Absence of Detectable Freely Diffusing Radicals during the Formation of an Aromatic Grignard Reagent

Full text of DOI:10.1021/jo0001028

5
014
J . Org. Chem. 2000, 65, 5014-5015
Absen ce of Detecta ble F r eely Diffu sin g
Ra d ica ls d u r in g th e F or m a tion of a n
Ar om a tic Gr ign a r d Rea gen t
on the metal surface, and must react there. These
systems usually give only partial retention of configura-
6
tion, together with significant racemization. Walling has
compared the abilities of the Garst and the Walborsky
models to account for known product mixtures from the
reactions of alkyl halides and magnesium. Ashby and
Oswald7 have published Grignard runs under varied
conditions; they interpret their data as support for return
of radicals from solution to the Mg surface to form RMgX.
These studies of RMgX formation based upon observation
of changes in stereochemistry or structure of course are
limited to aliphatic halides. In fact, we are not aware of
any studies of the formation of aromatic Grignard
reagents until the 1999 publication by Garst and co-
workers. Another such investigation is reported here,
based upon the known very fast rearrangement of the
o-allyloxyphenyl radical (2 in Scheme 1) reported by
Robert I. Walter*
Chemistry Department, University of Illinois at Chicago,
Chicago, Illinois 60607
mhry@aol.com
Received J anuary 25, 2000
The Grignard reagent has probably been the most
widely used intermediate in organic chemistry since its
introduction by Victor Grignard in 1900. Despite this
wide use, it is not possible to assign a specific structure
for a particular reagent. This is because both RMgX and
1
8
R
2
Mg are formed during preparation of the reagent and
Abeywickrema and Beckwith and prepared by them by
are connected by the equilibrium described by the
Schlenks. The position of this equilibrium depends on
reduction of o-allyloxychlorobenzene with tributyltin
hydride. The argument (shown in the scheme) is that the
free radical formed by transfer of an electron to halide
1, followed by loss of the halide ion and subsequent
2
the solvent, the structure of the alkyl group, and the
nature of the halogen. The mechanisms of reactions that
involve RMgX also have been extensively studied, and it
has been recognized that the formation of the reagent
mixture is a separate problem from its reactions. We are
concerned here with the former.
2
rearrangements, can be trapped by carbonation with CO ,
the carboxylic acid so formed isolated and purified, and
its structure determined. If at any point in the prepara-
tion the free radical 2 is formed and can move away from
the metal surface, that radical will undergo very fast
rearrangement, and the trapped carboxylic acid will have
the structure of 2,3-dihydro-3-benzofuranacetic acid, 5.
If intermediate 2 never leaves the Mg surface, the
carbonation product will be unrearranged 2-allyloxyben-
zoic acid, 4. The product isolated in a quite pure state is
Since the conventional reaction system in which RMgX
is formed is heterogeneous, studies by the usual methods
of chemical kinetics have been rare prior to the recent
extensive investigation by Whitesides and his collabora-
3
tors. Their results are particularly significant in that
they demonstrate directly the presence of a previously
hypothetical intermediate free radical during the forma-
tion of Grignard reagent from cycloheptyl bromide and
magnesium metal. There is agreement among recently
active research groups that formation of RMgX is initi-
ated by donation of an electron from magnesium metal
to an alkyl halide molecule at the metal surface. The
surface radical anion so formed will lose a halide ion to
give the initial free radical intermediate in the formation
1
3
the latter, identified by its proton-decoupled C NMR
spectrum.
Exp er im en ta l Section
o-Allyloxyiodobenzene 1 was prepared using the modified
9
Williamson synthesis proposed by Goering and J acobson. We
found a significant increase in yield (to 90%) when the mole ratio
of potassium carbonate to o-iodophenol used in the synthesis
was increased above unity.
4
of RMgX. Garst and co-workers have investigated sub-
sequent steps in Grignard reagent formation by examin-
ing the equilibrium free radical distribution from this
initial radical assumed to be released from the magne-
sium surface under ether. Subsequent free diffusion of
this radical product through the ether solvent while
undergoing permissible radical rearrangements gives an
equilibrium solution from which solutes return to the Mg
to form RMgX and RMgX (rearranged). This equilibrium
system can be modeled quantitatively; it would be
quenched to product mixtures that correspond well to
those observed experimentally. Studies by Walborsky and
P r ep a r a tion a n d Ca r bon a tion of th e Gr ign a r d Rea gen t
fr om 1. This Grignard was prepared in the conventional manner
in ether solution from a small excess of magnesium metal and
halide 1. Reaction generally started easily; a few crystals of
iodine helped in stubborn cases. After reaction at the metal
surface stopped, the solution was stirred for 60 m at ca. 35°;
usually a second straw-colored phase separated. This two-phase
mixture was cooled below -5°, and dry CO
rapidly until heat evolution stopped. The ether phase was diluted
with CH Cl , washed with 1 M HCl, and then extracted twice
2
was passed in
2
2
with molar bicarbonate solution. These alkaline extracts were
combined and acidified, and the precipitated solids were filtered
and dried. This product gives an NMR spectrum with many
relatively weak impurity lines. It was recrystallized by extraction
into pentane in a Soxhlet apparatus. Yields of extracted recov-
5
collaborators on retention of configuration during reac-
tions of chiral cyclopropyl-based halides led them to
conclude that RMgX forms on the magnesium surface
without migration of any intermediate through the
solution. Any free radical intermediate formed is always
1
3
ered solids were 49-63%, mp 62.7-62.9 °C. C NMR in CDCl
3
(
assignments aided by off-resonance decoupling): 165.9 s, Cd
O; 157.3 s, C
2
; 134.7 d, -C-H; 133.3 d, C-H; 131.1 d, C-H;
1
21.9 d, C-H; 119.8 t, dCH
2
; 118.0 s, C ; 113.1 d, C-H; 70.5 t,
1
(
(
(
1) Grignard, V. Compt. Rend. 1900, 130, 1322.
2) Schlenk, W.; Schlenk, W., J r. Berichte 1929, 62, 920.
3) Root, K. S.; Hill, C. L.; Lawrence, L. M.; Whitesides, G. M. J .
(6) Walling, C. Acc. Chem. Res. 1991, 24, 255.
(7) Ashby, E. C.; Oswald, J . J . Org. Chem. 1988, 53, 6068.
(8) Abeywickrema, A. N.; Beckwith, A. L. J . J . Chem. Soc., Chem.
Am. Chem. Soc. 1989, 111, 5405.
(
4) Garst, J . F.; Swift, B. L. J . Am. Chem. Soc. 1989, 111, 241-250.
Comm. 1986, 464-465.
(9) Goering, H. L.; J acobson, R. R. J . Am. Chem. Soc. 1958, 80, 3278.
(5) Walborsky, H. M. Acc. Chem. Res. 1990, 23, 286-293.
1
0.1021/jo0001028 CCC: $19.00 © 2000 American Chemical Society
Published on Web 07/14/2000

Notes
J . Org. Chem., Vol. 65, No. 16, 2000 5015
Sch em e 1
ment by carrying the halogen atom and an allyloxy group
on atoms ortho to one another on an aromatic ring. This
limits the range of structures which could be tested, but
we expect that other aromatic halides that have the
necessary structural features would show corresponding
behavior. Both phenyl rings and alkenes are known to
adsorb on some metal surfaces under high vacuum
conditions,12 so it is possible that both functional groups
in 2 adhere to the Mg surface and thus disable this
cyclization probe for a radical intermediate during forma-
tion of RMgX. The Grignard reagent from 1, once formed,
appears to have normal stability; note that our Grignard
reagent solution was stirred for an hour before carbon-
ation. This result cannot apply to preparation of aliphatic
Grignard reagents, where rearrangements of the alkyl
group are known to be common. Garst and co-workers
have recently reported13 a detailed study of ArMgX
formation from aryl halides, including comparisons of
product structures in these reactions with those of alkyl
halides. They showed that products formed from aliphatic
halides by side reactions and successfully predicted by
the model proposed earlier by Garst (with different co-
workers) generally are absent when an aromatic halide
is used. These authors conclude that aromatic free
radicals are not intermediates along the major pathway
to ArMgX. They probed this point further by observing
formation of Grignard reagents from o-(3-butenylphenyl)
halides, which could ring close by a radical rearrange-
ment analogous to 2 f 3. This rearrangement has been
OCH . (The five unassigned C-H shifts are due to carbon atoms
which cannot be unambiguously assigned.) These values are
identical to the published chemical shifts for the unrearranged
2
1
0
acid 4 and significantly different from those for the expected
1
1
rearrangement product 5.
Note that the isomerized acid 5 would be expected to carry
through the purification sequence with 4; its absence indicates
that little, if any, isomerization occurs. Since radical 2 is known
to isomerize very fast in solution, we infer that no freely diffusing
radical is present at any stage of the formation of this Grignard
reagent from an aromatic halide.
8
reported to have a first-order rate constant equal to 7.4
7
-1
×
10 sec , or about 1/16 the value for 2 f 3. They have
observed variable (usually small) amounts of 1-meth-
ylindane among the products of these reactions, suggest-
ing variable minor participation by radical intermediates.
Their most recent ideas about the mechanisms of Grig-
nard reagent formation are summarized in a recent
review.14
Discu ssion
The absence of rearrangement during preparation of
the Grignard reagent from halide 1 implies that at no
time during the reagent’s synthesis is the free radical 2
liberated to move independently in the ether solvent,
where it should undergo very fast rearrangement to 3.
We have this result for only one aromatic halide, and this
must meet the structural requirements for rearrange-
Ack n ow led gm en t. The author is much indebted to
Professor A. L. J . Beckwith for a discussion some years
ago of the use of radical 2 as a mechanistic probe. The
research of Ted Field for his senior thesis at UIC (on a
different problem) led the author to formulate this
investigation. Incorporation of the recent material in
refs 13 and 14 was kindly suggested by a referee. This
work was supported by the Chemistry Department of
the University of Illinois at Chicago.
(
10) Stanetty, P.; Kesler, H.; Purstinger, G.; Monatsh. Chem. 1990,
1
21, 888.
(
11) Bravo, P.; Gentile, A.; Ticozzi, C. Gazz. Chim. Ital. 1984, 114,
8
9-91.
(
(
12) Bent, B. E. Chem. Rev. 1996 96, 1361-1390.
13) Garst, J . F.; Boone, J . R.; Webb, L.; Lawrence, K. E.; Baxter, J .
T.; Ungvary, F. Inorg. Chim. Acta 1999, 296, 52-66.
14) Garst, J . F.; Ungvary F. In Grignard Reagents: New Develop-
ments; Richey, H. G., Ed. Wiley-Interscience: New York, 2000; pp 185-
75.
(
2
J O0001028