Mechanism of mercuration of ferrocene: General treatment of electrophilic substitution of ferrocene derivatives

Article abstract of DOI:10.1021/om960815+

The mercuration of (pentadeuteriocyclopentadienyl)cyclopentadienyliron(II), 1, occurs preferentially at the C₅H₅ ring. The ratio of C₅H₅:C₅D₅ attack ranges from 5.2-6.5, depending upon the mercurating agent employed. In contrast to the Friedel-Crafts acetylation of 1, mercuration does not give rise to intramolecular interannular proton transfers. Under certain conditions, a complex, 6, can be isolated that contains both carbon-mercury and iron-mercury bonds. Taking these facts into account, a mechanism for the mercuration of ferrocene is proposed whereby precomplexation of the mercurating agent to the iron atom precedes the rate determining formation of the carbon-mercury bond with concomitant loss of H⁺. Combined with our previous work concerning the acetylation and proton exchange of ferrocene, this study allows the formulation of a general mechanism of electrophilic substitution of ferrocene, based on the reactivity of the electrophile and the basicity of the iron atom of the resulting ferrocene product.

Full text of DOI:10.1021/om960815+

1114
Organometallics 1997, 16, 1114-1122
Articles
Mech a n ism of Mer cu r a tion of F er r ocen e: Gen er a l
Tr ea tm en t of Electr op h ilic Su bstitu tion of F er r ocen e
Der iva tives
Allan F. Cunningham, J r.
Additives Division, Ciba Specialty Chemicals Research Marly SA, P.O. Box 64,
1723 Marly, Switzerland
Received September 24, 1996X
The mercuration of (pentadeuteriocyclopentadienyl)cyclopentadienyliron(II), 1, occurs
preferentially at the C5H5 ring. The ratio of C5H5:C5D5 attack ranges from 5.2-6.5, depending
upon the mercurating agent employed. In contrast to the Friedel-Crafts acetylation of 1,
mercuration does not give rise to intramolecular interannular proton transfers. Under
certain conditions, a complex, 6, can be isolated that contains both carbon-mercury and
iron-mercury bonds. Taking these facts into account, a mechanism for the mercuration of
ferrocene is proposed whereby precomplexation of the mercurating agent to the iron atom
precedes the rate determining formation of the carbon-mercury bond with concomitant loss
of H+. Combined with our previous work concerning the acetylation and proton exchange
of ferrocene, this study allows the formulation of a general mechanism of electrophilic
substitution of ferrocene, based on the reactivity of the electrophile and the basicity of the
iron atom of the resulting ferrocene product.
Sch em e 1
In tr od u ction
Interest in the special reactivity and properties of
ferrocene and its derivatives has not subsided, even
more than four decades after the discovery of this
metallocene.1 One of the special features of this com-
pound is its aromaticity,2 which is instrumental in the
synthesis of a wide variety of substituted ferrocenes via
electrophilic substitution. Not surprisingly, the mech-
anisms of these useful transformations have also at-
tracted considerable attention.3 In principle, electro-
philic attack can occur at two different sites: at a carbon
atom of a cyclopentadienyl ring either on the face
opposite to the metal-carbon bonds (exo attack) or on
the same face as the metal-carbon bonds (endo attack,
Scheme 1). In addition, intermediates such as A, having
an iron-electrophile bond, may be involved in either
pathway.
and 1,1′-bis(tributylstannyl)ferrocene led to mixtures of
three isomeric acetylferrocenes.4a Since such product
mixtures could not result from a simple endo displace-
ment of a trimethylsilyl or tributylstannyl group by the
acylating agent, the nature of the intermediate was
probed via deuterium labeling. This revealed that
attack of the acyl chloride-AlCl3 complex occurred at
the exo face of the cyclopentadienyl ring and, subse-
quent to proton transfer from carbon to iron, afforded
two isomeric cationic iron hydrides (Scheme 2).
A
second heteroannular or homoannular proton transfer
resulted in the elimination of trimethylsilyl or tributyl-
Previous work in our laboratory has focused upon the
mechanistic course of the Friedel-Crafts acylation of
ferrocene.4 Acetylation of both 1,1′-bis(trimethylsilyl)-
(3) (a) Rosenblum, M.; Santer, J . O.; Howells, W. G.; J . Am. Chem.
Soc. 1963, 85, 1450. (b) Rosenblum, M.; Abbate, F. W. J . Am. Chem.
Soc. 1966, 88, 4178. (c) Mangravite, J . A.; Traylor, T. G. Tetrahedron
Lett. 1967, 4461. (d) Bitterwolf, T. E.; Ling, A. C. J . Organomet. Chem.
1972, 40, 197. (e) Cerichelli, G.; Floris, B.; Illuminati, G.; Ortaggi, G.
J . Org. Chem. 1974, 39, 3948. (f) Floris, B.; Illuminati, G. Coord. Chem.
Rev. 1975, 16, 107. (g) Cerichelli, G.; Illuminati, G.; Ortaggi, G.;
Giuliani, A. M. J . Organomet. Chem. 1977, 127, 357. (h) Fung, C. W.;
Roberts, R. M. G. Tetrahedron 1980, 36, 3289. (i) Roberts, R. M. G.;
X Abstract published in Advance ACS Abstracts, February 1, 1997.
(1) (a) Kealy, T. J .; Pauson, P. L. Nature 1951, 168, 1039. (b) Miller,
S. A.; Tebboth, J . A.; Tremaine, J . F. J . Chem. Soc. 1952, 632.
(2) Woodward, R. B.; Rosenblum, M.; Whiting, M. C. J . Am. Chem.
Soc. 1952, 74, 3458.
S0276-7333(96)00815-1 CCC: $14.00 © 1997 American Chemical Society

Mechanism of Mercuration of Ferrocene
Organometallics, Vol. 16, No. 6, 1997 1115
Sch em e 2
stannyl chloride and formation of the observed products.
The fact that the deuterium content of the products was
identical to that of the reactant indicated that direct
deprotonation of the iron hydride does not occur.
In order to exclude that the exo acylation and the
ensuing ring to metal to ring proton transfers were
specific to the aforementioned ferrocene derivatives, we
prepared and studied the acetylation of the pentadeu-
terated ferrocene 1.4b Here again, the exo attack of the
acylating agent could be easily demonstrated. Whereas
1H NMR analysis showed that acylation occurred in-
discriminately at both cyclopentadienyl rings, the deu-
terium content of the acetylferrocene obtained was 4.82
atoms (eq 1). Clearly, a metal deuteride is engendered
requires that protonation also occurs via exo attack of
H+. Further evidence for the exo protonation of fer-
rocenes has been obtained through the study of the
formation of dihydrogen from [1.1]ferrocenophane in
acidic media.3j Treatment of this dimetallocene with
BF3-D2O results in its oxidation and the generation of
predominately H2. If direct protonation of the metal
center was operative, the exclusive generation of D2
would be expected.
Armed with this diagnostic tool for identifying on exo
attack, we then considered the electrophilic mercuration
of ferrocene. The interaction of various mercury(II)
compounds with ferrocene has been the subject of a
number of publications. Although the precise structure
of ferrocene-mercuric salt adducts is yet to be resolved,
the presence of an iron-mercury bond has been detected
by infrared,5 Mo¨ssbauer,3i,6,7 solid state 13C NMR,7 and
ultraviolet spectroscopy.8 Indeed, the kinetic evaluation
of the acetoxymercuration of ferrocene in acetic acid
seemed to indicate the rapid formation of a related 1:1
adduct of ferrocene:mercuric acetate (A, E ) Hg-
(OCOCH3)2) followed by the rate determining formation
of the carbon-mercury bond (endo attack).3h However,
apart from the reasonable correlation of the observed
reaction rate with the kinetic expression for the pro-
posed transformation, only a shift in the 1H NMR signal
of ferrocene upon addition of mercuric acetate and a
transient color change to red upon mixing concentrated
solutions of the two reactants could be offered as proof
for precomplexation of mercuric acetate to the iron atom
of ferrocene. In contrast, the lack of a direct correlation
between the rate of mercuration and the number of
ferrocene alkyl substituents has been used as evidence
against prior precomplexation.3f Whereas the addition
which undergoes preferential loss of HCl with the
concomitant formation of a C-D bond. In fact, evidence
for both proton and deuteron transfer products was
obtained through mass spectral analysis: the m/e 122
(C5H4D1Fe) and m/e 125 (C5H1D4Fe) fragments of the
product were enriched relative to acetylferrocene-d0 and
acetylferrocene-d9. Therefore, the occurrence of in-
tramolecular proton transfers is not an isolated phe-
nomena and is indicative of an exo attack of an
electrophile.
(5) Morrison, W. H.; Hendrickson, D. N. Inorg. Chem. 1972, 11, 2912.
(6) (a) Roberts, R. M. G.; Silver, J .; Azizian, J . J . Organomet. Chem.
1986, 303, 387. (b) Watanabe, M.; Motoyama, I.; Sano, H. Bull. Chem.
Soc. J pn. 1986, 59, 2103. (c) Silver, J . J . Chem. Soc., Dalton Trans.
1990, 3513.
With respect to the process of hydrogen exchange of
ferrocene, the principle of microscopic reversibility
Silver, J .; Morrison, I. E. G. J . Organomet. Chem. 1981, 209, 385. (j)
Mueller-Westerhoff, U. T. Angew. Chem. 1986, 98, 700. (k) Mueller-
Westerhoff, U. T.; Haas, T. J .; Swiegers, G. F.; Leipert, T. K. J .
Organomet. Chem. 1994, 472, 229.
(4) (a) Cunningham, A. F., J r. J . Am. Chem. Soc. 1991, 113, 4864.
(b) Cunningham, A. F., J r. Organometallics 1994, 13, 2480.
(7) (a) Watanabe, M.; Masuda, Y.; Motoyama, I.; Sano, H. Chem.
Lett. 1987, 1981. (b) Watanabe, M.; Masuda, Y.; Motoyama, I.; Sano,
H. Bull. Chem. Soc. J pn. 1988, 61, 3479. (c) Watanabe, M.; Masuda,
Y.; Motoyama, I.; Sano, H. Hyperfine Interact. 1988, 40, 355.
(8) Traverso, O.; Chiorboli, C.; Mazzi, U.; Zucchini, G. L. Gazz. Chim.
Ital. 1977, 107, 181.

1116 Organometallics, Vol. 16, No. 6, 1997
Cunningham
Sch em e 3
of one ethyl group to ferrocene engenders an ap-
proximately 2-fold increase in the reaction rate, the
introduction of the second ethyl group promotes a
further increase of only 25%. Mercuration via an endo
attack without precomplexation had been previously
advanced in an earlier study in which the rates of
various electrophilic substitution reactions were shown
to correlate with σ+ constants determined for the endo
and exo attack of ferrocene derivatives. According to
this mechanistic interpretation, weak electrophiles such
as Hg(II) should preferentially attack the more electron-
rich endo face of the cyclopentadienyl ring.3c
We now report our work concerning the mercuration
of (pentadeuterocyclopentadienyl)cyclopentadienyliron
(1) which unequivocally confirms that mercuration
proceeds via precomplexation and subsequent endo
transfer of the electrophile. In addition, these results
allow the formulation of a general mechanism for the
electrophilic substitution of ferrocene and its deriva-
tives, based not only on the strength of the electrophile
but also on the basicity of the iron atom.
F igu r e 1. 1H NMR spectrum of chloromercurioferrocene-
d0 (left) and 2a (right) between 4.0 and 4.6 ppm, with
constant amplitude of the unsubstituted cyclopentadienyl
singlet.
the mercuration of ferrocene with Hg(OAc)2,3c although
the exact reaction conditions were not specified.
Resu lts a n d Discu ssion
Although a large isotope effect is observed upon
mercuration, the question of whether mercuration pro-
ceeds via an endo or exo attack remains open. Due to
the absence of the C5H5Fe fragment in the mass
spectrum of chloromercurioferrocene, 2a had to be
converted into ferrocene to allow for the analysis for
proton transfers. Toward this goal, 2a was reduced with
LiAlH4 in THF followed by hydrolysis with H2O to
provide 3a and reduced similarly with LiAlD4/THF-d8/
D2O to furnish 4a . Comparison of the mass spectrum
of 3a with that of 1 allows for the identification of the
C5H4D1Fe fragment (m/e 122) that would arise from an
exo attack of the mercurating agent of the C5D5 ring
followed by interannular deuteron transfer. This would
lead to an enrichment of this fragment relative to the
C5H5Fe fragment (m/e 121), which is obtained from the
other three potential mercuration products upon reduc-
tion with LiAlH4 (Scheme 4, only reduction with LiAlH4
is shown). As can be seen, the percentage of this
fragment in the spectra of 3a and 1 is practically
identical; therefore, no deuteron transfer has occurred
Treatment of a 0.33 M solution of 1 in 1:1 toluene:
ethanol with 0.5 equiv of mercuric acetate followed by
ligand exchange with LiCl led to the isolation of chlo-
romercurioferrocene 2a in 35% yield (based on 1,
Scheme 3).9 An additional 58% of unreacted 1 could be
recovered; the balance of the material was presumably
converted to the insoluble 1,1′-bis(chloromercurio)fer-
1
rocene. Inspection of the H NMR clearly shows that
mercuration occurs preferentially on the C5H5 ring of 1
(Figure 1). In the absence of interannular proton
transfers, indiscriminate mercuration of 1 would afford
a spectrum identical with that of undeuterated chloro-
mercurioferrocene: the ratio of each broad triplet (repre-
senting the two pairs of protons of C5H4HgCl) to the
singlet (C5H5) would be 1:2.5. In the case at hand, the
ratio of C5H5 attack to C5D5 attack is 6.5:1, again if
potential interannular proton transfers are disregarded.
An isotope effect of 3.2 has previously been reported for
(9) Fish, R. W.; Rosenblum, M. J . Org. Chem. 1965, 30, 1253. These
authors employ methanol/benzene.

Mechanism of Mercuration of Ferrocene
Organometallics, Vol. 16, No. 6, 1997 1117
Sch em e 4
Sch em e 5
(Figure 2). A similar analysis reveals, as would be
expected, that no proton transfer occurs during the
mercuration of 1. Given the lack of intramolecular
proton (deuteron) transfers, the mass spectra of 3a and
4a also permit a second, albeit less accurate, evaluation
of the ratio of C5H5 attack to C5D5 attack. Inspection
of the molecular ions of 3a and 4a indicate that the
former is composed of 12% ferrocene-d4 and 88% fer-
rocene-d5 whereas the latter consists of 83% ferrocene-
d6 and 17% ferrocene-d5. This translates into a C5H5:
C5D5 attack ratio of ∼5.9: 1, which is in unison with that
F igu r e 2. Comparison of the mass spectra of 3a (lined
bar), 1 (solid bar), and 4a (dotted bar) in the m/e 121-126
region (normalized to 100%).
water.11 The activation energy determined for this
demercuration was 19.1 kcal/mol. Indeed, demercura-
tion should not require the formation of a metal proto-
nated ferrocene intermediate since, according to the
mechanism proposed, it must proceed via an exo attack
by H+ of the carbon atom bound to mercury (the
microscopic reverse). The desilylation and deborylation
of ferrocene derivatives under mildly acidic conditions
must also proceed in a similar manner.3e
At this juncture, the course of the reductive demer-
curation of 2 merits comment. There is some confusion
in the literature regarding the mechanism of the reac-
tion of aryl- and alkylmercuric halides with LiAlH4.
Publications in the 1950’s concerning phenylmercuric
chloride and diphenylmercury claimed that the ethereal
LiAlH4 reduction of both of these compounds afforded
tetraphenylaluminate.12 Subsequent hydrolysis of the
1
obtained through analysis of the H NMR spectrum of
2a , 6.5:1. It is important to note that recovered 1
exhibited a mass spectrum identical to that of the
original sample.
These results imply that the acetoxymercuration of
ferrocene proceeds via a rate determining endo attack
of the mercurating species with a concerted or stepwise
exo loss of H+, although it is yet unclear whether a
species with an iron-mercury bond serves as an inter-
mediate (Scheme 5). The preferred endo attack of the
mercurating agent may be the result of the instability
of the intermediate 5, which would arise from an exo
attack and a ring to iron transfer of a proton, toward
demercuration. Whereas ferrocene can only be proto-
nated with strong acids such as BF3-H2O,10 the hy-
drolysis of chloromercurioferrocene can be accomplished
under mildly acidic conditions, 0.01 M HCl in dioxane-
(11) Nesmeyanov, E. G.; Perevalova, E. G.; Gubin, S. P.; Kozlovskii,
A. G. J . Organomet. Chem. 1968, 11, 577.
(12) (a) Barton, D. H. R.; Rosenfelder, W. J . J . Chem. Soc. 1951,
2385. (b) Winstein, S.; Traylor, T. G. J . Am. Chem. Soc. 1956, 78, 2597.
(c) Traylor, T. G. Chem. Ind. 1959, 1223.
(10) Curphey, T. J .; Santer, J . O.; Rosenblum, M.; Richards, J . H.
J . Am. Chem. Soc. 1960, 82, 5249.

1118 Organometallics, Vol. 16, No. 6, 1997
Cunningham
latter yielded benzene. A more detailed study, pub-
lished over 20 years later, discounted the formation of
intermediate aluminates by demonstrating that the
benzene content of the product mixture was the same
before and after hydrolysis.13 To account for the product
mixture of diphenylmercury, benzene, and biphenyl,
these authors suggested an initial electron transfer from
LiAlH4 to phenylmercuric chloride, which provides the
corresponding radical anion. Subsequent loss of chlo-
ride anion and elemental mercury provides a phenyl
radical. Interaction of the phenyl radical with phen-
ylmercuric chloride eventually yields diphenylmercury,
whereas hydrogen abstraction from LiAlH4 affords
benzene. Biphenyl, obviously, is obtained via dimer-
ization of the phenyl radicals. Absent from this analysis
is phenylmercuric hydride: the formation of this product
was deemed thermodynamically unfavorable. This,
however, is in stark contrast to the recent preparation
and NMR analysis of phenylmercuric hydride and
related species at low temperature.14 In fact, the
decomposition of phenylmercuric hydride was shown to
afford diphenylmercurysone of the products observed
in the aforementioned study.
The mercuration of ferrocene with mercuric trifluo-
roacetate is best carried out in the presence of mercuric
oxide.15 Mercuric oxide reacts with the trifluoroacetic
acid released upon mercuration to regenerate mercuric
trifluoroacetate, thereby preventing proton exchange
and/or oxidation of ferrocene. Reaction of 1 with 0.1
equiv of Hg(O2CCF3)2 and 0.1 equiv of HgO under the
same conditions employed for the acetoxymercuration
described above furnished 2b, after ligand exchange
with lithium chloride (Scheme 3). The yield of 2b was
70% based on the total mercury(II) employed and 14%
1
based on 1. The latter was recovered in 74% yield. H
NMR analysis of 2b indicated a high preference for
attack of the C5H5 ring of 1. Nonetheless, mercury
trifluoroacetate is slightly less selective than the cor-
responding acetate, as would be expected: the ratio of
C5H5 to C5D5 attack is 6.2:1. In light of the results
obtained with mercuric acetate, the high isotope effect
suggests a mechanism based on the endo attack of the
mercury(II) species. This is also confirmed by mass
spectral analysis of 3b and 4b: no evidence of a proton
transfer can be detected.
16
The mercuration of 1 with Hg(O3SCF3)2 was inves-
Our observations indicate that, at least in the case of
chloromercurioferrocene, the reduction is not uniquely
radical in nature. In separate trials, 2a was treated
with LiAlD4 in THF-d8 and hydrolyzed with either H2O
or D2O (eq 2). Workup with H2O afforded ferrocene
tigated in a similar manner (Scheme 3). The high
isotope effect of 5.2, although less pronounced than in
the two previous examples, 6.2 with Hg(O2CCF3)2 and
6.5 with Hg(O2CCH3)2, indicates that the desired change
in the mechanism from an endo to exo attack did not
occur. Apparently, regardless of the counterion bound
to mercury, the energy required for the formation of 5
is much too high in relation to that for the endo attack
of mercury(II) with direct exo loss of a proton.
having a deuterium content of 5.04 atoms, whereas a
much higher deuterium content (5.76 atoms: +0.72 D)
resulted from the D2O workup. In contrast, reduction
of 2a with LiAlH4 followed by H2O hydrolysis led to
ferrocene having 4.84 deuterium atoms. This suggests
that approximately 22% of 2 undergoes reduction via
the ferrocenyl radical, either through hydrogen (deute-
rium) abstraction from the reducing agent or from THF.
The vast majority of 2, however, is converted to an
anionic derivative, be it an alanate or ferrocenyllithium,
which is quenched with water.
Having established that acetoxymercuration of fer-
rocene is achieved via the endo attack of the electro-
phile, we then considered the use of mercurating agents
of higher electrophilicity. It is conceivable that a very
reactive mercury(II) species would attack exo to the
cyclopentadienyl ring to afford the iron hydride 5. The
high instability of the mercury(II) fragment could
disfavor the reverse reaction that regenerates the start-
ing materials and lead to the preferential loss of a
proton (via iron to ring transfer) to afford 2.
In a last attempt to force a change in the mechanistic
course of mercuration, we reevaluated the reaction of 1
with Hg(O2CCF3)2 in toluene. The replacement of an
alcohol by an aromatic solvent leads to higher conver-
sions in the vinyl mercuration of 1,1-diarylethylenes.17
This effect has been ascribed to the enhanced electro-
philicity of the mercury species due to weaker solvent-
mercury interactions. Addition of 1.0 equiv of solid
Hg(O2CCF3)2 to a solution of 1 in toluene at -42 °C led
to the precipitation of deep red solid over a 2 h period
(Scheme 3). Quick transfer of this suspension into
vigorously stirred aqueous LiCl resulted in a color
change from red to yellow-orange. After purification, a
19% yield of 2d was obtained along with the recovery
of 80% of unreacted 1 (∼100% mass balance). Analysis
of 2d in the usual way suggested that under these
conditions, the endo mode of electrophilic substitution
was again preferred. The observed isotope effect was
6.6, slightly higher than that of the corresponding
(15) Han, Y.-H.; Heeg, M. J .; Winter, C. H. Organometallics 1994,
13, 3009.
(16) Deacon, G. B.; Tunaley, D. J . Organomet. Chem. 1978, 156, 403.
(17) Sokolov, V. I.; Bashilov, V. V.; Reutov, O. A. J . Organomet.
Chem. 1978, 162, 271.
(13) Singh, P. R.; Khana, R. K. Tetrahedron Lett. 1983, 1411.
(14) (a) Bellec, N.; Guillemin, J .-C. Tetrahedron Lett. 1995, 6883.
(b) Kwetkat, K.; Kithching, W. J . Chem. Soc., Chem. Commun. 1994,
345.

Mechanism of Mercuration of Ferrocene
Organometallics, Vol. 16, No. 6, 1997 1119
Sch em e 6
reaction in 1:1 ethanol:toluene at room temperature. It
appears that the enhanced electrophilicity of Hg(O2-
CCF3)2 in pure toluene is more than compensated for
by the decrease in temperature. Attempts to carry this
reaction out at room temperature, however, led to the
oxidation of 1.
To develop a better understanding of the course of the
mercuration of ferrocene with Hg(O2CCF3)2 in toluene,
the reaction was repeated on a larger scale with fer-
rocene-d0. The red solid was filtered at low temperature
and washed several times with cold toluene (-42 °C).
The cold toluene filtrate was treated with LiCl, as
previously described. From the organic phase, ferrocene
was recovered in 22% yield, as well as a trace of
chloromercurioferrocene (<0.5%). Concentration of the
aqueous phase afforded a white solid containing less
than 0.1 mg of mercury, representing far less than 0.1%
of the mercury employed in the reaction. Therefore, the
red solid contains ferrocene derivatives and mercury
species in a 4:5 ratio. Taking into account that mercu-
ration must already occur at -42 °C (chloromercurio-
ferrocene is detected in the filtrate) and that mercura-
tion proceeds in ∼20% yield (previous experiment), the
composition of the red solid must be a 3:1 mixture of
the ferrocene-Hg(O2CCF3)2 complex and the trifluoro-
acetoxymercurioferrocene-Hg(O2CCF3)2 complex 6 (eq
3).18 When the solid is allowed to warm, a color change
from red to blue occurs at 0 °C (oxidation). Treatment
of the blue oily substance with aqueous LiCl then affords
11% chloromercurioferrocene (based on total ferrocene
employed) and 26% of recovered ferrocene; the remain-
ing 41% of the ferrocene derivatives undergo oxidation.
Combining these results with those obtained with 1
allows for the conception of a mechanism for the
mercuration, which includes precomplexation of the
electrophile at the metal center. The treatment of
ferrocene with a mercury(II) species establishes an
equilibrium between free ferrocene and complex A
(Scheme 6). In the rate determining step, formation of
the carbon-mercury bond is achieved via a metal to ring
transfer of Hg(II) and dissociation of a mercury ligand.
This process is either concomitant with or prior to an
exo loss of a proton to afford the substituted product.
The aforementioned lack of correlation of the rate of
mercuration with the number of ferrocene alkyl sub-
stituents can be attributed to increased steric repulsions
during the iron to carbon migration of the mercurating
species.3f It should also be noted that the kinetic study
in question was monitored by UV spectroscopy, and no
thorough product analysis was carried out. As ferrocene
readily undergoes polymercuration,9,15 monomercurated
ferrocene can also compete with ferrocene for complex-
ation with the free mercurating agent to furnish B.
Depending upon the reaction conditions (at low tem-
peratures or in apolar solvents), the precipitation of
dimers (oligomers) of complexes A and B may occur. In
fact, the red complexes of ferrocene and its derivatives
with various mercuric halides have been isolated and
studied by various techniques.5-7 In alcoholic solvents,
the precipitation of dimers of complexes A and B or
compounds of other stoichiometry is disfavored and the
mercuration proceeds as expected. It is apparent,
however, that the solubility of ferrocene-HgX2 adducts
depends on both the solvent and the nature of X. The
mercuration of ferrocene with Hg(O2CCF3)2 in toluene
is unique in that both dimerization (oligomerization) of
the complexes and mercuration occur simultaneously.
Under identical conditions, the reaction of with Hg(O3-
SCF3)2 affords only dimers (oligomers) of complex A (X
) O3SCF3).
(18) Both mercuric trifluoroacetate and ferrocene are perfectly
soluble in toluene at this temperature; therefore, it is unlikely that
the precipitate contains any uncomplexed mercury species.
In view of our previous work on the electrophilic

1120 Organometallics, Vol. 16, No. 6, 1997
Cunningham
Sch em e 7
substitution of ferrocene, this mercuration study per-
mits a general understanding of the mechanisms of such
reactions. In effect, the course of these reactions is very
similar to that first proposed by Traylor almost 30 years
ago, with the important exception that the iron atom
plays an essential role (Scheme 7). Highly reactive
electrophiles, typified by RCOCl-AlCl3 (and possibly
RX-AlCl3), undergo an exo attack of the cyclopentadi-
enyl ring of ferrocene to afford an iron hydride inter-
mediate C. Various cationic iron hydride species,3d,10,19
including those related to the intermediate proposed for
acetylation of ferrocene, C (E ) C+(OH)CH3),4b,20 have
been characterized by 1H NMR. These reactions, being
relatively exothermic, should have early transition
states characterized by a low extent of carbon-electro-
phile bond formation and, therefore, exhibit no isotope
effect. As shown by Rosenblum,3b the attack of such
electrophiles may also occur at the metal center (endo)
if an exo attack is excluded by steric constraints.
Deprotonation of C then proceeds via metal to carbon
proton transfer with the exo loss of H+. As previously
discussed, the occurrence of these interannular proton
transfers is proof in itself that the proton exchange of
ferrocene proceeds via an exo attack of H+. Further
evidence is provided by the formation of predominately
H2 through oxidative deuteration of ferrocenophane in
BF3-D2O.
an exo attack, is unstable with respect to elimination
of the electrophile.
Whereas this interpretation is valid for ferrocene
itself, the mode of electrophilic substitution of ferrocene
derivatives depends strongly upon the nature of the
substituent(s). It has been shown by a similar labeling
experiment that the acetylation of acetylferrocene with
CH3COCl-AlCl3 (strong electrophile) must occur via an
endo attack, although it proceeds without an isotope
effect (exothermic reaction with an early transition
state).4b The generation of the iron hydride 8a that
would result from an exo attack is disfavored due to the
low basicity of the iron atom. This is in accord with
With electrophiles that are weaker than H+, for
example mercury(II) or thallium(III), substitution occurs
via precomplexation of the electrophile to afford A
followed by the rate determining endo attack of a
cyclopentadienyl ring of ferrocene. As such reactions
are less exothermic, the extent of bond formation in the
transition state is much greater than that with strong
electrophiles and an isotope effect is observed. The endo
mode is favored not only because of the interaction of
these soft electrophiles (soft acids) with the easily
oxidized iron atom of ferrocene (soft base) but also since
the cationic iron hydride C, which would result from
the lack of exchange of cyclopentadienyl protons of 1,1-
diacetylferrocene with triflic acid-d1 (which should
21
occur via 8b) and the resistance of diacetylferrocene to
further acetylation.22 The addition of two electron-
donating methyl groups to diacetylferrocene, however,
suffices to stabilize the intermediate 8c: a triacetylfer-
rocene has been isolated from the reaction of 1,1′-
(21) Roberts, R. M. G.; Silver, J .; Wells, A. S. Inorg. Chim. Acta 1986,
119, 171. Here again, the bis(aluminum chloride) complex of diacetyl-
ferrocene protonated at the iron is being compared to the triprotonated
species.
(19) Bitterwolf, T. Inorg. Chim. Acta 1986, 117, 55 and other papers
in this series, including ref 3d.
(22) (a) Only diacetylferrocenes were isolated: Nesmeyanov, A.;
Leonova, E. V.; Kochetkova, N. S.; Malkova, A. I. J . Organomet. Chem.
1975, 96, 271 (10 equiv ov CH3COCl-AlCl3, CH2Cl2, RT). (b) Rausch,
M. D.; Fischer, E. O.; Grubert, H. J . Am. Chem. Soc. 1960, 82, 76 (3
equiv CH3COCl-AlCl3, CH2Cl2, reflux). The reported tetraacetylation
of ferrocene with acetic anhydride and trifluoroacetic acid (Sweeney,
US Patent 2,852,542, February 29, 1956) has been revealed to be irre-
producible: Bell, W; Glidewell, C. J . Chem. Res., Synop. 1991, 44.
(20) (a) Rimmelin, P.; Sommer, J .; Sandstro¨m, J .; Seita, J . J .
Organomet. Chem. 1976, 114, 175. (b) Olah, G. A.; Mo, Y. K. J .
Organomet. Chem. 1973, 60, 311. The essential difference between the
two species is that the acid complexed to the carbonyl: the acetylation
intermediate is an aluminum chloride complex of acetylferrocene
protonated at the iron.

Mechanism of Mercuration of Ferrocene
Organometallics, Vol. 16, No. 6, 1997 1121
dimethylferrocene with excess CH3COCl-AlCl3.23 Ar-
guments refuting alternative mechanisms based on the
endo attack of strong electrophiles without precomplex-
ation have been set forth in a previous publication.24
To define the exact point at which the mechanism for
electrophilic substitution changes from exo to endo by
the proton affinity or oxidation potential of the resulting
ferrocene product is difficult: the stability of the iron
hydride intermediate depends not only upon the Lewis
acid species employed for the reaction but also on the
strength of the complex formed between the Lewis acid
and the ferrocene substituent(s).
In conclusion, the reaction of ferrocene with various
mercurating agents has been shown to proceed via
precomplexation of the mercury(II) species at the metal
center followed by the rate determining formation of the
carbon-mercury bond (endo attack). For the first time,
direct evidence was obtained for the formation of an
intermediate ferrocene-HgX2 adduct whose conversion
to a ferrocene having a carbon-mercury bond was
competitive with the precipitation of a mixed ferrocene-
HgX2/ferrocene(HgX)-HgX2 dimer (oligomer). The
mechanistic details of the mercuration complemented
earlier studies based on the acetylation and protonation
of ferrocene derivatives and allowed the formulation of
a general treatment of electrophilic substitution for this
metallocene. Important aspects of this mechanistic
interpretation are the reactivity of the electrophile and
the basicity of the iron atom of the resulting ferrocene
product.
(m/e 126 and 125) or a deuteron (m/e 122 and 121) transfer
has occurred.
Red u ction of 2a w ith LiAlH 4. To a solution of ∼0.02 g of
2a in THF (0.5 mL) at room temperature was added ∼0.05 g
of LiAlH4. Gas was evolved, and the gray suspension was
allowed to stir for 0.5 h. The reaction mixture was quenched
carefully with a minimum amount of H2O and then diluted
with hexane. The organic phase was removed from the solids
and placed directly on a small column of silica gel. The
ferrocene 3a formed was obtained via elution with hexane: MS
(70 eV) m/e (194-188) 194 (0), 193 (1.5), 192 (13.5), 191 (100),
190 (14), 189 (8), 188 (2); m/e (126-121, normalized to 100)
126 (81), 125 (23.5), 124 (10), 123 (5.5), 122 (17), 121 (100).
Red u ction of 2a w ith LiAlD4. This was carried out as
described for the reduction with LiAlH4, with the exception
that THF-d8 was used as the solvent and that quenching was
accomplished with D2O. 4a : MS (70 eV) m/e (194-188) 194
(1.5), 193 (14.5), 192 (100), 191 (20.5), 190 (8), 189 (2), 188
(0.5); m/e (126-121, normalized to 100) 126 (100), 125 (12.5),
124 (11.5), 123 (16.5), 122 (86.5), 121 (17.5).
Mer cu r a tion of 1 w ith Mer cu r ic Tr iflu or oa ceta te.
This reaction was performed as described above for mercuric
acetate, with the exception that 0.04 g (0.0001 mol) of mercuric
trifluoroacetate and 0.02 g (0.0001 mol) of yellow mercuric
oxide were employed. Chromatography gave 0.14 g (74%
recovery) of unchanged 1 and 0.06 g (14% based on 1, 74%
based on mercury(II)) of 2b: 1H NMR (CDCl3) δ 4.47 (br t, 2,
J ) 1.5 Hz, mercury satellites at 4.53 and 4.41), 4.24 (s, 0.81),
4.11 (br t, 2, J ) 1.5 Hz, mercury satellites at 4.21 and 4.01)
ppm.
Compounds 3b and 4b were obtained as described above
for 3a and 4a . MS (70 eV) of 3b: m/e (194-188) 194 (0), 193
(1), 192 (13.5), 191 (100), 190 (19), 189 (8), 188 (2); m/e (126-
121, normalized to 100): 126 (88), 125 (26), 124 (11), 123 (7),
122 (18), 121 (100%). MS (70 eV) of 4b: m/e (194-188) 194
(1), 193 (13.5), 192 (100), 191 (26.5), 190 (9), 189 (3), 188 (1);
m/e (126-121, normalized to 100) 126 (100), 125 (15), 124 (12),
123 (16), 122 (82), 121 (21%).
Mer cu r a tion of 1 w ith Mer cu r ic Tr iflu or om eth ylsu l-
fon a te. This reaction was performed as described above for
mercuric acetate, with the exception that 0.38 g (0.002 mol)
of 1, 0.10 g (0.0002 mol) of mercuric trifluoroacetate, and 0.06
g (0.0003 mol) of yellow mercuric oxide in 6 mL of 1:1 toluene:
ethanol were employed. Chromatography gave 0.30 g (80%
recovery) of unchanged 1 and 0.02 g (14% based on 1, 9% based
on mercury(II)) of 2c: 1H NMR (CDCl3) δ 4.47 (br t, 2, J ) 1.5
Hz, mercury satellites at 4.53 and 4.41), 4.24 (s, 0.95), 4.11
(br t, 2, J ) 1.5 Hz, mercury satellites at 4.21 and 4.01) ppm.
Given the high isotope effect upon mercuration, the analysis
for proton transfers (reduction of 2c with LiAlH(D)4 and mass
spectral analysis) was deemed unnecessary.
Exp er im en ta l Section
Gen er a l Meth od s. The solvents were used as obtained
from the suppliers with the exception of toluene which was
distilled from CaH2. The preparation of (pentadeuteriocyclo-
pentadienyl)cyclopentadienyliron(II) has been previously
described.4b Routine monitoring of the reactions was carried
out with glass-backed TLC plates of Merck 60 F254 silica gel.
Flash column chromatography was performed on Merck 60H
F254 silica gel. 1H NMR spectra were recorded on a Brucker
250 AC (250 MHz) or a Brucker 300 AC spectrometer (300
MHz). Chemical shifts are reported in ppm relative to internal
tetramethylsilane. Mass spectra were recorded on a Finigan
MAT 212-SS300 spectrometer at 70 eV.
Mer cu r a tion of 1 w ith Mer cu r ic Aceta te. To a solution
of 0.19 g (0.001 mol) of 1 in 1:1 toluene:ethanol (3 mL) was
added 0.16 g (0.0005 mol) of solid mercuric acetate. The
resulting solution was allowed to stir for 24 h at room
temperature. At this time a solution of 0.20 g (0.047 mol) of
LiCl in H2O (10 mL) was added, and the reaction mixture was
stirred vigorously for 0.5 h. The reaction mixture was diluted
with H2O (∼50 µL) and extracted with CH2Cl2 (3 × 20 mL).
The combined organic extracts were dried (Na2SO4) and
concentrated to give a yellow solid. Chromatography (SiO2,
2:1 hexane:CH2Cl2) gave 0.11 g (58% recovery) of 1 and 0.15 g
(35% based on 1, 70% based on mercuric acetate) of 2a : 1H
NMR (CDCl3) δ 4.47 (br t, 2, J ) 1.5 Hz, mercury satellites at
4.53 and 4.41), 4.24 (s, 0.77), 4.11 (br t, 2, J ) 1.5 Hz, mercury
satellites at 4.21 and 4.01) ppm. The mass spectrum of
recovered 1 is identical to that of the starting material: MS
(70 eV) m/e (194-188) 194 (0), 193 (1.5), 192 (13.5), 191 (100),
190 (4), 189 (7.5), 188 (1); m/e (126-121, normalized to 100)
126 (100), 125 (13.5), 124 (10), 123 (5), 122 (17.5), 121 (98%).
The italicized values are used to determine whether a proton
Mer cu r a tion of 1 w ith Mer cu r ic Tr iflu or oa ceta te in
Tolu en e a t -42 °C. To a solution of 0.19 g (0.001 mol) of 1
in toluene (5 mL) maintained at -42 °C by means of a dry
ice-acetonitrile bath was added 0.43 g (0.001 mol) of solid
mercuric trifluoroacetate. As the mercuric trifluoroacetate
slowly dissolved, a deep red solid formed. The reaction
mixture was allowed to stir for 2.5 h at -42 °C, at which time
it was poured into a vigorously stirred solution of 0.20 g (0.047
mol) of LiCl in H2O (10 mL). The solution continued to stir
for 0.5 h. The toluene phase was separated, and the aqueous
phase was extracted with toluene (2 × 20 mL). The combined
organic extracts were dried (MgSO4) and concentrated to give
a yellow solid. Chromatography (SiO2, 2:1 hexane:CH2Cl2)
gave 0.17 g (80%) of unchanged 1 and 0.08 g (19%) of 2d : 1H
NMR (CDCl3) δ 4.47 (br t, 2, J ) 1.5 Hz, mercury satellites at
4.53 and 4.41), 4.24 (s, 0.75), 4.11 (br t, 2, J ) 1.5 Hz, mercury
satellites at 4.21 and 4.01) ppm.
Compounds 3d and 4d were obtained as described above
for 3a and 4a . MS (70 eV) of 3d : m/e (194-188) 194 (0), 193
(1), 192 (13.5), 191 (100), 190 (22), 189 (8.5), 188 (2.5); m/e
(126-121, normalized to 100) 126 (87), 125 (28), 124 (11), 123
(23) Nesmeyanov, A. N.; Perevalova, E. G.; Beinoravichute, Z. A.;
Malygina, I. L. Dokl. Akad. Nauk. SSSR 1958, 120, 1263.
(24) See ref 3a, footnote 17.

1122 Organometallics, Vol. 16, No. 6, 1997
Cunningham
(6.5), 122 (18), 121 (100). MS (70 eV) of 4d : m/e (194-188)
194 (1), 193 (14), 192 (100), 191 (25.5), 190 (9), 189 (2.5), 188
(1); m/e (126-121, normalized to 100) 126 (100), 125 (15), 124
(12), 123 (15), 122 (82), 121 (23.5).
at -42 °C, was also poured into a vigorously stirred solution
of 1.00 g (0.0235 mol) of LiCl in H2O (50 mL). The solution
continued to stir for 0.5 h. The organic phase was separated,
dried (MgSO4), and concentrated to give an orange solid.
Chromatography (SiO2, 2:1 hexane:CH2Cl2) gave an additional
0.20 g (22%) of ferrocene and a trace (<0.01 g) of chloromer-
curioferrocene. The aqueous phase was concentrated to give
0.83 g of a white solid. The mercury content of this solid was
110 mg/kg, i.e., the solid contains less than 0.01 mg of mercury
(it is predominately LiCl).
A similar reaction was carried out with mercuric trifluo-
romethylsulfonate. The red suspension was directly poured
into LiCl/H2O. Only ferrocene, no trace of chloromercuriofer-
rocene, could be detected by TLC analysis, and the blue color
of the aqueous phase indicated that oxidation of ferrocene had
occurred.
Mer cu r a tion of F er r ocen e-d 0 w ith Mer cu r ic Tr iflu o-
r oa ceta te in tolu en e a t -42 °C. To a solution of 0.93 g
(0.005 mol) of ferrocene in toluene (25 mL) maintained at -42
°C by means of a dry ice-acetonitrile bath was added 2.13 g
(0.005 mol) of solid mercuric trifluoroacetate. As the mercuric
trifluoroacetate slowly dissolved, a deep red solid formed. The
reaction mixture was allowed to stir for 2.5 h at -42 °C and
then filtered under argon through a externally-cooled (dry ice)
sintered glass filter. The red solid (6) was washed several
times with cold (-42 °C) toluene and then dried under vacuum,
while warming to room temperature. At ∼0 °C, the red solid
became a blue viscous oil; the weight of the oil was 2.70 g.
This oil was dissolved in CH2Cl2 (50 mL) and poured into a
vigorously stirred solution of 1.00 g (0.0235 mol) of LiCl in
H2O (50 mL). The solution continued to stir for 2.5 h. The
organic phase was separated, dried (Na2SO4), and concentrated
to give an orange solid. Chromatography (SiO2, 2:1 hexane:
CH2Cl2) gave 0.25 g (27%) of ferrocene and 0.24 g (11%) of
chloromercurioferrocene. The filtrate, which was maintained
Ack n ow led gm en t. The author is indebted to Ciba-
Geigy Ltd. for the permission to carry out and publish
this work and to Dr. U. Mareis for the measurement of
the mass spectra.
OM960815+