcentrationsshow

twosignalsat0.9V/1.7Vand2.2V/3VvsSCE.

By analyzing thevariation of the current magnitude of the

firstsignal(1.5/1.7VvsSCE),versustheEt3N•3HFconcentration,

for three potential scan rates (Fig. 3(a)), one can observe the

appearanceofaplateauforEt3N·3HFconcentrationshigherthan

2_×10−2_molL−1_{.Thisbehaviorcouldbeduetothetriethylamine}

trihydrofluorideadsorptionontheplatinumsurface.

Becauseof thelow electrode area, theadsorbed quantityof

Et3N·3HFcouldbeassumed negligible,consequentlyits

concen-trationinthebulkcanbeassumedconstant.

ALangmuirisothermmodelwasexpectedtogovernEt3N·3HF

adsorption, and following analysis aims to demonstrate this

assumption. Et3N·3HF+adsorptionsite() kf ⇄ kb (Et3N·3HF)adsorbed (1)

The interfacial concentration of the adsorbedfree Et3N·3HF

canbeexpressed(usingthecoverageratio (2)), accordingthe

Fig.1.Potentialscanratedependencyonthecurrent–potentialcurves,obtainedongoldrotatingdiskelectrode(r=1mm)immersedindimethoxyethanesolventcontaining tri-fluorinatedtriethylamineintwoconcentrations:(a)Et3N·3HF0.1molL−1;(b)Et3N·3HF1molL−1.CE:Pt;25◦C;nostirring.Curves1–6:0,0.05,0.1,0.2,0.5and1.0Vs−1 respectively.

Fig.2.Current–potentialcurvesobtainedfordifferenttri-fluorinatedtriethylamine concentrations,onplatinumdiskelectrode(r=1mm)immersedinacetonitrile sol-ventcontainingBu4NClO40.2molL−1;CE:Pt;25◦C;r=0.1Vs−1;nostirring.Curves 1–6:Et3N·3HF10−3,8×10−3,2×10−2,3.5×10−2,5×10−2and7×10−2molL−1 respectively.

= [occupied_[availablesites]_sites]

= interfacialconcentration (molm

−2₎

interfacialconcentrationatsaturation max(molm−2)

(2)

=K ₁ maxc

+KC (3)

whereKistheadsorptionequilibriumconstant=kf/kbandcisthe

freeEt3N·3HFconcentration,i.e.itsinitialconcentration.

ThecurvesinFig.2(aswellascurvesinFig.6,nextsection)show thatEt3N·3HFisanirreversiblesystem:thepotentialofthe

sec-ondsignalshiftstowardsmoreanodicvalueswithtriethylamine

trihydrofluoride concentration. Consequently, the peak current

magnitudeforamonolayeradsorbedspeciescanbeexpressed[17]

byrelation(4). Ipeak=

ngl·nls·˛·F2·r· ·S

2.718·R·T (4)

where ngl is the global number of exchanged electrons, nls is

thenumberofelectronsexchangedwithinthelimitingstep,˛is

theanodicelectrontransfercoefficient,FistheFaradayconstant (96480Cmol−1_{), rispotentialscanrate(Vs}−1_{),Sistheworking}

electrodegeometricalarea(m2_{),Tistheabsolutetemperature(K)}

andRisthegasconstant(8.31Jmol−1K−1).

CombinationofEqs.(3)and(4)ledtorelation(5),allowingthe accesstotheadsorptionequilibriumconstantK.

1 Ipeak = 1 + 1 ·K·c (5) where = ngl·nls·˛·F 2 ·r·S· max 2.718·R·T

Fig.3(b)showsthelineardependenceofthecurrentreverseagainst thereverseofEt3N·3HFinitialconcentration,whichisinagreement

withtheLangmuirmodel.Assuming:ngl=2(permoleofproduced

fluorine),nls=1and ˛=0.5,onecanestimatefromtheprevious

resultstheEt3N·3HFadsorptionequilibriumconstantK,aswellas

thenumberofsitesavailabletoadsorption max.

Fig.3.Et3N·3HFconcentrationinfluenceonthecurrentofthesignallocatedinthepotentialrangefrom1.5to1.7VvsSCE.Curve(2)wasobtainedfromFig.2;curves(1)and(3) wereobtainedfromotherI/Ecurves,representedinthenextsectionornorepresentedhere.(1)–(3):0.01,0.10and0.20Vs−1_{,respectively.(a):Isignalat1.5–1.7V=f([Et3N·3HF])} fordifferentpotentialscanrates;(b):(Isignalat1.5–1.7V)−1=f({[Et3N·3HF]}−1)fordifferentpotentialscanrates.

Table1

ResultsobtainedfromFig.3(b),onthefirstanodicsignalofthetri-fluorinatedtriethylamine,relatedtoitsadsorptionontheplatinumanode. Potentialscanrate,

r(Vs−1₎

{Isignalat1.5–1.7V}−1=+ /[Et3N·3HF] R2 K(concentrationwas expressedinmolL−1₎ max(×10−4molm−2) (A−1_{) ((molL}−1₎−1₎ 0.01 33.6×103 _{420 0.988 80 6.6} 0.10 22.4× 103 _{215 0.990 104 1.3} 0.20 17.0×103 _{212 0.991 65 0.6}

TheaverageestimatedequilibriumconstantfortheEt3N•3HF

adsorptionat25◦Cis:K=(0.8_±0.2)_×102_{,arelativelylowvalue}

demonstratingarelativelylowaffinityofthiscompoundforthe

platinumanode.Theinterfacialconcentrationatsaturation( max)

oftheEt3N·3HFadsorbedonplatinumdecreaseswhenthepotential

scanrateincreases(Table1),indicatinganunusualbehaviorofan adsorbedspecies.

However,thefirstsignalappearstobea peak forthelower

concentrations(Fig.2,curves1–4),whileincreasingthe concen-tration(Fig.2,curves5and6),causestheshapeofthesignalto

becomeaplateauwithpracticallyconstantcurrent,similartoa

limitingcurrent.Thelimitingcurrentisinaccordancewithano

limitativeadsorptionstep,meaningarapidestablishmentofthe

adsorptionequilibrium,becausethehigherconcentrationofthe

Et3N·3HFinthebulk.Thisoppositebehaviorwillbediscussedin

thenextsections,becauseotherphenomenacouldtakeplace.

Indeed Et3N·3HF is a complex resulted from association of

hydrofluoricacidandEt3Naccordingtoanequilibrium.

Neverthe-less,astudyperformedby19_{FNMRonthestabilityofEt}

3N·3HF

1Minacetonitrile(Fig.4),showedthatresultingcomplexseem

tobeinstable.ObtainedspectrashowthatthesignalofEt3N·3HF

(−87ppm/CF3COOH)completely disappearsafter16h, and four

additionalsignalsappears_{(−74,−73,−66}and₋₄₈ppm).Itseems

thatEt3N·3HFdecomposestovariousproducts(Et3N·2HF,Et3N·HF,

Et3Nandhydrofluoricacid).Asimilarbehaviorwasobservedfor

othersolvents.

Ontheotherhand,studiesofFuchigamietal.[2–4]shownthat

whenfluorinationwascarriedoutusingEt3N·3HF,attheendof

thereactionfreeEt3Nwasdetected,meaningthatthecomplex

dissociatesduringthefluorination.

ThesecondEt3N·3HFoxidationsignal(Fig.2)appearstobea

peaklocatedintherangeof2.2V/3VvsSCE;itscurrentmagnitude

Fig.5. CurrentmagnitudedependenceofEt3N·3HFconcentrationforthesecond oxidationsignal(curvesofFig.2),locatedintherange2.2–3VvsSCE.(1)–(3):0.01, 0.10and0.20Vs−1_{respectively.}

showsalineardependenceagainstthetriethylamine

concentra-tion(Fig.5),meaninga diffusionlimitedreaction.Theintercept ofthesesstraight linesis notzero,becausetheresidualcurrent wasnotremoved;indeed,blankeliminationisdifficultbecausethe Et3N·3HFactbothaselectrolyteandelectroactivespecies,

influenc-ingthemigrationcurrentandcausingchangesintheblank.

Inaddition,thissignalshiftstowardsmoreanodicpotentials

anditspeakpotentialseemstobestronglyinfluencedbyEt3N·3HF

concentration(Fig.2),probablybecauseofpHvariationinducedby apartialdecompositionofEt3N·3HF,leadingtofluorohydricacid.

Inconclusion, triethylaminetrihydrofluoridecanbeoxidized

bothasanadsorbedandfreeforms.Adsorptiononplatinum

elec-trodeobeystotheLangmuirisothermmodel, whilethesecond

electrochemicalsignaltranslatesamasstransferlimitation.

t=0 t=1h t=2.5 h t=16 h -87 -40 -50 -60 -70 -80 -90 -100 -110 δ /ppm -86 -73 -48 -40 -50 -60 -70 -80 -90 -100 -110 δ /ppm -84 -73 -48 -40 -50 -60 -70 -80 -90 -100 -110 δ /ppm -74 Massif, three signals -66.4 to -66.6 -73 -48 -40 -50 -60 -70 -80 -90 -100 -110 δ /ppm

Fig.6. Current–potentialcurvesobtainedforvariouspotentialscanrates,onplatinumdiskanode(r=1mm)immersedinacetonitrilecontainingBu4NClO40.2molL−1;CE: Pt;25◦_{C;nostirring.1–4:0.017,0.083,0.170and0.500Vs}−1_{respectively.(a)Et3N·3HF8×10}−3_molL−1_{;(b)Et3N·3HF2×10}−2_molL−1_.

3.2.2. Influenceofthepotentialscanrateonthetriethylamine trihydrofluorideoxidation

InordertoaccesstotheEt3N•3HFkineticparameters,its

electro-chemicalbehaviorintheacetonitrile–Bu4NClO4mediawasstudied

(Fig.6),byplottingthecurrentpotentialcurvesforvarious poten-tialscanrates.Aspreviously(Fig.2),thecurrent–potentialcurves

obtainedfor differentpotentialscan rates showtheadsorption

wave(mainlypresentintheexpectedform,i.e.apeak)followed

bythediffusionpeak.

For thefirst signal,a linear dependence (6)of thepotential

scanrateonthecurrentmagnitude,accordingtorelation(4),was

obtainedandconfirmstheelectrochemicalreactionlimitationby

theEt3N·3HFadsorptionontheelectrodesurface.

Ipeak=0.121+0.0183r, R2=0.989 (6)

Theinterceptofthisrelationisnotzero,becausetheresidual cur-rentwasnotremoved(samereasonasinthecaseofFig.5).

Thepeakpotentialofthesecondsignal(2.6–2.8VvsSCE,Fig.6)

shiftstowardsmoreanodicvalueswhenthepotentialscanrate

increases;itsvariationversusthenaturallogarithmofthescanrate, accordingtotherelation(7),allowshavingsomeinformationabout thesystemreversibility/irreversibility:

Epeak=E0′+ R×T ˛×nls×F



0.78+ln D1/2 k0 + ln



˛×nls×F×r R×T



1/2



(7) whereE0′

istheapparentstandardpotentialoftheelectroactive system(V)andk0_{istheintrinsicheterogeneouselectronictransfer}

constant(ms−1).

Fig.7.Secondanodicpeakpotentialvariationagainstthelogarithmofthe poten-tialscanrateforvariousEt3N·3HFconcentrations.ResultsfromFig.6andother curvesnot presented.Curves1–4:Et3N·3HF8×10−3_{, 2×10}−2_,3.5×10−2 _and 7×10−2_molL−1_{respectively.}

Resultsfor Et3N·3HFconcentrationsin therange8× 10−3 to

7×10−2_molL−1_{(Fig.7)showalinearcorrelation,meaningthat}

theelectroactivesystemisirreversible.

The average value of slopes obtained in Table 2 _{(¯x =}

2.303(RT/2˛nlsF))allowsestimating˛nls∼0.36.

OnecanobservethatanincreasedEt3N·3HFconcentrationdoes

notaffecttheslope,whiletheinterceptincreases.Changesinthe interceptvalue(Table2)withtheEt3N·3HFconcentrationarean

unclassicalobservationandtheywillbediscussedattheendof

thisparagraph.

Takingintoaccountthe˛nlsvaluedeterminedpreviously,the

Et3N·3HFdiffusioncoefficientcanbedeterminedusingthelinear

dependenceofthepotentialscanratesquarerootwiththe

cur-rentmagnitude(8)andresultsforfourEt3N·3HFconcentrations

areindicatedinTable3. Ipeak=2.99×105·ngl·

p

˛·nls·D·S·c0·

√

r (8)

whereDistheEt3N·3HFdiffusioncoefficientinCH3CNcontaining

Bu4NClO40.2molL−1(m2s−1);c0isthetriethylamine

trihydroflu-orideinitialconcentrationinthebulk(molL−1_{);ngl=2(permole}

ofproducedfluorine).

The relative good linearity of the four curves Ipeak=f(√r)

confirms the kinetics limitation by the Et3N·3HF diffusion.

Nevertheless, determinedEt3N·3HFdiffusion coefficientslightly

((2.85₋1.24)/2.85=0.56)changeswithconcentration,a logarith-micincreaseofthisparameterwasobserved(Fig.8),relation(9): DEt₃N·3HF (m2s−1)=4.77×10−10

+7.29×10−11_{·ln[Et3N·3HF] (molL}−1_{), R}2

=0.995 (9)

Thisvariation,evenifnotimportant,seemsunusual.Apossible rea-sonexplainingthischangecouldbethemigrationcurrent.Indeed,

Fig.8. Et3N·3HFdiffusioncoefficientvariationwithinitialtri-fluorinated triethy-lamineconcentration.

Table2

ResultsobtainedfromFig.7,onthesecondanodicsignalofthetri-fluorinatedtriethylamine,relatedtoitsdiffusion-controllednature.

[Et3N·3HF](molL−1) Epeak(V)=+xlog(r)(Vs−1)accordingtorelation(7) R2 ˛nls ˛nls

 x ¯x 8×10−3 _{2.639 0.084} 0.0805 0.999 0.35 0.36 2×10−2 _{2.830 0.087 0.984 0.34} 3.5×10−2 _{2.960 0.074 0.965 0.39} 7× 10−2 _{3.048 0.077 0.969 0.37}

becausethechosenTBAPconcentration(0.2molL−1),theions aris-ingfromEt3N·3HFdissociation,especiallyF−,couldmigrateand

consequentlyincreasethemigrationanodiccurrent,whichcannot

beneglectedinthewholeexaminedrangeofEt3N·3HF

concentra-tion:8_×10−3–2_×10−2molL−1.

Thetransferencenumberoftheelectroactivespecies,tEt3N·3HF

canbeestimatedaccordingfollowrelation: tEt₃N·3HF=

z2

anionarisesfromEt₃N·3HFdissociation×F2×uanionarisesfromEt3N·3HFdissociation×[Et3N·3HF]

X

allionsj

z2

j ×F2×uj×[j]

Becausewedonotknowmobilityofionsarisingfromthese

elec-trolytesinacetonitrile,wewillusethefollowingapproximations: - bothTBAPandEt3N·3HFareassumedtobebinaryelectrolytes,

-mobilitiesujofalljionsarisingfromtheseelectrolyteshavethe

sameordermagnitude.

Thetransferencenumbercanbewrittenasfollow:

tEt₃N·3HF=

[Et3N·3HF]

2_×[Et3N·3HF]+2×[TBAP]

anditsvalues(Table3),intherangefrom0.01to0.13,arerelatively low.

Notethatthesevaluesaresimpleestimations,becausethe dis-sociationofEt3N•3HFispartialandtheconcentrationofthearising

anionislowerthantheEt3N•3HFconcentrationusedtoestimate

thetransferencenumber.Intheseconditions,theestimated trans-ferencenumberishigherthantherealone.

Forabinaryelectrolyte,theproduct(1/(1−tj))translatesthe

contributionoftheionjmigrationflux,inthediffusionfluxofthis ion.Consequently,relation(8)canberewritten:

Ipeak= 1 1₋t_j·2.99×10 5 ·ngl·

p

˛·nls·D·S·c0·√r (8bis)

Usingtherelation(8bis),anestimationoftheEt3N·3HFdiffusion

coefficienttakingintoaccountthemigrationcurrentwasproposed andresultsareindicatedinTable3.Asprevious,obtainedvaluesof DchangewithEt3N·3HFconcentration;nevertheless:

-differencesintheDvalues,areinanarrowrange(1.1_×10−10

to 2.2_×10−10_{) instead 1.2×10}−10 _{to 2.8×10}−10 _e.g.

((2.2−1.1)/2.2=0.5),

-effect of the migration term is relatively low:

6%<1{(DD−DM/D)/DD}<22%.

ThevariationofDvsconcentrationseemsunusual,as

passiva-tionproblemscouldnotexplainitcompletely.IncreasedEt3N·3HF

concentrationleadstoamoreimportantpassivationand

conse-quently,causescurrent anddiffusioncoefficienttodecrease.As

explained above, Et3N·3HF can decompose in free fluorohydric

acidandtriethylamine;eitherEt3N·3HFandHFoxidationcouldbe

attributedtotheobtainedsignal.ThusdeterminedDusingtheslope ofrelation(8bis)isanapparentdiffusioncoefficientbecausethe valueoftheinitialconcentrationwasusedinsteadthereal concen-trationofthecorrespondingelectroactivespecies(whichislower onetheinitialconcentration)

Diffusion coefficient variation versus the concentration of

Et3N·3HFcanalso(partially)explainthechangesontheintercept

value(Table2)previouslyobserved.Consequently,determination oftheintrinsicheterogeneouselectronictransferconstantk0_seems

tobedifficultbecauseboththeconcentrationoftheelectroactive speciesandtheE0′

valueareunavailable.Onewilltrytoaccessto thisparameterinthenextsection.

3.2.3. Influenceoftheangularvelocityoftheplatinumanodeon thetriethylaminetrihydrofluorideoxidation

Inorder tobetterunderstandthis system, (estimationofk0_,

comparisonof theresultsfor D),current–potential curveswere

plottedat thesteadystate,forvarious Et3N·3HF concentrations

andangularvelocitiesofarotatingplatinumdisk.Results(Fig.9) showasimilarshapecurveforalltheexaminedconcentrations.The currentmagnitudeforthefirstoxidationsignal(1.2–2VvsSCE), previouslyattributedtotheEt3N·3HFadsorptionontheelectrode

surface,seemstobeaffectedbythestirring,forthelowexamined concentrations(Fig.9(a)and(b));noeffectwasobservedforhigher concentrations:(Fig.9(c)and(d)).

Theoretically,adsorptionisnotaffectedbystirring.

Neverthe-less, a slowadsorption ratecan influencethe time requiredto

reachtheequilibrium.ForaEt3N·3HFconcentrationhighenough,

the adsorption equilibrium is rapidly reached and the stirring

doesnotaffectthecurves.Conversely,for concentrationsupto

2_×10−2molL−1, thetime necessarytosaturatingthe electrode

is higher and the stirring can have an effect. The low value

of Et3N·3HF adsorption equilibrium constant (K) and the max

Table3

Et3N·3HFdiffusioncoefficientcalculatedwithrelation(8)and˛nlsvaluesfromTable2;ngl=2.ResultsobtainedfromFig.6. [Et3N·3HF] (molL−1₎ Slope Ipeakat2.6/2.8V=f(r1/2) R2 _D Et3N·3HF(×1010m2s−1) (withoutmigration contributiontothecurrent)

Estimatedtransferencenumber ofEt3N·3HF,tEt3N·3HF=[Et3N· 3HF]/(2[Et3N·3HF]+2[TBAP])

DEt3N·3HF(×1010m2s−1) (withmigration

contributiontothecurrent)

DD−DM/D_D D (%) 8×10−3 _9.91×10−5 _{0.964 1.24 0.02 1.16 6.3} 2×10−2 _3.08×10−4 _{0.982 1.97 0.04 1.70 13.7} 3.5×10−2 _6.18×10−4 _{0.998 2.27 0.07 2.10 7.5} 7×10−2 _1.35×10−3 _{0.989 2.85 0.13 2.21 22.4}

Fig.9.Angularvelocity(ω)dependenceofthecurrent–potentialscurvesobtainedonplatinumrotatingdisk(r=1mm),immersedinacetonitrilesolvent,containingBu4NClO4 0.2molL−1_{;CE:Pt;r=4×10}−3_Vs−1_;25◦_{C.Curves1–5:ω=100rpm,250rpm,500rpm,1000rpmand5000rpmrespectively.(a)Et3N·3HF8×10}−3_molL−1_{;(b)Et3N·3HF} 2×10−2_molL−1_{;(c)Et3N·3HF3.5×10}−2_molL−1_{;(d)Et3N·3HF7×10}−2_molL−1_{respectively.}

variationwiththepotentialscanrate(Table1),areinaccordance withtheseresults.

Anotherreasonthatcouldexplainthisbehaviorisrepresented

byatemperaturevariationattheelectrodesurface,arisingfrom

Et3N·3HFdecomposition,aphenomenondampedbythestirring.

Thesecondoxidationsignalappearstobeapeakandits

poten-tialshiftstomoreanodicvalues_(∼2.3to_∼3VvsSCE)whenthe

Et3N·3HFconcentrationincreases,factthatconfirmsitsirreversible

behavior.Foraclassicalelectroactivesystem,this‘diffusionlimited’ signal,atthesteadystate,hastoshowalimitingcurrentplateau;

theobtained‘peak’-shapedcurvetranslatespassivation

phenom-enacaused bya thin layerofgaseous fluorineproduced atthe

electrodesurface.Usingthe‘net/exemptedoftheblank’maximum

currentofthissignal,asanapproximationfortheexpectedlimiting current(Fig.9),linearcorrelations(Table4)wereobtainedforthe fourEt3N·3HFconcentrations,followingLevichrelation:

Ilim=0.62·ngl·F·S·D(2/3)·c0·−(1/6)·√ω (10)

where ngl is assumed tobe 2(by molecule ofproduced F2), ω

is the electrode angular velocity (rads−1_{) and  is the media}

kinematic viscosity, assumed to be the acetonitrile viscosity

(4.25_×10−4m2_s−1).

ThelinearcorrelationIlim=f(√ω)obtained,couldmeanthat

passivationdoesnotstronglyaffecttheelectrodesurface,because

a strongstirring could contribute toremove bubbles fromthe

electrodesurface.Slopesofthefourcurvesallowestimationof tri-ethylaminetrihydrofluoridediffusioncoefficientD,andtheresults arepresentedinTable4.Aspreviously,Dwasestimatedtakinginto accountthemigrationcurrentcontributionandtheresultsarealso showninTable4.

The obtained values of diffusion coefficient allow several

remarks:

-theydecreasewhenEt3N·3HFconcentrationincreases,

-comparisonofDvaluesdeterminedbythetwomethods

(influ-enceofboththepotentialscanrateandangularvelocityofthe rotatingdiskelectrode(Table4),takingintoaccountonlythe dif-fusioncurrent)showratiosdecreasingintherangefrom14to0.5

withintheEt3N·3HFconcentrationexaminedrange.

- for low Et3N·3HF concentrations (<2×10−2_molL−1_{), the}

obtaineddiffusivitiesare relativelyhigh, and thatis in accor-dancewitha‘lowsize’electroactivespecies,i.e.HF(orF−),arising fromEt3N·3HFdecomposition,especiallyinthediffusionlayer

area.AssumingthatEt3N·3HFdecompositionisanequilibrium

reaction,stirringcouldcontributetofavorthis equilibriumby

a standardization of both theconcentration and temperature

especiallyinthediffusionareaclosetotheelectrode.Forthese

‘lowEt3N·3HFconcentrations’,thedeterminedDvaluemustbe

correctbecausemostoftheintroducedEt3N·3HFdecomposes,

and the concentration, used to determine D (Levich slope,

Table4),isveryclosetoitsinitiallyintroducedconcentration.

In addition,low Et3N·3HFconcentrations leadtolow current,

consequentlytoalowfluorinequantityproducedattheinterface, andsoaminimizedeffectofpassivation.

-for Et3N·3HF concentrations higherthan 2×10−2molL−1,the

influenceofstirringontheEt3N·3HFdecompositionremainslow.

TheconcentrationoftheelectroactivespeciesHF(orF−),arising

by Et3N·3HF decomposition,is lower than theinitial

concen-trationofEt3N·3HF,usedtoestimateD,consequently, itisan

apparentdiffusioncoefficient,anditsvaluewasassumedtobe

closeto1.7_×10−9_m2_s−1_.

In order toestimate Et3N·3HF intrinsic heterogeneous

elec-tronictransferconstantk0′

,analysisofcurvesfromFig.9,using theKouteck ´y–Levichequation,wascarriedoutforthesecond tri-ethylamineoxidationsignal,locatedinthepotentialrange2/3Vvs SCE: 1 I = 1 Ik+ 1 0.62nglFSD2/3c0−(1/6)ω1/2 (11) whereIisthecurrentmagnitudeintheactivation/diffusion ‘poten-tialrange’ofthecurrent–potentialcurves(i.e._∼2.6to_∼2.8V)and Ikrepresentsthecurrentfreeofmass-transfereffects:

Ik=nFSck0exp

h

_˛nlsF

RT (E−E

0′

)

i

(12)

Results (Fig. 10) show a linear variation of 1/I=f(1/ω1/2_{) for}

the examined potentials and for a Et3N·3HF concentration

Table4

Tri-fluorinatedtriethylaminediffusioncoefficientinacetonitrile(calculatedfromresultsofFig.9,usingIlim=f(ω1/2)slopeforfourdifferentinitialconcentrations)andtheir comparisonwiththediffusioncoefficientvaluescalculatedwithIpeak=f(r1/2)slope(Table3).

[Et3N•3HF] molL−1 Slope Ipeak=f(ω1/2) R2 _D Et3N·3HF(×1010m2s−1) (withoutmigration contributiontothecurrent)

Estimatedtransference numberofEt3N·3HF, tEt3N·3HF

DEt3N·3HF(×1010m2s−1) (withmigration

contributiontothecurrent)

DLevich Dpotentialscanrate

8× 10−3 _1.573×10−5 _{0.998 17.2 0.019 16.69 14.4}

2×10−2 _3.629×10−5 _{0.999 15.2 0.045 14.21 8.4}

3.5×10−2 _2.766×10−5 _{0.986 4.4 0.074 3.89 1.9}

7×10−2 _2.859×10−5 _{0.999 1.6 0.130 1.32 0.6}

examinedconcentrations;eachconcentrationallowsgettingaset

ofparallelslines1/I=f(1/ω1/2_{),andtheslopeisdifferentforeach}

set(similarbehavior totheresultsindicatedin Table4).The Ik

determinationwascarriedoutfromtheinterceptofthesestraight lines.

AnalysisoftheIkvalues,obtainedfromtheinterceptofcurves

1/I=f(1/ω1/2_{),wasperformedaccordingto:}

lnIk=ln(nglnFSc0k0 ′ )+˛·_RTnls·FE (13) where k0′ =k0_exp

h

−˛n_RTlsFE0′

i

(14) Eq.(14)istheapparentintrinsicheterogeneouselectronictransfer constant.Despiteacertaindispersionofpoints,asatisfactorylinear correlationofln(Ik)=f(E)wasobtainedforEt3N·3HFconcentrations

intherange8_×10−3 to3.5_×10−2molL−1 (resultspresented in

Table5).

Thekineticparametersk0′

and˛nlswerecalculated,andthe

obtainedvaluesallowsomeremarks:

-˛nls changes with Et3N·3HF concentration; this dispersion is

mainlycausedbythefactthatresultsareverysensitivetothe cur-rentmagnitude:indeed,a5%changeincurrentvaluescanleadto achangeintheinterceptofthecurves1/I=f(ω1/2_)ofabout20%.

Thiscouldexplaintheobtainedratios ˛n_{ls(Kouteck ´y}–Levich)

˛nlspotentialscanrateTable2

inarangefrom0.6to1.2.Nevertheless,theaveragevalue˛nls=

0.35isveryclosetotheoneobtainedintheprevioussection. -Asimilarvariationwasobtainedfortheapparentintrinsic

hetero-geneouselectronictransferconstantk0′

:itsvalueschangeversus

Fig.10. Variationofthecurrentreverseagainstthereverseoftheangularvelocity squareroot.Et3N·3HF2×10−2_molL−1_{(Fig.9(b)).Thecurrentvaluesarechosenfor} potentialsintheactivation/diffusionpartoftheI/Ecurve.1–3:2.585,2.611and 2.637VvsSCErespectively.Thesameinvestigationwasperformedforallexamined concentrations(8×10−3_to7×10−2_molL−1_).

Et3N·3HFconcentrationandareverylowbecauseincludingthe

exponential term: _{exp [−(˛nls}F/RT)E0′_{] Moreover, in addition}

tothecurrent sensitivitypreviouslyindicated, changes inthe

k0′

values couldbedue tothevariationoftheapparent

stan-dard potentialof theelectroactive system(HF/F2).E0′ closely

dependsofthesolutionpH,whichdecreasesbecauseHFarises

fromEt3N·3HFdecomposition.Consequently,therealvalueofthe

intrinsicheterogeneouselectronictransferconstantk0_ishigher

thantheoneofk0′

.Letusnotethatk0′

value1.54_×10−11ms−1

obtainedforthehighestconcentrationis‘themostcorrect’,orat least‘theclosest’totherealvalueofk0_{,astheapparent}

stan-dardpotentialE0′

isthelowestone(thelowestinfluenceofthe exponentialterm).Inordertoovercomethesedifficulties,andto accesstomoreprecisekineticparameters,apossibilitycouldbe tooperateelectrolysesinabuffermedia,forexampleamixture ofpolarandnon-polarsolvents(i.e.HF+heptane).This

exper-imentalinvestigationwillconstitutea futureworkinorderto

gofurthertotheknowledgeofthissystem.Neverthelesscurrent estimatedvalueofk0′

includesapparentstandardpotentialE0′

andisveryusefultoestimateoverpotentialsandconsequentlyto performpreciseenergybalanceforfluorinationinthe microreac-tor.

3.3. Electrochemicalstudyofanisoleonplatinumanode

Electrochemicalkineticstudieswerecarriedoutusinga

plat-inum disk anode, in order to understand the anisole anodic

behavior,inabsenceandinpresenceoftriethylamine trihydroflu-oride.Thisstudyexpectstohelpinapproachthefinalgoalofour project,‘thedirectanodicfluorinationofvaluableadductsusingan electrochemicalmicroreactor’.

ThevoltammogramsinFig.11(a),obtainedforvariouspotential scanrates,showthatanisolecanbeoxidizedforpotentialshigher

than1.5VvsSCE;neverthelessthecurvesshapeisunusual.The

firstpeak-shapedsignalappearsat1.6to1.8VvsSCE,meanwhile

forhigherpotentials(1.9–2.7VvsSCE),thecurrentbecomes practi-callyconstantandthecurvepresentsaplateau.Additionalanisole oxidationsmighttakeplace,causingincrease(insteaddecrease)of

currentmagnitude.

Thepeakpotentialshiftstowardsmoreanodicvalues,

accord-ingtoalinearvariationversusthescanratelogarithm(15)and

confirmsanirreversibleanisoleoxidation:

Epeak=Ea0′+ RT ˛anlsF



0.78+ln D 1/2 a k0 a +ln



˛anlsF RT



1/2



+1.1515_˛RT anlsF log(r)=1.8551+0.0822log(r) R2 =0.997 (15) wherek0

astheanisoleintrinsicheterogeneouselectronictransfer

constant.Thecurveslopeallowsestimationof˛anls=0.36.

Rela-tion(16)illustratesalineardependenceofthepotentialscanrate squarerootonthemagnitudeofthe‘peak’-current(i.e.potentialsin

Table5

Variationofln(Ik)withthepotentialEforthreeEt3N·3HFconcentrations.

[Et3N·3HF](molL−1) ln(Ik)=X+Q·E(relation(13)) R2 k0′(ms−1) ˛nls(˛nls=0.35)

˛n_{ls(Kouteck ´y −Levich)} ˛nlspotentialscanrateTable 2

X Q

8×10−3 _{−49.799 16.436 0.986 4.86×10}−23 _{0.42 1.2}

2×10−2 _{−43.982 13.903 0.997 6.54×10}−21 _{0.35 1.1}

3.5×10−2 _{−36.883 10.793 0.996 4.52×10}−18 _{0.28 0.6}

ResultsfromFigs.9and10(andfiguressimilarto10,nopresent,foralltheexaminedconcentrations).

therange1.65–1.85VvsSCE),andconfirmsthekineticslimitation byanisolediffusion. Ipeak=2.99×105·ngl

p

˛anlsDaSc0 √ r=1.62×10−4√_{r R}2 =0.991 (16)

Diffusion coefficient of anisole Da was estimated at

∼2.0×10−8_m2_s−1_{, assuming the transferred electron}

num-berngl=2, and usingthecurve slope aswellas thepreviously

calculatedvalue˛anls.Thisvalueappearstobehighforarelatively

‘bigsized’moleculelikeanisole;acetonitrilelowviscosity,aswell asitslowabilitytosolvateanisole,couldbeoneexplanationfor

this high value.In addition, as indicate previously,the current

magnitude‘capturedatthepeak’ishigherthantheexpectedone,

becausethepresenceofmorethanoneoxidationsteps.

Inordertoconfirmthisdiffusioncoefficientvalue,andtoaccess tothekineticparametersofthesystem,anothersetofexperiments wascarriedout.Fig.11(b)showscurrent–potentialcurvesobtained atsteadystateforvariousangularvelocities(ω)oftherotating

plat-inumdiskanode.Theoxidationsignalbeginsat_∼1.6VvsSCEand

isfollowed byadiffusionplateau,presentinga limitingcurrent

(mainlyforω<1000rpm).Forhigherangularvelocities,the

cur-rentmagnitudereachesamaximuminapotentialrangefrom2

to2.2VvsSCE,anddecreasesformoreanodicpotentials.Astrong stirringleadstoacurrentincreaseandconsequently,more impor-tantquantitiesofanisoleareinvolvedintothereactioncloseto theinterface.Intheseconditions,partialelectrodepassivationcan occurandthecurrentdecreases.

Using the maximum current as an approximation for the

expectedlimitingcurrent(Fig.11(b)),andapplyingtheLevich rela-tion(10),anon-linearcorrelationofIlim=f(ω1/2)isobtainedwithin

theωexaminedrange(Fig.12(a)),inaccordancewiththeprevious

passivationexplanations.However,forlowangularvelocities(ω

lowerthan1000rpm),passivationeffectsarereducedandalinear correlationofIlim=f(ω1/2)isobtained(17).

Ilim(A)=8.82×10−6ω1/2(rads−1) 0.5

, R2

=0.9999 (100_≤ω≤500rpm) (17)

Anisole diffusion coefficient deduced from the slope is

Da∼1.6_×10−8_m2_s−1 _{(assuming ngl=2) a value that seems}

to be more precise than the previous one, because the low

passivationeffectsatlowstirringrates.

Analysis of curves in Fig. 11(b) was carried out using the

Kouteck ´y–Levich Eq. (11), by plotting the variation of the

cur-rent reverse against thereverse of the angular velocity square

root.Thecurrentwascapturedintheactivation/diffusionrangeof current–potentialcurves,forpotentialsfrom1.7to1.75VvsSCE, andresults(Fig.12(b)),showalinearvariationof1/I=f(1/ω1/2_)for

thefourexaminedpotentials.

LogarithmicanalysisoftheobtainedIkvalues(fromthe

inter-ceptofcurvesFig.12(b))versustheanodicpotential,accordingto relation(13),allowsdeterminationofthekineticparametersk0′ a

and˛anlsandresultsshowalinearcorrelation(18).

ln(Ik)=15.144E−12.241 R2=0.994 (18)

where ˛anls=0.39 and ka0′=k0exp

h

−(aanlsF) RT E0 ′

i

∼8× 10−6_ms−1.

Theapparentintrinsicheterogeneouselectronictransfer

con-stant (k0′

a) value is low, indicating an irreversiblebehavior for

anisoleoxidation.However,arigorousanalysisrequiringadditional investigationstocalculatetheapparentstandardpotentialE0′

ofthe system,hastobecarriedoutinordertoobtaintheintrinsic het-erogeneouselectronictransferconstant(k0_{).Forthepursuitofthe}

fluorinationwithinamicroreactorofadductslikeanisolethek0′ a

valuewillbeused.

3.4. Electrochemicalstudyofanisole/triethylamine

trihydrofluoridemixtureonplatinumanodeandpreliminaryset offluorination

Themain goalofthis partof thestudyis tounderstandthe

electrochemicalbehaviorofthesystemanisole/triethylamine

tri-hydrofluoridebeforeapproachingtheanisolefluorinationwithin

anelectrochemicalmicroreactor.Becausethetriethylamine

trihy-drofluorideisusedasafluorinatingagentaswellasanelectrolyte, itsconcentrationwillbehigher(8_×10−3molL−1)thananisole con-centration(10−3_molL−1_).

Current–potentialcurves,obtainedbyvaryingboththe

poten-tialscanrate(Fig.13(a))andtheangularvelocityoftherotating diskelectrode(Fig.13(b)),showthepresenceofthreedistinct sig-nals,inthepotentialranges1–1.6V,1.7–1.9Vand2–2.6VvsSCE respectively.

Fig.11.Current–potentialcurvesobtainedonplatinumrotatingdisk(r=1mm),immersedinacetonitrilesolventcontaininganisole10−3_molL−1_{andBu4NClO40.2molL}−1_; CE:Pt;25◦_{C.(a)Curvesplottedatdifferentpotentialscanrates;nostirring.1–4:0.017,0.083,0.170and0.500Vs}−1_{respectively,and(b)curvesplottedatdifferentangular} velocities;r=4× 10−3_Vs−1_{.1–5:100,250,500,1000and3500rpmrespectively.}

Fig.12.Angularvelocitydependenceontheanisoleoxidationcurrent;resultsobtainedfromFig.11(b).(a)Limitingcurrentvariationagainstthesquarerootoftheangular velocityforanisoleoxidation;and(b)variationofcurrentreverseagainstthereverseoftheangularvelocitysquareroot.Thecurrentwascapturedintheactivation/diffusion rangeoftheI/Ecurve.1–4:1.702,1.712,1.725and1.751VvsSCE,respectively.

Fig.13.Current–potentialcurvesobtainedonplatinumrotatingdisk(r=1mm),immersedinacetonitrilesolventcontaininganisole10−3_molL−1_{,Et3N·3HF8×10}−3_molL−1 andBu4NClO40.2molL−1;CE:Pt;25◦C;nostirring.(a)Differentpotentialscanrates,1–4:0.017,0.083,0.170and0.500Vs−1respectively,and(b)differentangularvelocities, 1–3:100,250and500rpmrespectively;r=4×10−3_Vs−1_.

Fig.14.Reactionmechanismproposedforanisoleelectro-fluorination.

Thesevoltammogramscorrespondtotheadditionof

triethy-laminetrihydrofluorideandanisoleoxidationcurves.Thesecond

signal(1.7–1.9V)wasattributedtoanisoleoxidation, whilethe

first(1–1.6V)andthethird(2–2.6V)wereattributedtothe tri-ethylaminetrihydrofluoride.

Thecurrentmagnitudeforthetriethylaminetrihydrofluoride

oxidationinpresenceofanisoleseemstobelowerthanthecurrent

magnitudefor thetriethylamine trihydrofluoride alone. Indeed,

acomparison ofthesecurrentsforcurves obtainedinthesame

conditions(no.3Fig.13(b), I_{∼ 1.71×}10−4Aand no.3Fig.9(a),

I_∼1.31_×10−4_{A) shows significant differences. An explanation}

couldbethatanisoleoxidationleadstocationradicalsthatare fur-therreactingwithtriethylaminetrihydrofluoride(orfluoride).This

couldbeinaccordancewiththereactionmechanismsproposedby

HouandFuchigami[3]fordimethoxyethane(Fig.14).

Whenoperatingwithanisolealone,forrelativelystrong

stir-ringrates,itsoxidationtocationradicalsseemsleadingtoapartial

anodepassivation,causinga decreaseinthecurrentmagnitude

(Fig.11(b)).Thepassivationseemstodisappearwhentriethylamine

trihydrofluoride is present, probably because of the

electrore-generated radicals consumption, avoiding theirduplication and

consequentlytheelectrodepassivation.Thisfactallowsusto pur-sueinthedesigningoftheelectrofluorinationprocesswithoutneed ofaspecialattentiontothepassivationproblems.

This study shows that for anisole electro-fluorination it is

important tomaintaintheanodic potential at2.2–2.3V vs SCE

duringfluorination,inordertoavoidtri-fluorinatedtriethylamin

oxidation togaseous fluorine.A preliminarysetofexperiments

wasperformedusinga filter-presselectrochemicalmicroreactor

[18,19]inordertofluorinatebothanisoleanddimethoxyethane

in various working conditions ₍₋₁₀◦_C<T<40◦_{C). The results}

didnot showanisole fluorination (either onthe aromatic ring

nor the methoxy group). Nevertheless, dimethoxyethane was

‘mono-fluorinated’ (which seems to be more difficult than

anisole fluorination)and two reaction productswere obtained

(FCH2OCH2CH2OCH3chemicalyield:62%andCH3OCHFCH2OCH3

chemicalyield:38%),whenachargeof7F/molewasused.Amore

importantinvestigationofthefluorinationofanisolewillconstitute thegoaloffutureworks.

4. Conclusions

Theaimofthispreliminarystudywastoexaminethe

electro-chemicalbehavior of anisoleand triethylamine trihydrofluoride

(fluorinatingagent)inordertohelptoapproachthefinalgoalof ourstudy,thedirectanodicfluorinationofvaluableadductsusing

Several kinetic studies, performed in various conditions,

showthat both compounds canbe oxidized.Triethylamine

tri-hydrofluoride ‘doubly’ reacts, under both adsorbed and free

forms, at two different potentials. Adsorption is governed by

the Langmuir model, with an adsorption equilibrium constant

K=(0.8±0.2)×102 _{estimatedat25}◦_{Candindicatingarelatively}

lowaffinityofthetriethylaminetrihydrofluoridefortheplatinum anode.

Onesignalwasobserved foranisoleoxidation, even ifmore

thanonereactionseemstooccur.Electrodepassivation,observed

duringanisoleoxidation,disappearinpresenceoftriethylamine

tri-hydrofluoride,phenomenathatcouldbeanimportantadvantage

forthefluorinationprocess.

Triethylaminetrihydrofluoridediffusioncoefficientwithinthe workingmedia,determinedbytwodifferentways(1.2_×10−10_<D

(m2_s−1_)<1.7×10−9_{),appearstobe‘high’andvariesversusthe}

tri-ethylaminetrihydrofluorideinitialconcentrationintheexamined

range(8_{× 10}−3_–3.5×10−2_molL−1_{).OnecanproposeforEt3N·3HF,}

anapparentdiffusioncoefficientvaluescloseto1.7×10−9_m2_s−1_,

determinedforthelowestTriethylaminetrihydrofluoride

concen-trationsinvolvedinthisstudy.

Anisole diffusion coefficient was estimated by

two methods and similar results were obtained:

(1.6_{× 10}−8_<D(m2_s−1_)<2.0×10−8_{). As for Et3N·3HF, anisole}

diffusioncoefficientis relativelyhighfor a relatively‘bigsized’ moleculeanditcouldbejustifiedbylowacetonitrileviscosity.

Somekineticparameters(˛nlsandk0′)ofbothanisoleand

tri-ethylaminetrihydrofluoride werealsoproposed.For theanodic

electronic transfer coefficient, in the case of both anisole and

triethylaminetrihydrofluoride,resultsledtothesameand repro-duciblevalue:˛nls∼0.36validatedbytwodifferentstudymethods.

Though, the value of the intrinsic heterogeneous electronic

transferconstant(k0_{)couldnotbecalculated,becauseitrequires}

additionalinvestigations(E0′

ofthesesystems).However,results

showsomevaluesoftheapparentintrinsicheterogeneous

elec-tronictransferconstantforbothanisole(k0′

a∼8×10−6ms−1)and

triethylamine trihydrofluoride (k0′

a in the range 4.86×10−23 to

4.52×10−18_ms−1_{).Thesevaluesarelow(eveniftheycontainan}

exponentialtermofthestandardpotential),indicatingirreversible systems;fortriethylaminetrihydrofluoride,theobtainedvalueis

verylowandvarieswithconcentration,becausethepH

modifica-tioncausedbythetriethylaminetrihydrofluoridedecomposition.

Preliminary results of preparative electrolyses, carried

out in conditions defined by the electrochemical kinetic

study, are encouraging [18,19], especially for fluorination of

dimethoxymethane,demonstrating thatelectro-fluorinationcan

beperformedinthealphapositionofaheteroatom.

Acknowledgement

WewouldliketothankMariaPilarGarcíaLeónforher

contri-butionsandveryhelpfuldiscussions. References

[1]V.Suryanarayanan,M.Noel,JournalofFluorineChemistry92(1998)177; N.Ilayaraja,M.Noel,JournalofElectroanalyticalChemistry638(2010)39. [2]Y.Hou,T.Fuchigami,ElectrochemistryCommunications1(1999)445. [3]Y.Hou,T.Fuchigami,TetrahedronLetters40(1999)7819.

[4]Y.Hou,S.Higashiya,T.Fuchigami,JournalofOrganicChemistry64(1999)3346. [5] D.Baba,H.Ishii,S.Higashiya,K.Fujisawa,T.Fuchigami,JournalofOrganic

Chemistry66(2001)7020.

[6]K.Suzuki,H.Ishii,T.Fuchigami,TetrahedronLetters42(2001)4861. [7]K.M.Dawood,T.Fuchigami,TetrahedronLetters42(2001)2513. [8] H.Ishii,Y.Hou,T.Fuchigami,TetrahedronLetters56(2000)8877. [9] N.Ilayaraja,M.Noel,JournalofElectroanalyticalChemistry632(2009)45. [10]N.Ilayaraja,S.Radhakrishnan,N.G.Renganathan,ChemistryandMaterials

Sci-ence16(2010)137.

[11] H.Löwe,W.Ehrfeld,ElectrochimicaActa44(1999)3679.

[12] W.Ehrfeld,V.Hessel,H.Lehr,ChemistryandMaterialsScience194(1998)233. [13]N.deMas,A.Gunther,M.A.Schmidt,K.F.Jensen,IndustrialEngineeringand

ChemicalResearch48(2009)1428.

[14] B.A.Shainyan,Y.S.Danilevich,RussianJournalofOrganicChemistry42(2006) 231.

[15]J.S.Jaworski,M.Cembor,M.Orlik,JournalofElectroanalyticalChemistry582 (2005)165.

[16] G.Dabosi,M.Martineau,G.Durand,Analysis6(1978)289.

[17] A.J.Bard,L.R.Faulkner,ElectrochemicalMethods,Fundamentalsand applica-tions,2nded.,JohnWiley&SonsInc.,NewYork,2001.

[18]C.Kane,T.Tzedakis,AIChEJournal54(2008)365.

[19] T.Tzedakis,C.Kane,A.Launay.‘Microréacteur’.Frenchpatentno.04.12305, 2004,PCT2006,No.WO 2006/053962;extendedtoUSA01122004no. 60/631,877.