Indian Journal of ChemistryVol. 4513. October 2006. pp, 2270-2280

Papers

Grignard reagent and Green Chemistry: Mechanistic studies to understand themolecular origins of selectivity in the formation of RMgX

Hassan Ilazimeh, Jean-Marc Maualia, Mircillc Atrolini, Nicolas Bodineau, Kishan HandooCaroline Murchi-Dclapicrrc, Erie Peralez & , 1iehel Chanon"

Laboraioirc AM3, Case 561, UMR CI\'RS (>I7~, Faculte des Sciences ct Techniques de Saint-Jerome. Univcrsitc PaulCezanne (Aix-Marscillc 111).133<)7 Marseille Ccdcx 20. France

E-Illai I: michcl.chanonticuniv.u-Jmrs.fr

Received 12 June 2()06

Controlling selectivity is one or the important keywords in the development of Green Chemistry. The mechanisticstudy of the formation of RMgX with the aim of understanding at which stcp(s) the selectivity is settled shows that there isroom 1'0[' basic research in a field too often considered as the simple development of new recipes. One of the most widelyused reactions in synthetic studies still offers fascinating conundrums when one tries to answer the apparently simplequestion: at which stcp(s) docs the formation of by-products coexisting with RMgX occur') I lero the illustration of thisstatement is given by studying the patterns of reactivity of aliphatic and aromatic radical clock halides toward magnesium orpotassium. The importance of using the concepts developed by elcctrochcmists to rationalize the selectivity observed atcathodes or anodes I'm undcrstanding selectivity observed in the reactive dissolution of metals is stressed.

Keywords: Green chemistry, Grignard reagent. mechanistic study

IPC Code: Int. CIS C071349

Since its discovery more than a century ago, theGrignard reagent has played a central role insynthesist. March's treatise, a great classic of organicchemistry, connects RMgX with the preparation ofthe following classes or compounds: acetals, alcohols.aldehydes, alkencs, alkyl halides, arencs, allcnes,amides, borancs, carboxylic acids, cnamincs, esters,imidcs, mercaptans, oximes, etc.". Even more strikingis the "evergreen" character or this reaction. Table Iselects some of the breakthroughs associated with thisreagent over more than a century. This table isremarkable within the perspective of green chemistry.Some opponents to the importance 01' looking atsynthetic chemistry under a greener perspective offerthe argument: chemistry is a mature science, there isnot much room left for improvements. If one searchesin the Chemical Abstracts data bank at the keywordsGrignard reagent one obtains 183 19 answers. Despitethe unavoidable conclusion that thousands or chemistsperformed this reaction, it clearly appears that onecentury was necessary to see the emergence of drasticinnovations.

The situation or understanding the mechanism orformation or RMgX is along the same line. The firstproposition that radical species could be involved as

intermediates in this formation was done by Gombergin 1927'.1. Since this date several prestigious teamshave invested much time and ingenuity to disentanglethe complexities of this mechanisrn'" ..1~. Themechanism round in the textbooks of chemistry isreported in Scheme I.

Tahle I - Griunard rcaucnt: an cvcrcrccn (jri~nard.[Clement,'1 Knolt.-l- i ormant." Villicras." l~it:kc.7 Ku7n'lda.'lmamoio." Knochel [0.[ '.

I ()OO Grignard discovers the rcugcru

1934 Clement: cutrainmcnt of aryl halides by alkyl halidesMc,C(,Mgl3r

195:1 Knott. Normant : TII I' as solvent for prcp.ui n~l'yMg13r, and vinylM~Llr

1967 Villieras: general approach to magucsium curbcnoidsby halogen magnesium cxchanpc

"1972 Ricke: highly reactive Mg metal: even PhMgr' maybe directly obtained

197:> Kumadu: cross coupling reactions or RMgX with RX

19~5 lmamoto: cerium chloride [or inducing selectively 1-:2addition of RMgX to ketones

20()O Knochel: halogcn-magucxium exchange reaction forpreparing Ionctionulizcd RMgX

HAZIMEH et 01.: (;RIGNARD REAGENT AND GREEN CHEMISTRY

.. 0+RX + \1 g .. R + X + Mg

'+ ..Mg + X .. \1 gX

R + MgX ... RMgX

Scheme I

When RX is an alkyl halide, the electron transfer isconcerted with the cleavage of the C-X bond, whereassome aryl halides radical anions are long lived enoughto be considered as intermediates rather thantransition states:". It is noteworthy that even recentlythis electron transfer mechanism has beenI II d,(,·17c la enge . . .Even if logical, this scheme does not account

directly for a certain number of observations. In termsof selectivity, one has to explain why the alkyl iodideswhich arc the best acceptors in the halide series arcgenerally the ones which give the lowest yields orRMgX. Along the same line the tertiary halides,which arc the best acceptors in the series primary,secondary, tertiary provide the lowest yields ofRMgX I.J. For the aryl halides, one would guess thatpara nitro substituted halides are going to yield easilythe corresponding ArMgX but they do not yieldeasily. One has to understand too, why very smallamounts of selected compounds can inhibit theformation of RMgX under certai n ci rcumstances".We will encounter, In the following, otherobservations which demand improvements orrefinements of Scheme l.Scheme II summarizes the principle of the tools

that we have used throughout our studies. Radicalclocks are specifically designee! organic halides. Thedesign includes an internal radical trap spatiallydisposed to trap intramolccularly as rapidly aspossible a radical if this one is formed on the routegoing from the substrate (here organic halide) to theproducts (here RMgX)""J. I f this trapping occurs, acyclized product will be formed as well as the mainexpected one which is linear. The presence of thiscyclic product constitutes a proof that at least part ofRMgX is formed via a route involving a radical. Torender this proof compelling one has to show thatanother intermediate (carbanion or carbonium) wouldnot produce the cyclic product under the conditions ofcxperirnent". On the other hand, the non observationof a cyclic product does not make it possible todirectly discard the participation of the radical in the

2271

formation or the linear compound. It may simplyoccur that, at the bifurcation point, the rate constant ofthe step competing with the intramolecular trapping ishigh enough to prevent any production of cyclizedproduct. For this reason, in the design of radicalclocks, the searchers have designed structures inwhich the intramolecular trapping (or, more generally,rearrangement) displays rate constants as high aspossible!'. This has been particularly true for theradical clocks designed to study biological problems.The label "radical clock" is associated with thestructures in which the rate constant of cyclization (orrearrangement) may be precisely measured. In thiscase, the ratio cyclized C/uncyclized U product tellsalso the rate at which the radical intermediate reacts togo toward the linear product. The intermolecular stepwhich competes with the intramolecular trapping Illaybe: coupling of the radical with another paramagneticspecies, atom transfer reaction (bi molecularhomolytic substitution), G- or fl-scission, addition onany double or triple bond, formation of an unevenbone! (I e-bond by reaction with a Lewis acid or :k-bond by reaction with a Lewis base, a special case (If

uneven bond beinz the bond between the radical and adiamaanetic met~~ surface 12), reducti vc or ox idati vcelectron transfer.

In Scheme I, the intermolecular step supposed tocompete with the intramolecular trapping is acoupling of the carbon centred radical with the MgXparamagnetic species situated either on the metalsurface or in the solution. It should be clear thatradical clocks provide information only about thesteps which form by-products via a radical route. It issupposed that the alkyl halides studied here do notreact with the formed RMgX under our experimentalconditions. This supposition could no longer hold forbenzylic or allylic halides.J3.J5. The limits of theradical clocks methodology to disentanglemechanistic complex ities have been cri tical! ycI isc usscd"'.

The aim of this report is to extract information frommechanistic data to gain insights on some of thefactors ruling the selectivity in the formation ofRMgX. To do so, we will first show how the use ofalkyl halide types of radical clocks led us to the studyof aromatic halide types of radical clocks. Theinsights gained on these aromatic radical clocks willthen provide fresh views 10 return to factors possiblyplaying a role in the selectivity of the Grignardreagent formation from alkyl halides. To accelerate

2272 INDIAN J. CHEM., SEC 13,OCTOBER 2006

o-.~k4

X abstractionor addition to».> generalized Acid,

/ Base or ElectronTransfer (+ or·).:

k2 j

kproradical ~

0-'x - I""""'"

X abstraction (X = halogen, H. etc.)or Addition to generalized Acid. Base (A,B)or Electron Transfer ( + or· )

--------~~---------( '\

C= C..._____X Il

! 1products products

the reading of readers more interested into yields thaninto mechanisms, the parts specifically dealing withyields will be printed in italics.

Results and Discussion

Alkyl halides radical clocksBickelhaupts group was the first group to use

radical clocks in the study of Grignard reagentformation mechanism. Scheme III shows that cyclicproducts are indeed formed when the classical radicalclock 6-bromo-l-hexene (key = 2.3 tOS

S·I, 2S °C)reacts with magnesium in THP7. The amount ofcyclized product is, however, drastically lower thanthe quantities observed when this radical clock reactswith BU3SnH48. This very low amount was first takenas an evidence converging with the Kharasch-Walborsky hypothesis that the carbon centred radicalCR' in Scheme I) is adsorbed on the metal surface as

.. f d144<J Thi d . h Isoon as It 15 orrne '. IS a sorption on t e metasurface would prevent the cyclization. The cyclizedproduct would result from an equilibrium adsorbedversus "free" carbon centred radical generally shiftedtoward the adsorbed species. A variation of thisrepresentation would be in cage (the cage includingthe metal surface) versus out of cage reactivity.Another representation was later proposed by Garst.

i··',11:

Scheme II

This author, extending the theoretical treatments ofdiffusion at the interface solid-liquid:", showed thatthe observed small amount of cyclized and dimericcompounds do not demand the intervention ofadsorptionSl.52. The controversy on this problem issummarized in two Accounts and an useful discussionof this issue may be found in van Klink PhOs3

.5s

, It isremarkable that the intervention or non intervention ofadsorption of radical species on the metallic surfacesto rationalize the overall selectivity has given rise tothe same kind of controversies. A good illustration ofthis point is provided by a look at some of the reportsdealing with the mechanism of Kolbe reaction"?'.

On our side, we synthesized a set of radical clocksdisplaying two characteristics: I) they were cyC\izingfaster (key = 107 S'I, 6S "C) than the 6- bromo-l- hexeneone, 2) the series chloro, bromo, iodo was examined'".The second point could have been considered as awaste of time. Indeed, if one returns to Scheme I, itappears that the competition intramolecular trappingof the radical versus its intermolecular reactionsshould not, at first sight, depend on the nature of X.Actually, the experimental results gathered InScheme IV show that this expectation was notfulfilled. Everything was going on as if a kind ofmemory effect was operating.

• I

. ;

HAZIMEH et al.: GRIGNARD REAGENT AND GREEN CHEMISTRY 2273

Br

~

MaBr 6gB

'~ +

85.5 2.5

86.0 2.5

DEE

THF

DEE 5.8

RRIOlal

THF 2.1

2.5 0.7 9

viscosity in cP

DEE: diethyl ether 0.19

THF : tetrahydrofuran 0.39

Yields are given in percents.

x C V C/U VIC

1.2 0.7 4

i 56 44 1.3 0.8

Scheme III

second case was reported by Bunnett and thenconjointly studied by this author with Beckwith usingaromatic centred radical clocks on reaction with thesolvated electron (KiNHit-BuOH)67.68. Their resultswere later reinterpreted by Andrieux and Saveant'".Scheme V shows that for this series of radical clocksthe highest amount of cyclized product is obtainedwith the iodo substrate.With X = Br and J, 17-18.5%of the dimer of the cyclized radical are formed (notconsidered in Scheme V). This order fits with theorder seen in Scheme IV.

As a working hypothesis, we first applied theconcepts proposed by Andrieux and Saveant torationalize the changes in selectivity reported inScheme IV. To do so, one had to admit that, in THF,the magnesium surface was able to produce solvatedelectron7o. The concentration of solvated electronwould rapidly decrease from the metal surface to thebulk solution. There would, therefore, be a gradient ofconcentration for the solvated electron. When the

0.6 1.8Br 36 64

CI 15 85 0.2 5.7x

Yields are given in percents.Scheme IV

Such a kind of memory effect had been reported inreactions involving electron transfer. The first casewas the SRN1 mechanism'r''?'. Kornblum had observedthat in a competition between the electron transferroute and the classical SN2 displacement, for paranitro-benzylic halides, the chloro substrate wasyielding the highest amount of radical reaction65.66.The opposite order is shown in Scheme IV. The

2274 INDIAN J. CHEM., SEC B, OCTOBER 2006

X C U C/U ((f l)Mg/THF OJ~IX~

58 8 82)W

Br 72 10 7.5 X = Cl, Br, I

Cl 78 18 4.4Yields are given in percents. Scheme VI

Scheme V

halide radical clocks would approach the metalsurface, the best electron acceptor (iodide) would bethe first to react with the solvated electron. Thepoorest electron acceptor (chloride) would have to gonearer the metal surface to give birth to the carboncentred radical. This carbon centred radical beingborn in a medium richer in solvated electron would betransformed more rapidly into a carbanion than theradical produced by the iodo substrate. As thecarbanion cyclizes far slower than the radical does,the apparent memory effect was rationalized.

This explanation was not quite compelling. Therewas no precedent in the literature for the presence ofsolvated electron in the vicinity of a magnesiumsurface submerged in THF. We decided to study thepattern of reactivity displayed by the radical clocksstudied by Bunnett and Beckwith when reacted withmagnesium in THF.

Aryl halide radical clocksScheme VI shows an unexpected result. When the

aryl halide radical docks react with magnesium inTHF almost no cyclized product is observed". This isso for the three different halides. This result isunexpected for several reasons: the first one is that thearyl radical has been measured to cyclize at least 10times faster (key= 4.0 108 s', 30De) than the alkylradicals studied in the preceding paragraph(Scheme IV) 71,72.If no other parameter intervenesone should expect a ratio cyclizedllinear higher thanthe one shown in Scheme IV. These unexpectedresults were confirmed by Garst's team". They foundthe same very low amount of cyclized product inTHF; this amount increased slightly when the reactionwas performed in diethyl ether. This team brought aquantitative plus in their report. Garst had been ableto propose a model accounting for the observedselectivity in the reaction of Mg with 6-bromo-l-hexene. Applying this model to the aryl radical clockshe expected more than 70% of cyclized productformation.

Why so little cyclization? Returning to Scheme Ithe simplest explanation would be to suppose that arylradicals have the possibility to react about 100 timesfaster than alkyl radicals with MgX radicals. Such ahypothesis has no special support from the literature.A more attractive explanation was offered by adiscovery of Bickelhaupt's group". To explain aconsistent and extensive set of rearranged by-productswhen selected aromatic halides reacted withmagnesium, this group had proposed that one of thereactions possibly competing with the intramolecularcyclization of the aryl radical was its reduction at themetal surface31. Aryl carbanions would then be on theroute aryl halides to ArMgX. Aryl radicals are knownto be far better oxidizing agent than alkyl ones 75.Within Bickelhaupt's proposition, the reason why thearyl radical clocks are yielding so little cyclizationcould be that they accept an electron from the metalsurface faster than they cyclize.

This was the explanation that we favoured", Forthe same observation, Garst proposed another one.One important difference between aryl and alkylhalides is the average lifetime of their radical anions.This lifetime is definitely longer for aryl halideradical anions although some of them could cleavewith rate constants in the range of those proposed foralkyl halides". Based on this point, Garst proposedthat these aryl radical anions are further reduced intodianions on the route toward ArMgX73. Thesedianions (intermediates or transition states) wouldthen cleave into an aryl carbanion and a halide anion.We have discussed elsewhere the reasons why we donot find this proposition compelling78.

One of these reasons is that, among the numerouspublications dealing with the reactivity of aryl halidesat a cathode only one proposes that dianions could beinvolved". Pulse radiolysis studies confirm thisgeneral trend'". Electrochemical studies of arylhalides provide, furthermore, a fresh view of thesuccession of events which rule the fate of an arylradical formed in the vicinity of the cathode when arylhalides are electrolysed. Saveant and Amatore

JJ

HAZIMEI! et al.: GRIGNAR]) REA(,ENT AN]) (iREEN CHEMISTRY

x-x ' Ar + X·Ar X +

Ar + SH Arl-I + S

step equivalent to radical cycli/.uion

cAr +

SH = solvent

Scheme VII

proposed a detailed model to describe what occurs inthe vicinity of a cathode when an aryl halide issubmitted to electrolyse~I.S(l. The succession of stepsis shown in Scheme VII.

The physical model partly overlaps with Garst'sdiffusion lllodefl.S7. The aryl halides radical anionswould be produced ill a thin layer (about 10 A) ofliquid covering the electrode. Thanks to their lifetime,they could diffuse away toward the bulk. During thisdiffusion, they would cleave into aryl radicals andhalide anion. The longest lived radical anions(chlorides) would cleave at the greatest distance fromthe metal surface; therefore they would have a smallerprobability of diffusing back to the metal surfacewhere they are transformed into carbanions. Thesecarbanions reacting with MgXc or MgX+ would yieldRMgX"'<)o The mechanism shown in Scheme Iwould therefore be replaced by the one shown inScheme VIII.

Thus. in contrast with the solvated electron modeldescribed in the preceding paragraph, the arylchloride radical clocks should yield the highestamount of cyclizcd product. Expressed in terms ofyields of RMgX this means that when the radical by-products were expected to be more important for thealkyl iodides in the series of halides, the reverseshould be true for the aryl halides. This statementcould be weakened because subtle di Iferencc revertsthis order. S.1.') I The electrochemical model leads to afresh view of the factors affecting the overallselectivity in the preparation of ArMgX. Because thefate of the aryl radical critically depends upon thedistance from the magnesium surface where it isformed, the overall selectivity amounts to an averagedvalue of a series of little flasks. Tn every flask, aspecific selectivity is settled as shown in Figure 1.For the aryl halides that we studied, the flask in closeproximity to the surface weights the most in theaverage ruling the overall selectivity. This is because

»

2275

. .. 0+RX + \1g .. R + X + \'lg

"+ ++ ..Mg + R .. :-.tg + R

++ ..\lgX+Mg + X •.

..MgX+R + .. R:-"I~X

Scheme VIII

the radical anions of the studied radical clods areshort lived'):'. Furthermore. we have shown that thereducing power developed in the vicinity of thesurface (heterogeneous electron transfer) may hehigher than the one obtained in the bulk(homogeneous electron transfer) even when thereducing reagent is the solvated clcctron". In short,the main reason why less cyclizcd products wereformed with the aryl radical clocks than with theslower alkyl radical clocks seems to he the highoxidizing power of the aryl radical.The mechanism proposed in Scheme VIII has

several consequences in terms of yields of ArMgXalthough the description that we gl\'e hereconsiderably oversimplifies the electrochemicalmodel'". The structure of the radical clock, theviscosity of the solvent. thc rate of cyclizarion of theradical, play also a role in ruling the ratio cyclizcd/linear compounds. We discuss these refinements Inanother rcport'" and a full treatment is given inArnatorcs review') I.Scheme VIII helps in understanding why ethers are

the best solvents for preparing GrignaI'd reagents. Ifnaked carbanions are intermediates on the routetoward RMgX. any species able to protonaic themwill compete with their reaction on MgX:,. Tables ofpK" values show that among the avuilahlc solvents.ethers arc the rare ones with a donor number hiuhcrthan alkanes displaying a very low acidity". ~

Several explanations were proposed to explain theeffect of entrainment used to prepare ArMgXproducts which could not be obtained by the simpleGrignard procedure <)5.%. Garst provided an originalone 60 years after the discovery of this effect. Theaddi tion of dibrornocthanc or any rcacti vc alky I hal ideto the medium would have. as main consequence. toenrich the medium in MgX:, '". This salt would haveseveral beneficial effects. One would be to increasethe polarity of the medium which accelerates the

0Selectron transfer steps'. The second would be to

This detail is: keep an excess of the MgX2 salt withrespect to the reducing alkali metal. This excess will,then, play its entrainment role in the step "preparationof the Grignard reagent". Otherwise, some of thehighly basic carbanion could abstract a proton fromthe medium. One remarkable feature of Rieke'smagnesium does not seem to be directly rationalizedwithin the framework of Scheme VIII. This feature isthe beneficial salt effect associated with the additionof potassium iodide to the solution 109. A tentativeexplanation could have to do with the corrosionmodel of Grignard reagent formatiorr". In this model,the metal surface is divided into coupled anodic andcathodic sites. The anodic dissolution of magnesiumcould be catalyzed by IK in way reminiscent of thesolid-liquid phase transfer catalysis'!", This pointdeserves further study using radical clocks.With this fresh view expressed in Scheme VIII and

Figure 1 it is tempting to return to alkyl halide radicalclocks and check if this scheme could be extended tothese substrates.

2276 INDIAN J. CHEM., SEC B, OCTOBER 2006

NtAIITHtlullrllcr TWO·"M V.T"U. COr.CtNTIIATtO"r.OINA NrOIUNOf VrlY "1O"II"4T&l"Il-=:&.0 UCIr.O '0 tjIt" TH,I" Tit An sro u,c "'TI01"4INTO A"," l CAIt tA" 01"4:500NIN •••T.1 TH' l.'AC "'VtTY.

'" err ' ••••11 THe II fit 0" TNt lUll' ACe TH t III CONCrNTI. •••TtON a lTia 1Mto ItTANTfUT TN' II:fOUCtlOO '0101I11 Of TH. NaOlU ••• 11 lTItONCHY ONlNtlHtOON U 11ATI0". eye LILA'" ION 0' TN f LAOICA l C LOCJ S••• ATO••• TltAN:!;r, Il MAYrv."I' "CO~ll""""'~ I MH I

Figure I-Fate of the radicals depending upon the place where they were born

simulate the effect of supporting salts inelectrochernistry'Pl'". The third would be to increasethe viscosity of the medium, slowing down thediffusion away of the aryl halide radical anion'". Thefourth, but not the least, would be to increase theconcentration of species able to react rapidly with thecarbanion to yield RMgX. This this effect" is usedcleverly to obtain quantitative yields of ArMgBr97.The direct preparation of C6HsMgF from the

reaction between fluorobenzene and magnesiumremained a challenge for 72 years 106. ActivatedRieke's magnesium made it possible'. What is sospecific about this activated magnesium? We believethat it displays several activating features within thescope of Scheme VIII. Hie first one is that the size ofthe metal particles obtained by reduction of an MgX2salt is very small. We checked by ESR studies that thereductive strength of Rieke's magnesium iscomparable to that of magnesium particles obtainedby metal vapour synthesis. Both being distinctlyhigher than the one associated to magnesiumtumings'l", This increased reducing strength couldhave to do with two main factors; the first one is thevery small size of particles which increases thesurface of contact solid-liquid, the second one is thatthe freshly prepared surface of these particles isunpolluted in comparison with usual magnesium

. 108 I hi . h h fturnings .. n t IS perspective, t e c ances 0

reducing fluorobenzene are increased. This,furthermore, holds for the chances of reducing theproduced aryl radical into an aryl carbanion on theroute toward RMgX. The second activating feature istruly, beneficial only if one experimental detail isrespected in the preparation of Rieke's magnesium.

Return to Alkyl Halide Radical ClocksThe cathodic reactivity of 6-bromo-l-hexene in

THF was not reported in the literature. Collaborationwith Combellas-Kanoufi group allowed filling thisgapill. The radical clock studied in the Grignardreagent formation by Bickelhaupt (Scheme III), whenreduced at a carbon cathode in THF as a solvent,yields a slightly higher yield of cyclic product (9 %).This hints at a certain similarity between thesuccession of events occurring near the cathode andnear the magnesium surface. We discussed the limitsof this similarity elsewhere III. In the present

HAZIMEH et al.: GRIGNARD REAGENT AND GREEN CHEMISTRY

discussion we stress only the points which arerelevant to our objective: understand some of thefactors at work in ruling the yields of RMgXformation.

Thinking only in terms of lifetimes of the halideradical anions would rule out the participation ofScheme VIII in the formation of alkyl halideGrignard reagents. Indeed, the electron transfer toalkyl halides is concerted with the production of alkylradicals 112. Therefore, the formed alkyl radicals areborn in a very close vicinity of the metal surface. Thiswould suggest that they have a better chance of beingreduced (Figure 1) rapidly; which means that lesscyclized product should form in comparison with arylhalide radical clocks. The opposite trend isexperimentally observed (compare Schemes IV andVI) despite the worsening fact that the rate constant ofcyclization of the aryl radical is higher than the onesof alkyl radicals'13,1'4. Saveant's group measured therates of reduction at a cathode of several alkylradicals'Y. They found that these rates are lower thanexpected probably because of the structuralreorganization demanded to pass from the quasiplanar carbon centred radical to the correspondingpyramidal carbanion. If Scheme VIII is to beextended to the alkyl halides one could have toconsider that in the fight between the opposite factors1 and 2: 1) lifetimes of radical anions and rates ofcyclization on one side, 2) respective rates ofreduction of alkyl and aryl radicals on the other, thesecond wins the competition, .

In the framework of Scheme VIII, this lastproposition could be only part of the explanation.During the cyclic voltammetric study of 6-bromo-l-heptene, Schmit discovered the interestingautocatalytic process shown in Scheme IX"5

RX

lK e

X 00

••R 0

R 0 + X

RX R 00

e

Scheme IX

2277

In this process, some of the carbanions formed inthe vicinity of the cathode, could act as reducingagents toward the RX molecules diffusing from thebulk toward the cathode (step 3). The consequence ofthis autocatalytic process would be to increase therelative quantity of radicals. Such an increase wouldbe associated with more cyciized products for alkylhalide radical clocks and more radical by-products inthe preparation of the Grignard reagent of thecorresponding alkyl halides. The best chances ofhaving an efficient step 3 would be for R carbanionswith good reducing properties and RX with goodacceptor properties. On the side of carbanions, thiscondition discards aromatic carbanions with respect toalkyl ones". On the side of alkyl halides, the bestacceptors are the iodides and, for the carbanions Rinvolved in step 3, the best reducing carbanions wouldbe the tertiary ones 75. This argument should be refinedif one remembers that too different rates of steps 2and 3 lower the turnovers. The systems leading to thebest turnovers are expected to display the highestquantities of cyclized products for alkyl halide radicalclocks and autocatalysis could contribute inexplaining the leaving group effects shown inScheme IV.

On the side of yields, the autocatalytic schemecould explain why alkyl iodides generally yield loweryields of RMgX than the corresponding chlorides 14. Inthe same line, tertiary halides give lower yields thanprimary ones. In a careful old work, Gilman's groupperformed an interesting series of experiments. For agiven halide, they would try to prepare RMgXaccording two procedures 116. In the first one, the alkylhalide was added very rapidly to a suspension ofmagnesium in diethylether; in the second, the halidewas very slowly added to the same suspension.Clearly, the first procedure corresponds to a higherconcentration of RX in the vicinity of the metal. The

Table II-Effect of the rate of halide addition on RMgX yield

Halide Slow addition Fast addition Difference(%) (%) (%)

EtBr 93 87 6i-PrBr 84 70 17t-BuBr 25 18 28Il-BuCI 91.2 90.6 IIl-BuBr 94 79 16Il-BuI 85.6 67.8 21

Difference = [(Slow addition - Fast addition)/Slow addition] x100

2278 INDIAN J. CHEM., SEC B, OCTOBER 2006

first procedure favours the autocatalytic participation.These authors observed no significant difference inRMgX yields between the two procedures when thesubstrates were alkyl chlorides. In contrast, when analkyl iodide was submitted to this set of experiments,a clear difference in yields was observed between thetwo procedures. In the series EtBr, II-BuBr, t-BuBrthe largest difference was found for the third one(Table II). The convergence is rather striking. To bemore compelling, one would have to check that thereported differences cannot be rationalized in theframework of a reaction between the startingsubstrates and RMgX, although it is generallyaccepted that, for the alkyl halides present inTable II, this reaction is slow'':', but it is known thatin the SN2-E2 competition, the importance of E2elimination is in the order I>Br>0117. Also, severalexamples of the Grignard reagent acting as a basehave been reported 118·120.

ConclusionWe started this presentation with an apparently

simple aim: establish connections between yields(selectivity) and the possible mechanisms for one ofthe most used reactions in organic chemistry. Evenwithout going into the details some very generalremarks emerge. The first one is that the science ofreactions is a very slow science; it took respecti vely34, 72, \00 years to pass i) from the alkyl halides tothe aryl ones in the preparation of Grignard reagents,ii) from relatively acti ve magnesium to a reallyreactive one, iii) from poorly functionalized substratesto functionalized ones. Grignard reaction is not anexception. Just consider the number of years whichseparate the discovery of the Diels Alder reaction andits first intramolecular extension to total synthesis. Inthe perspective of a Greener chemistry, it would beimportant to examine the question: How could weincrease the pace of innovation in this field 121·123?The second one is that simple guesses, not supportedby in depth experimental studies, may be hazardous.Seeing that aryl halides were, at first sight, less easyto transform into ArMgX than alkyl halides couldhave suggested that the yields of ArMgX would ingeneral lag behind AlkylMgX. The reverse isprobably true73. If the autocatalytic hypothesiscontributes in the explanation of higher quantities ofradical by-products going from alkyl fluoride to alkyliodides, one could again meet a counter-intuitivesituation. The best yields of alkylMgX could be, afteroptimization 124,obtained from the tluoroalkyl.

The gap between the various components ofchemical science is, one more time, illustrated by thispresentation. To better understand the roots ofselectivity of Grignard reagent formation, the toolswere available in Electrochemistry since 1980, theywere not used as such for more than 25 years. It isnow clear that every progress done in themodelisation of the molecular events occurring in thevicinity of the cathode or the anode is relevant to theoptimization of the yields of reactions involving thereactive dissolution of a metal in solution.

There is still a long way to reach a satisfactoryunderstanding of this chemistry of reactivedissolutions. This is the reason why we finish with aseries of question marks. What are the consequencesof the corrosion-like approach of Grignard reagentformation:" ? Why traces of inhibitors may stop areaction which is not a chain reaction? Why theinhibitor dioxygen inhibits some Grignards but doesnothing to others? What is the molecularrepresentation of what is called active sites on themetal surface? Why do the shape and density of pitsobserved at the magnesium surface depends on thenature of X for various RX 125?

ReferencesI Grignard V, C R Nebd Seances Aead Sci, 130,1900, 1322.2 March J, Advanced Organic Cheniistrv. Reaction»,

Mechanisms 111/1/ SII'I/CIIII'e; 3rd edn .. (Wiley and Sons, NewYork). 1985.

3 Clement H, Compt Rend, 198,1934,665.4 Knott R F, Ciric J & Breckenridge J G, CIIII ./ Chern, 31.

1953,615.5 Normanr H. Comp Relld, 239, 1954, 1510.6 Villieras J, /Jill! Soc Chilli FI', 1967, 1511.7 Riekc R I) & Hudnall PM, .1/\111 Chem Soc. 94, 1972, 7178.8 Kumada M, Taillao K. Sumitani T & Kis() Y, Kcnkvu

110"0"11 . /vsahi CO/'ll.l'II Kogvo Gijutsu Shoreikai. 22, 1973.171.

9 Irnamoto T, Pure /vppl Client. 62, 1990, 747.10 koulandcr M, Boymond L, Bcrilton L Lcprctrc A. Varchi

G. Avolio S, Laaziri W, Queguiner G. Ricci A, Cahiez (J 8:Knochel P, cu». EIII'./, 6, 2000, 767.

II Knochel 1', Dohle W, Gornmcnnann N, Kneisel F F, KoppF. Koru T, Sapounrzis I & Vu V A, I~ngell' Chcni 1111Ed. 42.20()3.4302.

12 Lee J-S, Velarde-Ortiz R, Guijarro A, Wurst J R & Rieke RD, .I Org ChC/lI, 65, 200(), 5428.

13 Gomberg M & Bachman W E. J Alii Chen: Soc. 49, 1927,236.

14 Kharasch M S & Reinrnurh 0, Grigtiard Reactions or N(}//metallic Substances; (Prentice-Hall, New York) 1954.

15 Walborsky H M & Young A E, J ;\111 Chern Soc. S6, 1964.3288.

16 Vink 1', Blomberg C, Vreugdenhil A D & Bickelhaupi F. JOrganoinct ChClII, 15, 1968,273.

HAZIMEH et al.: GRIGNARD REAGENT AND GREEN CHEMISTRY

17 Bodewitz H W H J, Blomberg C & Bickelhaupt F,Tetrahedron Leu, 1972,281.

18 Rogers R J, Mitchell H L, Fujiwara Y & Whitesides G M, JOrg Chern, 39, 1974, 857.

19 Bodewitz H W H J, Blomberg C & Bickelhaupt F,Tetrahedron Lett, 24, 1975, 2003.

20 Dubois J-E, Molle G, Tourillon G & Bauer P, TetrahedronLeu, 52, 1979, 5069.

21 Lawrence L M & Whitesides G M, J Alii Chem Sac, 102,1980,2493.

22 Walborsky H M & Banks B, Bull Sac Chilli Belg, 89, 1980,849.

23 Molle G & Bauer P, J Alii Chem Sac, 104,1982,3481.24 Garst J F & Swift B L, JAm Chem Sac, 111,1989,241.25 Walborsky H M & Rachon J, J Alii Chem Sac, Ill, 1989,

1896.26 Koon S E, Oyler C E, Hill J H M & Bowyer W J, J Org

Chern, 58, 1993, 3225.27 Bickelhaupt F, J Organomet Chem, 475, 1994, I.28 Hamdouchi C & Walborsky H M, in Handbook of Grignard

Reagents; edited by G S Silverman & P E Rakita, (Dekker,New York) 1996, pp 145-218.

29 Solovev V N, Sergeev G B, Nemukhin A V, Burt S K&Topoll A, J Pliys Cliem A, 101,1997,8625.

30 Sergeev G B & Shabatina T I, Surf. Sci., 500, 2002, 628.31 van Klink G P M, de Boer H J R, Schat G, Akkerman 0 S,

Bickelhaupt F & Spek A L, Organometallics, 21, 2002,2119.

32 Renou F, Mostafavi M, Archirel P, Bonazzola L & PernotP, J Phys Chem A, 107,2003, 1506.

33 Garst J F & Soriaga M R, Coor Chem Rev, 248, 2004, 623.34 Ashby E C & Oswald J, J Org Chem, 53, 1988, 6068.35 Save ant J M, in Advances ill Physical Organic Chemistry;

edited by T T Tidwell, (Academic Press, New York) 2000;Vol. 35. pp 117-192.

36 Beals 13 J, Bello Z I, Cuddihy K P; Healy E M, Koon-Church S E, Owens J M, Teerlinck C E & Bowyer W J, JPh),s Chern A, 106,2002,498.

37 Jasien P G & Abbondondola J A, Theochem-J Mol Struct,671. 2004, III.

38 Peralez E, Negrel J C & Chanon M, Tetrahedron u», 35,1994,5857.

39 Griller D & Ingold K U, Acc Chem Res, 13, 1980, 317.40 Newcomb M, Shen R, Choi S Y, Toy PH, Hollenberg P F,

Vaz AD N & Coon M J, J Alii Chent Soc. 122,2000,2677.41 Newcomb M & Toy PH, Acc Cliem Res, 33, 2000, 449.42 Bent B E, Chern Rev, 96,1996, 1361.43 Kundig E P & Perret C, Helv Chilli Acta, 64, 1981, 2606.44 Oppolzer W, Kundig E P. Bishop P M & Perret C.

Tetrahedron Lett, 23, 1982, 3901.45 Hill E A, In Grignard Reagents: New Developments; edited

by H G Richey, (Wiley,Chichester) 2000, pp 27-64.46 Newcomb M & Curran D P, Acc Chem Res, 21, 1988. 206.47 Bodewitz H W H J, Blomberg C & Bickelhaupt F,

Tetrahedron, 31,1975, 1053.48 Beckwith A L J & Moad G, J Chem Soc-Cliem COIIIIIIIIIl.

1974,472.

2279

49 Walborsky H M & Young A E, Baskerville Chem J, 14,1965, I.

50 Garst J F, J Phys Cliem, 84, 1980, 1995.51 Garst J F, Deutch J E & Whitesides G M, J Alii Chem Soc,

108, 1986, 2490.52 Garst J F, Swift B L & Smith D W, J Alii Chern Sac, III,

1989,234.53 Garst J F, Acc Chem Res, 24, 1991,95.54 Walborsky H M, Acc Chern Res, 23, 1990,286.55 Van Klink G P M, Thesis of Free University, Amsterdam,

1998.56 Hawkes G E, Utley J H P &Yates G B, J Chem Soc-Perkin

Trans, 2, 1976, 1709.57 Utley J H P & Yates G B, J Chern Sac-Perkin Trans, 2,

1978,395.58 Vassil'ev Y B & Grinberg V A, J Electroanal Chein, 336,

1992,281.59 Grinberg V A & Vassiliev Y B, Russ J Electrochem, 32,

1996,281.60 Nonaka T & Fuchigami T, in Organic Electrochemistry; 4th

Edn., edited by H Lund & 0 Hammerich, (Marcel Dekker,New York) 2001, pp 1051-1102.

61 Andrieux C P, Gonzalez F & Saveant J M, J ElectroanalChem, 498, 2001, 171.

62 Peralez E, Negrel J C & Chanon M, Tetrahedron, 51, 1995,12601.

63 Rossi R A, Pierini A B & Penenory A B, Chem Rev, 103.2003,71.

64 Rossi R A & Postigo A, Curr Org Chem, 7, 2003, 747.65 Kerber R C, Urry G W & Kornblum N. J Alii Cliem Soc, 86.

1964,3904.66 Kerber R C, Urry G W & Kornblum N, J Alii Cliem Soc. 87,

1965,4520.67 Bard R R, Bunnett J F, Creary X & Tremelling M J. J Alii

Clietu Soc, 102,1980,2852.68 Meijs G F, Bunnett J F & Beckwith A L J. J Alii Clietn Sac,

108.1986,4899.69 Andrieux C P & Saveant J M. J Alii Choll Soc, 115, 1993,

8044.70 Chanon M, Negrel J C. Bodineau N. Mattalia J M, Peralez E

& Goursot A, Macromol Symposia, 134, 1998. 13.71 Abeywickrerna A N & Beckwith A L J, J Chem Soc-Cliem

COIIIIIIIIIl, 1986, 464.72 Johnston L J, Lusztyk J, Wayner D D M, Abeywickreyma A

N. Beckwith A L J, Scaiano J C &Ingold K U, J Alii ChelllSoc, 10"1,1985,4594.

73 Garst J F, Boone J R. Webb L, Lawrence K E, Baxter J T (~Ungvary F, lnorg Chilli Acta, 296, 1999,52.

74 De Boer H J R, Akkerman 0 S & Bickelhaupt F, Angell'Chern, Int Ed Ellg/, 27,1988,687.

75 Fu Y, Liu L, Yu H-Z, Wang Y-M & Guo Q-X, J Alii ChernSoc. 127,2005,7227.

76 Bodineau N, Mattalia J M, Thimokhin V, Handoo K, NegrclJ C & Chanon M, Org Lett, 2. 2000, 2303.

77 Pause L, Robert M & Saveant J M, J Alii Chem Soc, 121,1999,7158.

78 Hazimeh H, Mattalia J M, Marchi-Delapierre C, Barone R.Nudelman N S & Chanon M, J Pltys Org Cheni, 18, 2005.1145.

2280 INDIAN J. CHEM .. SEC 13, OCTOBER 2006

79 Yanilkin V V, Berdnikov E A & l3uzykin 13 r, Russ .IElcctrochem, 36, 2000, 128.

80 Takeda N, Poliakov P V, Cook A R & Miller J R, J AmChern Sac, 126, 2004, 430 I.

81 Arnatore C. Thesis of University Paris 7. Paris, 1979.82 Amatore C, Saveant J M &Thiebault A, J Electroanal Chern

Interfacial Electrocliem, 103, 1979,303.83 Amatore C &Saveant J M, J Electroanal Chern Interfacial

Electrocliem, 102, 1979, 21.84 Arnatore C, Chaussard J, Pinson J, Saveant J M &Thiebault

A, JAm Chern Soc, 101,1979,6012.B5 Arnatore C, in Organic Electrochemistry: An introduction

and a Guide; 3rd Edn .. edited by M M l3aizer& H Lund, (MDekker, New York) 1991, pp 207 -232.

86 M'Halla F, Pinson J &Saveunt J M, JAm Chem Sac, 102,1980, 4120.

B7 Garst J F & Ungvary F, in Griguard Reagents: NewDevelopments; edited by H G Richey, (Wiley, Chichester)2000, pp 185-276.

88 Bank S &Bank J F, Tetrahedron Lett, 1969,4533.89 Bertini F, Grasselli P, Zubiani G & Cainelli G, Tetrahedron,

26, 1970, 1281.90 Lund H, Svith H. Pedersen S U & Daasbjerg K, Electrochim

Acta, 51,2005,655.91 Arnatore C, in Organic Electrochemistry; 4th Edn., edited

by H Lund & 0 Harnrnerich, (M Dekker, New York) 2001,pp 183-205.

92 Saveant J-M In Advances ill Electron Transfer Chemistry;edited by P S Mariano, (JAl Press Inc .. Greenwich) 1994;Vol. 4, pp 53-116.

93 Hazirneh H, Kanoufi F, Cambell as C, Mattalia J M, Marchi-Delapierre C & Chanon M,(unpublished).

94 Bordwell F G, Acc Cliem Res, 21,1988,456.95 Urion E, Compt Rend, 198,1934,1244.96 Grignard V, CR Hebd Seances Acad Sci, 198,1934,2217.97 Garst J F, Lawrence K E, Batlaw R, Boone J R &Ungvary

F, Inorg Chilli Acta, 222, 1994,365.98 Santamaria J, in Photoinduced Electron Transfer; edited by

M-A Fox & M Chanon, (Elsevier, Amsterdam) 1988, Part.13, pp 483-540.

99 Gritzner G, J Solid State Electrochem, B, 2004, 757.100 Oldham K 13, Cardwell T J, Santos J H & Bond A M, J

Electroanal Chem, 430,1997,39.

101 Khoo S 13, Foley J K & Pons S, J Electroanal Chernlnterfacial Electrochem, 215, 1986,273.

102 Kalinowski M K, Cliem Phys Lett, 7, 1970, 55.103 Perichon J & Buvet R, Bull Soc Chim Fr, 1968. 1282.104 Nath A & Bhattacharya A K, J Indian Che/II Soc. 41, 1964,

435.105 Vlcek A A, Collect Czech Client Conunun, 20, 1955.400.106 Ashby E C &Yu S H, J Org Chern, 36.1971,2123.107 Handoo K, Garcia G V, Negrel J C, Bodineau N, Marchi-

Dclapierrc C, Mattalia J M & Chanon M, (unpublished).108 Abreu J B, Soto J E, Ashley-Facey A, Soriaga M P, Garst J

F & Stickney J L, .I Colloid Interface set. 206, 1998, 247.109 Ricke R D & Bales S E, J Client Soc-Chem COIllIII1111, 1973.

879.110 Brarn G, Loupy A & Sansoulet J, lsr J Chelll, 26.1985,291.III Kanoufi F, Combellas C, Hazirnch H, Mattalia J M, Marchi-

Delapicrrc C &Chanon M, J Pltys Org Chem, 2006. inpress.

112 Andrieux C P, Gallardo I & d Saveant J M, J Alii Chern SOl',111,1989, 1620.

113 Newcomb M, Acta Chern Seand, 44. 1990,299.114 Annunziata A, Galli C, Marinelli M & Pau T, Eur J o-«

Cliem, 2001, 1323.115 Schmit J Lrn, Thesis of University of North Dakota, Grand

Forks, 1992.116 Gilman H, Zoellner E A & Dickey J B, .I Alii Cliem Soc, 51,

1929, 1583.117 March J, Advanced Organic Chemistry. Reactions,

Mectianistns and Structure; 5th Edn., (Wiley and Sons, NcwYork) 200I, p. 1321.

us Cameron D W & Meckel W, J cie». Sac C, 1968, 1615.119 Kosar W, in Handbook of Grignard Reagents, edited by G S

Silverman & P E Rakita, (M Dekker, New York) 1996, pp441-453.

120 Shen Y & Yao J, J Chern Res SYIIOP, 1996,394.121 Anastas P T & Williamson T C, Green Chem , 1998, I.122 Anastas P T & Kirchhoff M M, Acc Chern Res, 35, 2002.

686.123 Ahluwalia V K & Kidwai M, New Trends ill Green

Chemistry; 2nd Edn., (Anamaya, New Delhi) 2006.124 Akagah A, Poite J C & Chanon M, Org Prep Proced 1111,

17,1985,219.125 Hill C L, Vander Sande J B and Whitesides G M, J Org

Chern, 45, 1980, 1020.

"-