The Grignard Reagent

Preparation, Structure, and Some Reactions

Milton Orchin
University of Cincinnati, Cincinnati, OH 45221

The Grignard reagent, t h e product resulting from the re-

action of an alkyl or aryl halide with elemental magnesium
(I),
RX + Mg - RMgX CH3MgBr Etp -320

was named after Victor Grignard (1871-1935; Nobel Prize, CHsMgBr THF 3.5
shared with Paul Sabatier, 1912) and is probably the single
most important organometal reagent used in the laboratory CzHsMgBr Et20 482
synthesis of organic compounds in this century. Indeed a
case can he made for the proposition that the remarkable C2HsMgBr THF 5.09
versatility of the Grignard reagent and its ease of prepara-
C2HsMgCI THF 5.52
tion delayed the interest in organo-transition metal chemis-
try for decades. l'heavai~abili;~ of pure magnesium required
for the laboratory preparation of the Grignard reagent was
no doubt enhanced hy the commercial interest in magne-
sium, principally for use as an incendiary. The "Elektron"
was a 10-lb incendiary bomb developed by the Germans in comes from 'H NMR variable temperature studies @a). At
1918 and consisted of almost pure magnesium. room temperature only a single methyl signal is observed,
The successful preparation of the Grignard reagent in hut a t about -100 'C two methyl signals appear, indicating
good yield requires an ether solvent, usually diethyl ether or that the rapid equilibrium a t room temperature is frozen out
tetrahvdrofuran (or less commonly EtzN), whose solubiliz- a t the low temperature.
ing prbperties arise from complcx~tionot the ether oxygen There are several dimeric structures that can be written
at the maanesium center. Although practically all Grignard for the Grignard reagent inEtzO when X = F, CI, Br, or I and
reactions can be rationalized on t h e basis o f the strGture in T H F where X = F. Of these, the symmetrical, 3, and
RMgX, it was realized for many years that such a represen- unsymmetrical, 4, dimers are the most probable. However,
tation is a gross oversimplification and that the formation of the dimeric structure that best explains (66) the intermedi-
the reagent and its structure and reactions are very complex ate responsible for the Schlenk equilihrium is the alkyl
(2). bridging dimer, 5.

Structure of the Grlgnard Reagent

I t is now generally accepted that in diethyl ether and
tetrahydrofuran solution, Grignard reagents are best repre-
sented by an equilihrium mixture:
2RMgX + R,MG + MgXz The dramatic effect of solvent on the composition of the
1 2 Grignard reagent is illustrated by experimental values of K
(see table) found (7) for the Schlenk equilibrium in the
RzMg can be conveniently separated from the RMgX and solvents diethyl ether and tetrahydrofuran. The large differ-
MgXz (which are highly associated in EtzO) by addition of ence in K as a function of solvent is ascribed to solvation
dioxane, which precipitates all the halide. No substrate is effects, especially those associated with solvation of the
known that will react exclusively with RMgX hut not RzMg MgBrz; in THF, four solvent molecules are involved, but in
and vice versa, and the reaction with either alone or the EtzO, only two are involved.
mixture always produces the same product albeit with occa-
sional differences in rate (3). Dimeric and polymeric struc- Formation of the Grignard Reagent
tures make minor contributions, especially in dilute EtzO There appears to be general agreement that Grignard for-
solution (less than 0.3 M) and hence the equilibrium he- mation is initiated by a single electron transfer (SET) from
tween 1 and 2, the so-called Schlenk (4) equilibrium repre- surface magnesium to an adiacent RX, formina a radical
sents the hest formulation of the Grignard reagent. This anion that is associated witha surface kagnesi"m radiral-
equilihrium is consistent with the exchange observed he- cation. There also is general agreement that free radicals R.
tween radioactive 28MgBr~ and (CzH5)~Mg (5a) and between are formed somewh&e in the process and that RMgX is
25MgBr~ and (CzH5)zMg (56) when carefully purified magne- formed before RzMg with which i t rapidly equilibrates.
sium is used. No exchange is reported t o occur (5c) when less There remains, however, some controversy as to whether the
pure magnesium is used. The reason for the difference re- free radicals formed first diffuse directly and completely
mains a mystery. Impurities in the range of parts per million into solution and return t o the surface to form RMgX or
and/or traces of oxygen have been implicated but not dem- whether some fraction of the population of radicals remains
onstrated. I t is more likelv that ex~erimentaltechniques associated with the maenesium and combine to form RMeX
were imperfect in the no-exchange experiments. I.'nequivo. without any substantii liferime as free radicals in solutih.
ral evidence for the Schlenk equilihrium, The two pathways to HMgX areshown in Figure 1 where the

586 Journal of Chemical Education

subscripts indicate a surface species (s) and a species in this pathway as well as pathway d operate competitively.
solution (soh), respectively. Relatively little work has been done on the rate of forma-
Perhaps the most convincing evidence for the suggestion tion of the Grignard reagent and the identification of the
that radical formation can occur on the surface to form kinetically important rate-determining step although the
RMgX (pathway b, d) without formation of free R. (pathway initial one-electron reaction on the magnesium surface is a
c, e, d) is the behavior of chiral l-hromo-l-methyl-2,2-di- reasonable candidate. The rate increases with surface area of
phenylcyclopropane, 6. themagnesium (16) but does not seem to beaffected by trace
imnurities. In one studv the rate of reaction in ether of
cyhopentyl bromide with magnesium metal in the form of a
variable meed rotatine disk (17) was determined. The rate
was foundto be propo;tional the square root of the angu-
lar velocity of the disk demonstrating that the reaction is
mass transfer limited.
ReaCtlon of Grlgnard Reagents
The reaction of the bromide 6 with Me eives the corres~ond- There are literally thousands of examples of the reactions
ing RMgBr with an optical purity of ;rout 12%(8). were all of Grignard reagents with a variety of substrates. Perhaps
the bromide to proceed to reaction exclusivelv throueh the one of the most detailed studies involves the reaction with
free radical 7 inbo~ution,complete racemization wouli have benzophenone, e.g.:
occurred since in inde~endentstudies it has been shown (9)
that the rate of inversibn of the methylcyclopropyl radical (7
without the phenvl - . is about 101's-I at 71 O C . Fur-
groups)
thermore when 7 i s generated in solution from a variety of In a careful study of this reaction using both CH3MgBr and
chiral precursors, other than the Grignard reagent, the prod- (CH3)3CMgBr,a dual pathway mechanism, Figure 2, was
uct is racemic (lo), and side products resulting from dimer- proposed (18) that rationalized the rates and extent of con-
ization and disproportionation that are characteristic of free version of the ketone to 11, the precursor to the tertiary
radical reactions in solution are present in such reactions alcohol, and 12, the precursor to the pinacol, the two major
(11) but are absent in the preparation of RMgBr from 6. products. As shown in Figure 2, the former develops via a
The ~ . r e.~ a r a t i oofnthe Grienard reaeent from chiral nre- polar pathway, and the latter via a single electron radical
cursors isonly one of a number of probes for assessing radical pathway (Fig. 2).
formation. Other ~robesconsistof CIDNP studies (12).radi- The partitioning between the polar and the one electron
cal trapping expe;iments (13), and, particularly, the isomer- transfer pathways depends on the nature of the ketone, the
ization of rearraneement-prone radicals. This latter Drobe solvent, the R group, and the purity of the magnesium metal.
has been particul&ly instructive because the rate ofrear- Thus the reaction of CH3MgBr in EtnO (prepared from pure
rangement of several radical species are known from inde-

-
pendent experiments and the extent of isomerization RMgX
R'MgX during Grignard formation can be determined
experimentally as well.
single crystal Mg in excess CH3Br) with PhzCO proceeds
exclusivelyvia the polar mechanism, while tert-butylmagne-
sium chloride reacts with P h E O in EtzO by single electron
transfer. Interestingly, when 0.05 M FeC13 is added to the
Extensive experiments have been performed with 6-bro- CH3MgBr reaction the mechanism is converted from the
mo-l-hexene, 8. When the Grignard reagent is prepared polar pathway to the single electron transfer mechanism (see
from this halide, two RMgX compounds are isolated: below) with production of the pinacol.

-
The rate of isomerization R- .R' of the putative intermedi-
ate cyclohexenyl radical is known (14) to be about lo5 s-',
and preparation of the Grignard reagent from 8 leads to
about 95% of 9 and 5%of 10. When a log-log plot of the ratio
of R'MgXIRMgX to the rate constant for the rearrangement
is constructed using this point and two others similarly ob-
tained, a straight line with a slope of 0.5 can be used to
connect these points (16). This relationship is to be expected
from a diffusion theory model (15).Hence it was proposed
that in Grignard formation the radical R. is formed first,
diffuses into solution, has time to moreor less rearranee. and
both b a n d .R' return to thesurfare toreact with ~ g ~ , , a n d
form HMgX and R'MgX. This is the process of pathway c, e,
d, Figure I . The presently available evidence indicates that

Figure 1 . The mechanism of Orignard formation Figure 2. me two p h w a y s fw ma Grignard-ketone reaction.

Volume 66 Number 7 July 1989 587

The Effect of Transition Metals and Thelr Halldes on the
Grignard Reagent and Its Reactions
When catalytic quantities of CoC12are added to a solution
of a Grignard reagent containing free excess alkyl halide,
rapid coupling occurs:
P.M... r\ c0.1.. 4 c,.,

RMgX R X+ ac4
-
RR' MgX, + Figure 3. Free radical catalysis of the Grlgnard reaction.

as well as formation of RR, Rf-H, and R'R'. The steps in the

reaction may be written as follows using the reaction be-
tween ethyl hromide and phenylmagnesium bromide: RMgCl MCI, + - RMCI + MgC1,
CoCI, + PhMgBr - Ph. + CoCI + MgBrCl and with excess RMgX
2RMgCI + MClz - R,M + 2MgC1,
2Ph - Ph-Ph The most electro~ositivemetal (in this case Me) ends UD
comhined with halogen. ~ l e m e n t a magnesium
l Lduces th'e
halides of all metals below it in the electromotive series. and
similarly RMgX converts the inorganic halides of all metals

C,Hs + Solvent - C,H,

below magnesium to the alkyl derivative of the element.

-
Thus with CuXz the reaction may he written: CuClz
+
+
Ph. + Solvent - C,H,
2RMgX RzCu 2MgXz followed by reductive elimina-
tion: RzCu R-R +
Cu. This represents an overall two elec-
tron reduction of the metal, but this explanation of the
The generation of the free radical occurs in the first step. coupling has received relatively littlesupport,and it is much
The mechanism of this initiation can be viewed as exactly more likely the product arises through the radical mecha-
analogous to the one electron transfer pathway in the benzo- nism shown in the catalytic cycle with CoCl?.
phenone reaction with methylmagnesium bromide shown
earlier: Acknowledgment
Discussions with E. C. Ashby, H. M. Walhorsky, and J. L.
Belletire were very helpful.

A catalytic cycle involving CoC1 t o generate P h and CzHs

radicals in the above reaction is shown in Figure 3. One
would expect the radical pathway to PhC2H5to be competi-
tive with its formation from the S N displacement
~ in a
Wurtz-type coupling reaction:
BrMgPh + C&Br - +6 -6 +6
[BrMg. .Ph.. .CzHs.. Br]
-6

In the above discussion i t was assumed that some RX is

present in excess in the Grignard solution containing the
catalytic CoClz (or other transition metal halide such as
CuClz, FeCla etc). However, it is known that if CoClz is
added to a Grignard reagent in the absence of free RX,
coupling of the Grignard still proceeds. Presumably the
mechanism shown for the one electron transfer involving
RMgX and CoClz that leads to radical formation is again
operative. However, another possibility should be consid-
ered since i t is known (19) that RMgX can convert a transi-
tion metal halide to the alkyl derivative:

588 Journal of Chemical Education