Three-component synthesis of substituted η5-cyclopentadienyltricarbonylrhenium complexes: Scope, limitations, and mechanistic interpretations

Article abstract of DOI:10.1021/om990128t

We have investigated the scope and mechanism of the three-component synthesis of substituted CpRe(CO)₃ complexes which involves the reaction of nucleophiles with diazocyclopentadiene (C₅H₄N₂) and a fac-Re(CO)₃⁺ species. We found that only moderately strong nucleophiles (halogen, carboxylates, boronic acids) are suitable for this transformation and that it shows a great sensitivity to the steric and electronic features of the nucleophile. A Hammett-type ρσ analysis of the effect of para-substituents on the relative rate of this reaction with several benzoates showed that the reaction is accelerated by electron-donating substituents. A mechanistic analysis, based on structure/reactivity relationships and NMR experiments, indicated that the nucleophile initially reacts with the rhenium precursor. Then, in the rate-determining step, the resulting preassociated rhenium-nucleophile intermediate reacts with C₅H₄N₂ via a concerted S_N2-like transition state. The same general mechanistic pathway seems to be followed by two very different classes of nucleophiles, carboxylates and boronic acids, in the synthesis of acyloxy- and carbon-substituted CpRe(CO)₃ complexes, respectively. In particular, the lack of reactivity of boronic esters can be explained by the necessary preassociation step between the rhenium and a deprotonated hydroxy group of the nucleophile, which is possible only with free boronic acids.

Full text of DOI:10.1021/om990128t

Organometallics 1999, 18, 2519-2530
2519
Th r ee-Com p on en t Syn th esis of Su bstitu ted
η5-Cyclop en ta d ien yltr ica r bon ylr h en iu m Com p lexes:
Scop e, Lim ita tion s, a n d Mech a n istic In ter p r eta tion s
Filippo Minutolo and J ohn A. Katzenellenbogen*
Department of Chemistry, University of Illinois, 600 S. Mathews Avenue,
Urbana, Illinois 61801
Received February 23, 1999
We have investigated the scope and mechanism of the “three-component” synthesis of
substituted CpRe(CO)3 complexes which involves the reaction of nucleophiles with diazo-
cyclopentadiene (C5H4N2) and a fac-Re(CO)3+ species. We found that only moderately strong
nucleophiles (halogen, carboxylates, boronic acids) are suitable for this transformation and
that it shows a great sensitivity to the steric and electronic features of the nucleophile. A
Hammett-type Fσ analysis of the effect of para-substituents on the relative rate of this
reaction with several benzoates showed that the reaction is accelerated by electron-donating
substituents. A mechanistic analysis, based on structure/reactivity relationships and NMR
experiments, indicated that the nucleophile initially reacts with the rhenium precursor. Then,
in the rate-determining step, the resulting preassociated rhenium-nucleophile intermediate
reacts with C5H4N2 via a concerted SN2-like transition state. The same general mechanistic
pathway seems to be followed by two very different classes of nucleophiles, carboxylates
and boronic acids, in the synthesis of acyloxy- and carbon-substituted CpRe(CO)3 complexes,
respectively. In particular, the lack of reactivity of boronic esters can be explained by the
necessary preassociation step between the rhenium and a deprotonated hydroxy group of
the nucleophile, which is possible only with free boronic acids.
In tr od u ction
stable chemical species, and it must have a structure
that does not interfere with the receptor-binding pro-
cess. With regard to these issues, very favorable proper-
ties are shown by cyclopentadienyl tricarbonyl ReI and
TcI complexes CpM(CO)3 (1, Figure 1),4 especially when
they are compared to the more widely used inorganic
chelate complexes (2). First, CpM(CO)3 complexes 1 are
chemically and metabolically stable, which allows for
easy handling and effective in vivo distribution. Second,
they are lipophilic and relatively small, so that they are
less likely to interfere with the receptor-ligand binding
process, either through polarity or by steric hindrance.
Third, unlike many inorganic metal-oxo chelates (2),
they do not possess additional stereocenters, which often
complicate the purification process of radiopharmaceu-
ticals.5
Many metallic radionuclides are nowadays routinely
used in nuclear medicine.1 Some of the most widely used
ones are technetium-99m for diagnostic imaging and
rhenium-186 and -188 for radiotherapy.2 It is convenient
that the elements technetium and rhenium have very
similar chemical behaviors, so that the stable natural
isotopes 185Re and 187Re can be reliably studied as
models not only for the radioisotopes 186Re/188Re but also
for 99mTc.3
The synthesis of radiopharmaceuticals possessing
specific receptor-binding properties presents a special
challenge. Such agents are usually prepared by radio-
labeling organic molecules, which act as ligands for
specific biological receptors. A critical element in a
receptor-binding radiopharmaceutical is the nature of
the radionuclide attached to the ligand: it must be a
Despite these favorable characteristics, however, the
use of organometallic systems such as 1 conjugated with
bioactive molecules has, so far, been rather limited.6
Considerable effort has been devoted to the development
of new synthetic methods suitable for radiolabeling
biologically interesting molecules with this class of
* Corresponding author: Department of Chemistry, University of
Illinois, 461 Roger Adams Laboratory, Box 37-5, 600 S. Mathews
Avenue, Urbana, IL 61801. Telephone: (217) 333-6310. Fax: (217) 333-
7325. E-mail: jkatzene@uiuc.edu.
(1) J urisson, S.; Berning, D.; J ia, W.; Ma, D. Chem. Rev. 1993, 93,
1137-1156.
(2) (a) Dilworth, J . R.; Parrott, S. J . Chem. Soc. Rev. 1998, 27, 43-
55. (b) Hom, R. K.; Katzenellenbogen, J . A. Nucl. Med. Biol. 1997, 24,
485-498. (c) Schwochau, K. Angew. Chem., Int. Ed. Eng. 1994, 33,
2258-2267. (d) Lisic, E. C.; Mirzadeh, S.; Knapp, F. F., J r. J . Labeled
Compd. Radiopharm. 1993, 33, 65-75. (e) J ohn, E.; Thakur, M. L.;
DeFulvio, J .; McDevitt, M. R.; Damjanov, I. J . Nucl. Med. 1993, 34,
260-267.
(3) Boog, N. M.; Kaesz, H. D. Technetium and Rhenium. In
Comprehensive Organometallic Chemistry; Wilkinson, G., Stone, G. F.
A., Abel, E. W., Eds.; Pergamon Press: Oxford, U.K., 1982; Vol. 4, pp
161-242, and references therein.
(4) Katzenellenbogen, J . A.; Minutolo, F.; Spradau, T. W.; Skaddan,
M. B. In Technetium and Rhenium in Chemistry and Nuclear Medicine
5; Nicolini, M., Bandoli, G., Mazzi, U., Eds.; SGE Editoriali: Padova,
Italy, in press.
(5) See, for example: Hom, R. K.; Katzenellenbogen, J . A. J . Org.
Chem. 1997, 62, 6290-6297, and references therein.
(6) (a) Top, S.; El Hafa, H.; Vessie`res, A.; Quivy, J .; Vaissermann,
J .; Hughes, D. W.; McGlinchey, M. J .; Mornon, J .-P.; Thoreau, E.;
J aouen, G. J . Am. Chem. Soc. 1995, 117, 8372-8380. (b) Spradau, T.
W.; Edwards, W. B.; Anderson, C. J .; Welch, M. J .; Katzenellenbogen,
J . A. Nucl. Med. Biol. 1999, 26, 1-7.
10.1021/om990128t CCC: $18.00 © 1999 American Chemical Society
Publication on Web 06/06/1999

2520 Organometallics, Vol. 18, No. 13, 1999
Minutolo and Katzenellenbogen
Ta ble 1. Yield s Obta in ed w ith Differ en t
Heter oa tom Nu cleop h iles 5a -g (eqs 1 a n d 2)a
F igu r e 1. CpM(CO)3 complex (1) and a generic metal(V)-
oxo inorganic chelate structure (2).
Sch em e 1. Th r ee-Com p on en t Syn th etic Ap p r oa ch
to Su bstitu ted Cp Re(CO)3 Com p lexes
organometallic systems. The first success in this en-
deavor was achieved by the “double ligand transfer”
reaction,7 a process which, despite high efficiency, has
some drawbacks such as harsh conditions and substrate
limitations.
Our goal has been to develop fast, efficient, and
generally applicable synthetic methods to prepare ra-
diopharmaceuticals containing CpM(CO)3 functional
groups. As a starting point, a “three-component” ap-
proach (Scheme 1),8 which involved the simultaneous
formation of the Cp-metal and the Cp-nucleophile
bonds, was considered. Diazocyclopentadiene (C5H4N2)9
was expected to be an ideal Cp precursor. Because it
readily loses molecular nitrogen, it can be considered
to be a bifunctional synthon, able to donate six π-elec-
trons to the metal center and to undergo nucleophilic
substitution at the ring σ-site. We also wanted to use
metal precursors which could be easily obtained from
perrhenate and pertechnetate, because these are the
chemical forms delivered from generators of Re-186/188
and Tc-99m, respectively. We were particularly at-
tracted by fac-(CH3CN)3Re(CO)3+ and fac-(CH3CN)3Tc-
a
Conditions: 4 (0.050 mmol), nucleophile (0.10 mmol), diazo-
cyclopentadiene (0.060 mmol), CH3CN (2.5 mL), 80 °C for 45 min.
Yields are based on the initial rhenium precursor 3. c In this
b
reaction, 3 was directly used, without treatment with silver
+
triflate, to generate 4. The Re(CO)3 precursor contained 0.15
mmol of Br-.
substituted CpRe(CO)3 complexes).12 Here, we present
a detailed analysis of structure/reactivity relationships
in the nucleophilic component of this synthesis, spec-
troscopic evidence for reaction intermediates, and a
discussion of the possible mechanistic pathway followed
by this reaction.
+
(CO)3 species, whose efficient preparations had been
recently reported,10 because they already possess the
three CO ligands in the proper geometrical orienta-
tion. In addition to its easy synthetic accessibility, this
+
(CH3CN)3Re(CO)3 species proved to be much more
Resu lts a n d Discu ssion
reactive than pentacarbonyl halides,9a when used with
C5H4N2 in the synthesis of halogen-substituted CpRe-
(CO)3.11
Heter oa tom Nu cleop h iles. The reactions of a large
number of heteroatom nucleophilic species 5a -g (Table
1) with in situ generated [(CH3CN)3Re(CO)3]+TfO- (4,
eq 1) and C5H4N2 were examined in acetonitrile at 80
°C (eq 2).11
The nucleophilic species used in this sequence which
gave the most promising results, in terms of their
potential use in nuclear medicine, were organic car-
boxylates (which afford acyloxy-substituted CpRe(CO)3
complexes)11 and boronic acids (which afford carbon-
(7) (a) Spradau, T. W.; Katzenellenbogen, J . A. Organometallics
1998, 17, 2009-2017. (b) Wenzel, M. J . Labeled Compd. Radiopharm.
1992, 31, 641-650.
(8) Minutolo, F.; Katzenellenbogen, J . A. In Technetium and Rhe-
nium in Chemistry and Nuclear Medicine 5; Nicolini, M., Bandoli, G.,
Mazzi, U., Eds.; SGE Editoriali: Padova, Italy, in press.
(9) (a) Herrmann, W. A. Chem. Ber. 1978, 111, 2458-2460. (b)
Reimer, K. J .; Shaver, A. J . Organomet. Chem. 1975, 93, 239-252.
(10) (a) Alberto, R.; Schibli, R.; Schubiger, P. A. Polyhedron 1996,
15, 1079-1089. (b) Alberto, R.; Schibli, R.; Egli, A.; Schubiger, P. A.;
Herrmann, W. A.; Artus, G.; Abram, U.; Kaden, T. A. J . Organomet.
Chem. 1995, 492, 217-224. (c) Alberto, R.; Egli, A.; Abram, U.;
Hegetschweiler, K.; Gramlich, V.; Schubiger, P. A. J . Chem. Soc.,
Dalton Trans. 1994, 2815-2820. (d) Alberto, R.; Schibli, R.; Egli, A.;
Schubiger, P. A.; Abram, U.; Kaden, T. A. J . Am. Chem. Soc. 1998,
120, 7987-7988.
Initially, we obtained reasonably good yields with
halide ions (entries 1, 2, Table 1). However, the halogen-
substituted CpRe(CO)3 species derived from these mono-
valent nucleophiles would require additional function-
alization, via palladium-mediated C-C coupling reac-
(11) Minutolo, F.; Katzenellenbogen, J . A. J . Am. Chem. Soc. 1998,
120, 4514-4515.
(12) Minutolo, F.; Katzenellenbogen, J . A. J . Am. Chem. Soc. 1998,
120, 13264-13265.

η5-Cyclopentadienyltricarbonylrhenium Complexes
Organometallics, Vol. 18, No. 13, 1999 2521
Ta ble 2. Yield s Obta in ed w ith Differ en t Bor on ic
Acid s 7a -g (eqs 1 a n d 3)
F igu r e 2. Cyclic voltammogram recorded at a platinum
electrode in a CH3CN solution (0.5 mM, vs Ag/AgCl
reference electrode) of ferrocene-conjugated CpRe(CO)3
complex 6g (scan rate 0.1 V s-1).
tions,6a,13 in order for it to become attached to a suitable
receptor ligand. To overcome this limitation, we first
extended this transformation by the use of organic
carboxylates (5c-g), which gave good yields of acyloxy-
substituted CpRe(CO)3 systems (entries 3-7, Table 1).
This reaction tolerated the presence of protic hydrogens,
so carboxylate species could be generated in situ by
treating the corresponding carboxylic acid with triethy-
lamine (see Experimental Section). Alcohol, phenol,
carbonyl, and amide-NH substituents in the carboxylate
nucleophile are also tolerated. This tolerance of active
protons provides a great synthetic advantage, because
it allows this reaction to be conducted on unprotected
biologically active organic molecules.
a
Yields are based on the initial rhenium precursor 3 (eq 1).
b
Conditions A: 4 (0.050 mmol), boronic acid (0.50 mmol), Et3N
(1.0 mmol), diazocyclopentadiene (0.060 mmol), CH3CN (2.5 mL),
80 °C for 45 min. c Conditions B: 4 (0.050 mmol), boronic acid (0.10
mmol), Et3N (0.20 mmol), diazocyclopentadiene (0.060 mmol),
CH3CN (2.5 mL), 80 °C for 14 h.
oxidation of the iron center15 with E1/2(Fc/Fc+) ) +0.75
V (vs Ag/AgCl). Oxidation of the rhenium was not
observed at the potentials applied (up to +1.4 V). The
redox characteristics of 6g make it (or its derivatives)
a potential candidate for residualizing radionuclides in
hypoxic tissues,16 which are found in various pathologi-
cal states (solid tumors, stroke, ischemia, cardiovascular
obstructions, etc.).
Ca r bon Nu cleop h iles. To form a connection be-
tween the organometallic complex and the organic unit
in potential radionuclides that would be more robust
than the ester linking group present in 6c-g (Table 1),
we considered using carbon nucleophiles in the “three-
component” reaction. If successful, this would allow
us to form directly a stable C-C bond as the linking
group, which, unlike the esters, would not be prone
to undergo hydrolytic or metabolic cleavage in vivo. By
a preliminary screening of classical organometallic
species (organolithium, organosilicon, organocopper), we
found that none of them gave any of the desired Cp
complex, presumably because they react irreversibly
with the rhenium center or with its carbonyl ligands.
Only very weak organometallic reagents, such as bo-
ronic acids (7a -f, eq 3, and Table 2), proved to have
the proper reactivity for the synthesis of carbon-
substituted CpRe(CO)3 complexes 8a -f under our reac-
tion conditions.12
Although the formation constants of various nucleo-
philes with the [Re(CO)3]+ cation are not known, a
general characteristic of these reactive nucleophiles
(halides and carboxylates) that work well in this reac-
tion appears to be their “moderate” nucleophilicity. The
fact that stronger nucleophilic reagents failed to give
the desired Cp complex suggests that they underwent
irreversible interactions with the metal center or with
the coordinated carbonyl groups. Some “strong” nucleo-
philes that did not work in this reaction are primary
amines, thiols, hydrides, hydroxide ions,14 and alkox-
ides. On the other hand, “weaker” nucleophiles (e.g.,
alcohols, phenols, triflate, and nitrate ions) underwent
no reaction with the other two components.
The last entry of Table 1 shows that a novel bimetallic
complex (6g), containing a ferrocene moiety bound to
the CpRe(CO)3 unit, could be efficiently prepared by this
method in one step from the commercially available
carboxylic acid 5g. Cyclic voltammetry analysis of
complex 6g (Figure 2) showed the well-known reversible
(13) (a) Lo Sterzo, C.; Miller, M. M.; Stille, J . K. Organometallics
1989, 8, 2331-2337. (b) Lo Sterzo, C.; Stille, J . K. Organometallics
1990, 9, 687-694.
(15) Zanello, P. Electrochemical and X-ray Structural Aspects of
Transition Metal Complexes Containing Redox-Active Ferrocene
Ligands. In Ferrocenes; Togni, A., Hayashi, T., Eds.; VCH Publishers:
New York, 1995; Chapter 7.
(16) Uccelli, L.; Bolzati, C.; Boschi, A.; Duatti, A.; Morin, C.;
Pasqualini, R.; Giganti, M.; Piffanelli, A. Nucl. Med. Biol. 1999, 26,
63-67, and references therein.
(14) A remarkably high affinity of the solvated cation Re(CO)3+ for
OH- has been reported in: Egli, A.; Hegetschweiler, K.; Alberto, R.;
Abram, U.; Schibli, R.; Hedinger, R.; Gramlich, V.; Kissner, R.;
Schubiger, P. A. Organometallics 1997, 16, 1833-1840.

2522 Organometallics, Vol. 18, No. 13, 1999
Minutolo and Katzenellenbogen
A few representative examples of this transformation
are reported in Table 2. All of the aromatic boronic acids
generally gave reasonable yields (entries 1-4 and 6,
Table 2), although certain substituent effects are evi-
dent. Competition reactions confirmed that electron-
donating substituents, like the para-methoxy group in
7d , accelerate the reaction, whereas electron-withdraw-
ing substituents, like those in 7b and 7c, decrease the
reaction rate.17 A more detailed discussion of this
substituent effect is given later. Vinylic boronic acids
(7e, Table 2, and 9,18 Scheme 2) also afforded good yields
of the product complexes.
An important feature of this reaction is its tolerance
toward a wide variety of functional groups, such as free
alcoholic and phenolic hydroxy groups (Scheme 2),
carbonyls, and aryl bromides (Table 2). Moreover, no
catalyst is required by this transformation, unlike other
cross-coupling reactions involving boronic acids, which
always require a palladium catalyst.19 The one-pot
synthesis of estradiol analogue 10 (Scheme 2) starting
from the vinyl boronic 9 is a good example of the use of
this reaction to functionalize a biologically interesting
compound with a rhenium organometallic unit, in a
simple, fast, and efficient manner.
Su bstr a te Fea tu r es Affectin g Rea ctivity. As men-
tioned in the two previous sections, only heteroatom and
carbon nucleophiles of moderate strength proved to be
efficient in this three-component reaction. Other aspects
of the specific structures of the nucleophiles also affect
reactivity (Figure 3); a careful analysis of these factors
provides significant clues to the reaction mechanism.
The presence of strongly coordinating groups such as
pyridines (a, Figure 3) interferes with the coupling
reaction, probably because the metal atom is irreversibly
coordinated by the pyridine nitrogen atom and cannot
react further with the Cp precursor to form the cyclo-
pentadienyl complex. A similar effect is encountered
when even weaker coordinating groups are arranged in
a geometry that allows for chelation, as in b. In fact,
3-thiopheneboronic acid produced the desired complex
in good yield (entry 6, Table 2), whereas its regioisomer
2-thiopheneboronic acid, which can chelate with the
metal, did not produce any Cp complex.
F igu r e 3. Nucleophile structural features affecting reac-
tion yield (where relevant, nucleophilic sites have been
circled).
Sch em e 2. On e-Step Syn th esis of Estr a d iol
Der iva tive 10 fr om Bor on ic Acid 9
Another fundamental requirement is the presence of
protic groups (acidic hydroxy groups) which can gener-
ate anions in the presence of weak bases such as
triethylamine. Although the requirement for deproto-
nation is rather obvious for carboxylic acids, where
the nucleophilic center is the anionic oxygen, in the
case of organoboron reagents, where the ultimate nu-
cleophilic center is the carbon atom, it was not im-
mediately understood that deprotonation was also re-
quired. As shown in case c (Figure 3), only free boronic
acids worked in this reaction, whereas boronic esters
(which are usually reactive in Pd-catalyzed cross-
coupling reactions)19 are completely unreactive under
our conditions. These first three cases suggest that in
order for the reaction to proceed, it is essential to have
donor sites that can precoordinate reversibly with the
rhenium precursor, but that this precoordination must
not be too strong, otherwise it prevents the metal center
from subsequently becoming coordinated to the Cp
ring.
(17) A completely opposite trend is found in typical Pd-catalyzed
cross-coupling reactions involving the same class of phenylboronic
acids: Moriya, T.; Miyaura, N.; Suzuki, A. Synlett 1994, 149-151.
(18) (a) Kabalka, G. W.; Gooch, E. E.; Sastry, K. A. R. J . Nucl. Med.
1981, 22, 908-912. (b) Nakatsuka, I.; Ferreira, N. L.; Eckelman, W.
C.; Francis, B. E.; Rzeszotarski, W. J .; Gibson, R. E.; J agoda, E. M.;
Reba, R. C. J . Med. Chem. 1984, 27, 1287-1291.
(19) Miyaura, N.; Suzuki, A. Chem. Rev. 1995, 95, 2457-2483, and
references therein.

η5-Cyclopentadienyltricarbonylrhenium Complexes
Organometallics, Vol. 18, No. 13, 1999 2523
Sch em e 3. Th r ee Com p on en t Rea ction th r ou gh th e
Case d in Figure 3 also shows that the reactivity
depends on the hybridization of the carbon nucleo-
philes: aromatic and vinylic boronic acids generally give
good yields (Table 2), whereas poor yields were obtained
with aliphatic boronic acids.12 Furthermore, the last
example (e, Figure 3) shows that nucleophiles containing
hindered sp2-carbon atoms, such as ferroceneboronic
acid, do not react at all, whereas a carboxylate analogue
of ferrocene (entry 7, Table 1) shows an excellent
reactivity. These last two examples indicate that the
reaction is very sensitive to steric factors around the
active nucleophilic carbon center. Flat sp2-carbon nu-
cleophiles are preferred over bulkier sp3-counterparts,
but when one face of the sp2-hybridized nucleophile
becomes hindered, as in case e, the yield is lowered. All
the issues discussed in this initial analysis of structure/
reactivity relationships constitute a solid basis for
starting our mechanistic investigations.
Mech a n istic Con sid er a tion s: Rea ction In ter m e-
d ia tes. Understanding the mechanism of this unusual
transformation is clearly a challenge. Some guidance
is provided by related reactions, the closest precedents
being the reactions of tetrahalogenated diazocyclopen-
tadienes with pentacarbonylmanganese halides. In these
reactions, η1-metal pentacarbonyl adducts (eq 4) have
been isolated and characterized20 and were found to
transform readily to the more stable η5-metal tricarbo-
nyl complexes.
η1-sp ecies a n d th e η3-η5 Rin g Slip p a ge Sequ en ce
is started by association between two of the three
reagents, to form an “initial intermediate” (Scheme 3),
which then reacts with the third reagent to give the η1-
complex. In the literature examples shown above (eq
4),20 diazocyclopentadiene derivatives undergo insertion
reactions into polar Re-Cl and Re-Br bonds, losing
molecular nitrogen, and generating the first detectable
η1-intermediate. If the same process happens in our
reaction, then the initial step must be an association of
+
the (CH3CN)3Re(CO)3 precursor (4) with the anionic
nucleophile, forming a polar Re-Nu bond into which
diazocyclopentadiene subsequently inserts.
The proposal that the nucleophile reacts first with the
+
(CH3CN)3Re(CO)3 precursor is supported by IR and
NMR experiments. No association reaction and no
1
consumption of reagents were observed by H and 13C
NMR spectrometry or solution IR spectroscopy when
diazocyclopentadiene was mixed either with the nucleo-
phile or with the rhenium precursor alone. On the other
hand, significant spectroscopic changes associated with
both the metal and the nucleophile were observed when
the rhenium precursor was mixed with the nucleophile.
Figure 4 shows 13C NMR experiments in which a
solution of [1-13C]acetic acid in CD3CN (spectrum a; δ
172.72) was treated with triethylamine to generate the
acetate anion (spectrum b; δ 176.10). Then, 1 equiv of
[(CD3CN)3Re(CO)3]+TfO-, prepared like its protio-
analogue 4 (eq 1), was added, at which point the
carboxylate anion was transformed into four new species
(spectrum c; δ 176.78, 177.27, 177.49, 180.67). The
nature of these intermediates will be discussed below.
After addition of diazocyclopentadiene (spectra d, e, and
f), all these species, even the longer-lasting one at δ
180.67, were converted into the final η5-cyclopentadienyl
complex [13C]-6c (see Table 1), which showed a peak at
δ 168.90. The same behavior was found in an identical
experiment conducted with [1-13C]-benzoic acid (δ 167.84,
data not shown). Here again, several species appeared
after treatment of the labeled benzoate (δ 171.92) with
the rhenium precursor (δ 171.72, 171.91,22 172.60,
175.06), which, after addition of C5H4N2, were eventu-
ally all converted into the final complex [13C]-12b (see
Scheme 6), which showed a peak at δ 164.28.
Thermal ring slippage of η1-CpRe(CO)5 to η5-CpRe-
(CO)3 has recently been studied extensively (eq 5),21 and
evidence has been reported for the intermediacy of the
η3-CpRe(CO)4 complex in this process.
By analogy with these precedents, we believe that in
our case the last two species involved in the reaction
pathway before the final formation of the η5-complex
are the corresponding η1- and the η3-species (Scheme
3) in which the more labile donor solvent CH3CN
replaces two or one molecule of CO, respectively.
Contrary to the literature examples shown in eq 4 and
eq 5, we were not able to detect the η1- and the
η3-intermediates, presumably due to their greater sus-
ceptibility toward ligand (CH3CN) dissociation and ring
slippage.
A reasonable explanation for the formation of multiple
Although the last intermediates in this reaction can
be reasonably surmised from these precedents, what
happens at the beginning of this transformation still
needs to be clarified. In fact, it is likely that the reaction
intermediates from the interaction between carboxylate
+
anions and the solvated Re(CO)3 species is presented
in Scheme 4. The species involved in the equilibria
shown here have precedents in reactions involving ReI
(20) (a) Herrmann, W. A.; Huber, M. J . Organomet. Chem. 1977,
140, 55-61. (b) Day, V. W.; Stults, B. R.; Reimer, K. J .; Shaver, A. J .
Am. Chem. Soc. 1974, 96, 4008-4009.
(21) Young, K. M.; Miller, T. M.; Wrighton, M. S. J . Am. Chem. Soc.
1990, 112, 1529-1537.
(22) The chemical shift of this peak is very similar to the starting
benzoate; so, it is also possible that there is still some uncoordinated
benzoate and that, as opposed to the case with acetate, only three
coordinated species are formed.

2524 Organometallics, Vol. 18, No. 13, 1999
Minutolo and Katzenellenbogen
Sch em e 5. P r op osed Equ ilibr ia In volvin g Bor on ic
Acid s in th e In itia l Associa tion P r ocess w ith th e
Rh en iu m P r ecu r sor
Sch em e 6. Com p etition Rea ction s of
pa r a -Su bstitu ted Ben zoic Acid s 11a -d w ith th e
Rh en iu m P r ecu r sor 4 a n d Dia zocyclop en ta d ien e
F igu r e 4. 13C NMR spectra of (a) CH313COOH (0.097
mmol) in CD3CN (1.1 mL); (b) after addition Et3N (0.20
mmol); (c) after addition of [(CD3CN)3Re(CO)3]+TfO- (0.10
mmol); (d) after addition of C5H4N2 (0.50 mmol), 23 °C, 5
min; (e) sample d after 4.5 h at 23 °C; (f) sample e after 1
h at 70 °C.
Sch em e 4. Sp ecies Lik ely In volved in th e
Associa tion P r ocess of Aceta te a n d Ben zoa te
An ion s w ith th e Rh en iu m P r ecu r sor (Sp ectr u m c,
F igu r e 4)23
+
Re(CO)3 precursor are mixed together (spectrum c,
Figure 4).
So far, attempts to characterize these mixtures of
intermediates formed in acetonitrile solution by elec-
trospray ionization mass spectrometry (ESI-MS) have
failed, most likely because all of these species are
neutral and, therefore, difficult to detect by this tech-
+
nique. Only fragments derived from the fac-Re(CO)3
portion could be observed. Analysis of IR spectra of
acetonitrile solutions of 4 showed a splitting of the νas-
(CO) band (1935 cm-1) into two broad bands (1934 and
1912 cm-1) after addition of 1 equiv of tetraethylam-
monium acetate. This result is a clear indication of a
change in the coordination environment around the
metal center resulting from the Re-acetate association,
but since several species are probably simultaneously
and TcI pentacarbonyl species and potentially bidentate
ligands of general formula A2Y (A ) O, S, Se; Y ) CH,
CR, COR, CNR2, PR2, AsR2).23 As shown in Scheme 4,
up to four species (unidentate, chelate, syn- and anti-
dimers) can be formed, and, consistently, NMR shows
four different peaks when the carboxylate and the
(23) See ref 3, pp 188-190.

η5-Cyclopentadienyltricarbonylrhenium Complexes
Organometallics, Vol. 18, No. 13, 1999 2525
Ta ble 3. Rela tive Ra te Con sta n ts (kX/kH) F ou n d in
th e Com p etition Rea ction of pa r a -Su bstitu ted
Ben zoic Acid s 11a -d (Sch em e 6) a n d Ha m m ett
Con sta n ts σ (eq 6)24
c
entry nucleophilea
Xa
σb
kX/kH
log(kX/kH)
1
2
3
4
11a
11b
11c
11d
p-methoxy -0.27
1.14
0.00 (1.00)
0.057
(0.00)
-0.19
-0.51
H
p-acetyl
p-nitro
0.50
0.78
0.64
0.31
a
b
See Scheme 6. Reference 24 and eq 6. c Values equal to
1
product ratios determined by H NMR integration of the reaction
product mixture.
present at equilibrium (Scheme 4), nothing further can
be deduced from this experiment.
Similar multiple equilibria can be proposed for the
reaction with boronic acids. Interestingly, when the
+
addition of the Re(CO)3 precursor to (E)-1-nonenylbo-
ronic acid (7e, Table 2) was followed by H NMR, the
1
F igu r e 5. Hammett Fσ plot and least-squares curve-fit
equation for the competition reactions involving para-
substituted benzoates 11a -d (Scheme 6).
peaks for the vinylic hydrogens suggest the formation
of at least four different Re-boronate adducts. This
result can be explained by assuming the occurrence of
the multiple equilibria shown in Scheme 5, which are
analogous to the ones proposed for carboxylates (Scheme
4). The fact that only free boronic acids are reactive,
whereas boronic esters are completely unreactive, sug-
gests that this preassociation step is essential for the
reaction to proceed. Therefore, we believe that at least
one of the several associated intermediates shown in
Scheme 4 and Scheme 5 is the species which then reacts
with diazocyclopentadiene to give the final product.
Because boronic esters cannot form these kinds of
preassociated intermediates, they are unable to serve
as nucleophiles in the synthesis of the CpRe(CO)3
products.
Su bstitu en t Effects on Rela tive Rea ction
Ra tes: Tr a n sition Sta te Sp ecu la tion s. The experi-
ments described in the previous section suggest that
preequilibria involving coordination of the nucleophile
to the metal center are the first steps in the synthesis
of the CpRe(CO)3 products. However, thereafter, no
further intermediates could be detected until the final
η5-Cp complex is formed. The reaction between the
preassociated species (Scheme 4 and Scheme 5) and
diazocyclopentadiene is the rate-determining step in the
three-component reaction. Therefore, an analysis of
factors affecting the reaction rate can provide useful
information about the transition state involved. For this
purpose, competition reactions between different para-
substituted benzoic acids (11a -d , Scheme 6) were
conducted, and a Hammett free energy correlation
analysis (eq 6)24
electron-withdrawing substituents slow the reaction
(entries 3, 4). These results have been inserted into the
Hammett equation (eq 6), which correlates the loga-
rithm of the rate constants with the Hammett σ-con-
stants of the substituent. The reaction constant F gives
a measure of the sensitivity of the reaction rate to
changes in the electronic properties of the reagent.
A Fσ plot of the data from Table 3 is shown in Figure
5, together with a least-squares linear curve-fit and its
resulting equation. The calculated reaction constant F
has a value of -0.51 (Figure 5), which indicates a
notable sensitivity of the reaction to the substituent
effects. The negative F value means that the reaction is
accelerated by electron-donating substituents. This
result implies that nucleophilic attack by the benzoate
is likely involved in the rate-determining step (see
below).25
As was mentioned earlier, a similar competition
reaction was also conducted with different para-
substituted phenylboronic acids (7a -d , Table 2),12 and
here again electron-donating substituents such as para-
methoxy in 7d accelerated the reaction, whereas electron-
withdrawing substituents lowered the reaction rate. In
this case, it was not feasible to make a Hammett
correlation because of the simultaneous occurrence of
a side reaction, protodeboronation, whose rate is also
increased by electron-donating substituents.26 Never-
theless, a direct dependence of the reaction rate on the
(25) The statistical correlation factor for the Hammett linear plot
of our data is not excellent (r2 ) 0.901, Figure 5), but a certain deviation
from the linear correlation can be attributed to several factors: (i) the
nucleophilic species involved in the rate-determining step are not free
benzoates (for which σ values have been originally tabulated), but are
rhenium-coordinated benzoates (see Scheme 7) and therefore might
need adjustments in their σ values, which are currently not available;
(ii) solvent effects on σ values have been reported in the past (see refs
a-d reported below), especially when hydroxylic solvents such as water
(solvent in which σ values have been originally tabulated) are replaced
by non-hydroxylic solvents such as acetonitrile, used in our reaction.
These considerations can explain why there is not a more linear
Hammett correlation, although the significance of the nucleophile
strength-dependence of the reaction rate, and therefore the active
participation of the nucleophile to the rate-determining step, is not
questionable. See: (a) Ritchie, C. D.; Lewis, E. S. J . Am. Chem. Soc.
1962, 84, 591-594. (b) Taft, R. W. J . Phys. Chem. 1960, 64, 1805-
1815. (c) Amis, E. S. Solvent Effects on Reaction Rates and Mechanisms;
Academic Press: New York, 1966; pp 152-159. (d) Hine, J . Physical
Organic Chemistry; McGraw-Hill: New York, 1962; pp 88-93.
kX
log
) Fσ
(6)
( )
kH
of the effect of the electronic properties of the nucleo-
phile (substituent constants σ, Table 3) on the relative
reaction rates was carried out. The results, reported in
Table 3, clearly show that electron-donating substitu-
ents such as para-methoxy (entry 1) accelerate the
reaction with respect to unsubstituted benzoate, whereas
(24) (a) Hammett, L. P. J . Am. Chem. Soc. 1937, 59, 96-103. (b)
Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165-195, and
references therein.

2526 Organometallics, Vol. 18, No. 13, 1999
Minutolo and Katzenellenbogen
Sch em e 7. Ra te-Deter m in in g Step : Rea ction of
P r ecoor d in a ted Rh en iu m -Nu cleop h ile Sp ecies
w ith Dia zocyclop en ta d ien e to Give th e
Sch em e 8. Hyp oth etica l In ter m ed ia cy of
Meta lloca r ben oid Sp ecies 13
Sh or t-Lived η1-Com p lex In ter m ed ia te
if the rate-determining step were the formation of the
Re-carbene intermediate 13 (step 1), then we should
not expect the reaction rate to be accelerated by
increasing the strength of the nucleophilic reagent (Nu),
because step 1 does not involve any nucleophilic reac-
tion. Furthermore, in the formation of metal-carbene
species, the metal precursors react as electrophiles with
diazo compounds.28a Therefore, this step should be
accelerated by electron-withdrawing substituents on the
nucleophile. These considerations are not consistent
with what we found in the competition reaction data
reported in Table 3 and with the Hammett correlation
of Figure 5.29
Another possibility for a metallocarbenoid reaction
pathway would be if the transformation of 13 into the
η1-complex were the slow step (step 2, Scheme 8). In
this case, the carbene species 13 should be a rather
stable low-energy intermediate that we should be able
to detect. But, as shown by the NMR experiments
reported in Figure 4, no such intermediates were found
after addition of diazocyclopentadiene to the rhenium-
nucleophile mixture. Moreover, if these species were
formed, we should also see a reaction between the Re-
(CO)3+ species (4, eq 1) and diazocyclopentadiene in the
absence of the nucleophile. However, as mentioned
above, no reaction occurred between 4 and C5H4N2 in
the absence of the nucleophile under the reaction
conditions. All these considerations indicate that a
mechanism involving the occurrence of a metal-carbene
species can reasonably be ruled out.29
We believe that the mechanism that is most consis-
tent with all of our experimental observations involves
a transition state derived from a concerted formation
of the two bonds on the cyclopentadienylic C-1 carbon
with both the nucleophile and the metal center (Scheme
9).
In this mechanism, the bimolecular nucleophilic
substitution (SN2) character of the slow step, driven by
the oxygen atom of the carboxylate (a, Scheme 9) and
by the carbon substituent R′ of the boronic acid (b,
Scheme 9), can explain the remarkable sensitivity of this
reaction toward the nucleophilicity of the reagents. Such
SN2-like transition states are stabilized by extended
nucleophilic strength can also be confidently affirmed
for this class of carbon nucleophiles. Both of these
results are very useful in suggesting a possible transi-
tion structure involved in the rate-determining step.
The results described in the previous section suggest
that a rapid preassociation between the rhenium and
nucleophile species occurs before the rate-determining
step. We think that among the four possible adducts
(Schemes 4, 5), the reactive species for both carboxylates
and boronic acids are the unidentate adducts, i.e., those
formed by loss of one acetonitrile ligand (Scheme 7, n
) 1). On the basis of literature precedents (eqs 4, 5),20,21
we also hypothesized the occurrence of a short-lived η1-
complex (Scheme 3), which rapidly undergoes a ring
slippage sequence to form the final η5-complex. What
happens between (Scheme 7, rate-determining step) still
needs to be established. We have noticed that the
reaction rate increases when higher concentrations of
diazocyclopentadiene are used, although exact kinetic
experiments have not been run. Nevertheless, this rate
dependence indicates that diazocyclopentadiene is par-
ticipating in the rate-determining step.
One possibility is that the reaction proceeds through
a metal-carbene species. Such species can be formed
in different ways. However, it is very unlikely that,
under our conditions, diazocyclopentadiene directly
generates a free carbene by loss of molecular nitrogen,
because such a process generally requires rather high
temperatures or UV irradiation, and we have found
that the reaction proceeds even at room temperature
in the dark. Moreover, free carbenes are known to react
readily with acetonitrile, producing several nitrile-
containing addition products,27 none of which have been
detected under our conditions. Nevertheless, metallo-
carbenoid species of several transition metals can also
be formed upon direct interaction of diazo compounds
with the metal precursors (Scheme 8), and they are
known to be reactive intermediates for several trans-
formations.28
(28) (a) Doyle, M. P. Acc. Chem. Res. 1986, 19, 348-356. (b) Padwa,
A.; Hornbuckle, S. F. Chem. Rev. 1991, 91, 263-309. (c) Comprehensive
Organometallic Chemistry II; Abel, E. W., Stone, F. G. A., Wilkinson,
G., Eds.; Pergamon Press: Elmsford, NY, 1995; Vol. 12, Chapters 5.1-
5.5, and references therein.
(29) An alternative mechanism might involve as the rate-limiting
step the loss of a ligand (CH3CN) from the rhenium center of the
preassociated rhenium-nucleophile species, to generate an open
coordination site, prior to the association with diazocyclopentadiene.
One would expect in this case that electron-donating substituents on
the carboxylate would facilitate ligand (CH3CN) dissociation and would
accelerate the overall reaction. The subsequent reaction would need
to be rapid: carbenoid formation, nucleophilic attack, and rearrange-
ment. However, the fact that the reaction rate depends on the
diazocyclopentadiene concentration rules out the possibility that the
loss of acetonitrile from the metal-nucleophile intermediate is the rate-
determining step of the reaction.
However, there are a few points that rule against this
mechanism. In a metallocarbenoid pathway (Scheme 8),
(26) (a) Kuivila, H. G.; Reuwer, J . F.; Mangravite, J . A. J . Am. Chem.
Soc. 1964, 86, 2666-2670. (b) Kuivila, H. G.; Reuwer, J . F.; Mangrav-
ite, J . A. Can. J . Chem. 1963, 41, 3081-3090.
(27) Griller, D.; Hadel, L.; Nazran, A. S.; Platz, M. S.; Wong, P. C.;
Savino, T. G.; Scaiano, J . C. J . Am. Chem. Soc. 1984, 106, 2227-2235,
and references therein.

η5-Cyclopentadienyltricarbonylrhenium Complexes
Organometallics, Vol. 18, No. 13, 1999 2527
Sch em e 9. P r op osed Con cer ted Tr a n sition Sta tes
for th e Rea ction of Both Ca r boxyla tes (a ) a n d
Bor on ic Acid s (b)
Sch em e 10. Alter n a tive Step w ise Rea ction of th e
P r ea ssocia ted Re-Nu Sp ecies w ith C5H4N2
th r ou gh th e High -En er gy In ter m ed ia te 14
π-conjugation, as would operate in the cyclic diene
system of the Cp ring.
Although these mechanistic pathways (a and b,
Scheme 9) are almost identical for carboxylates and
boronic acids, the different product outcomes (oxygen-
substituted complexes from the former, carbon-substi-
tuted complexes from the latter) can be explained as
follows. Once precoordinated to rhenium, carboxylates
have only one free nucleophilic site, the uncoordinated
carboxylic oxygen, which eventually forms a bond with
the C-1 atom of the Cp ring of the final complex. On
the other hand, precoordinated boronic acids still have
two potentially available nucleophilic centers, a hydroxy
and a carbon substituent. In this case, because of the
great oxophilicity of boron, the carbon substituent is
much more easily released and attacks the C-1 atom of
the Cp ring, producing carbon-substituted complexes.
Boronic acids are also known to undergo transmetala-
tion processes in the presence of palladium or rhodium
catalysts in several recently reported reactions.19,30
Therefore, it is not impossible, even in our reaction, that
a boron/rhenium transmetalation step could occur before
the coordinated metal center reacts with diazocyclopen-
tadiene. However, transmetalation reactions are known
to occur readily for boronic esters,19 just as well as for
free boronic acids, and in our reaction boronic esters
proved to be completely unreactive (Figure 3, case c).
This result suggests that an initial transmetalation step
is unlikely in our case and supports the formation of
reactive intermediates in which the rhenium is coordi-
nated to the boronic acid through a deprotonated oxygen
atom (Scheme 9, eq b).
F igu r e 6. Distribution of the partial positive charge in
transition states for para-substituted benzoates and phe-
nylboronic acids (L ) acetonitrile).
trarily to what is observed experimentally. We can
therefore assume that the concerted transition state
illustrated in Scheme 9 is the most reasonable.
The partial positive charge distribution, and its
stabilization or destabilization by the aromatic sub-
stituent X in the concerted transition states involving
para-substituted benzoates and phenylboronic acids, is
shown in Figure 6. In both cases it is clear that the
electron donation by the aromatic substituent X in the
para position can effectively stabilize the transition
state and, thereby, accelerate the reaction. In the case
of carboxylates, the partial positive charge that develops
on the carboxylic carbon atom can be efficiently delo-
calized onto the aromatic ring and onto the para-
substituent X. Similarly, the partial positive charge that
develops on the aromatic ring of phenyl boronic acids
can be stabilized by the electron-donating substituent
in the para-position. In this latter case, the transition
state closely resembles the one involved in an aryl-1,2-
nucleophilic migration reaction, which, analogously, is
known to be accelerated by electron-donating aromatic
substituents.31,32
Some of the structure-reactivity relationships re-
ported earlier in Figure 3 can be explained on the basis
of the transition states proposed here. For example,
steric reasons can account for the lower reactivity of
aliphatic boronic acids compared to aromatic and vinylic
An alternative stepwise mechanism involving a high-
energy intermediate 14 (Scheme 10) can also be ruled
out, because the slow step, which is the initial electro-
philic addition of the metal center onto the cyclopenta-
dienylic C-1 carbon, does not involve the nucleophile.
In fact, the nucleophile transfer from the rhenium to
the Cp ring occurs in the fast step. Therefore, this
mechanism should not be accelerated by electron-
donating substituents present on the nucleophile, con-
(31) 1,2-Aryl migrations, as occur in nucleophilic displacements with
anchimeric assistance, have large, negative F values (-1.3 to -6.3).
Typical examples are provided in these references: (a) Kingsbury, C.
A.; Best, D. C. Bull. Chem. Soc. J pn. 1972, 45, 3440-3445. (b)
Thompson, J . A.; Cram, D. J . J . Am. Chem. Soc. 1969, 91, 1778-1789.
(c) Kim, C. J .; Brown, H. C. J . Am. Chem. Soc. 1969, 91, 4286-4287,
4287-4289, 4289-4291. (d) Lancelot, C. J .; von Rague´ Schleyer, P. J .
Am. Chem. Soc. 1969, 91, 4291-4294. (e) Lancelot, C. J .; Harper, J .
J .; von Rague´ Schleyer, P. J . Am. Chem. Soc. 1969, 91, 4294-4296.
(32) As noted in the text, we were unable to measure a F value for
the direct aryl migration as occurs in the boronic acid substitution,
because of competitive protodeboronation. In the case of carboxylates
11a -d , our F value of -0.51 is understandably smaller than the values
reported in ref 31, because the electron-donating effect in the nucleo-
phile has to operate through a carboxyl group.
(30) (a) Sakai, M.; Ueda, M.; Miyaura, N. Angew. Chem., Int. Ed.
Engl. 1998, 37, 3279-3281. (b) Takaya, Y.; Ogasawara, M.; Hayashi,
T.; Sakai, M.; Miyaura, N. J . Am. Chem. Soc. 1998, 120, 5579-5580.
(c) Sakai, M.; Hayashi, H.; Miyaura, N. Organometallics 1997, 16,
4229-4231.

2528 Organometallics, Vol. 18, No. 13, 1999
Minutolo and Katzenellenbogen
Exp er im en ta l Section
boronic acids (case d, Figure 3). In fact, a flat sp2-
hybridized carbon substituent fits very nicely between
the Cp ring and the tetracoordinated boron atom in this
“sandwich-like” transition state (Figure 6). An sp3-
hybridized carbon substituent is not flat and, therefore,
presents more serious steric hindrance, which would
increase the energy of the transition state, slowing the
reaction or making it less efficient. Even worse is the
case of ferroceneboronic acid (case e, Figure 3): although
it is an sp2-hybridized boronic acid, one face of the
carbon-nucleophilic center is so severely hindered by
the second Cp ring and the iron atom that it is
unreactive. In fact, it is rather difficult to imagine how
such a bulky reagent could even fit into the transition
state we have proposed. This mechanism also accounts
for stereospecific transformation of (E)-vinylboronic
acids to (E)-vinyl-substituted CpRe(CO)3. In fact, com-
plexes 8e (Table 2) and 10 (Scheme 2), both derived from
(E)-vinyl boronic acids, specifically maintain their trans
double-bond configuration, as shown by their vinylic
hydrogens’ coupling constants of 15.7 and 15.9 Hz,
respectively. No cis-isomer was ever observed in either
case.
Gen er a l Com m en ts. Unless otherwise specified, reagents
and solvents used in this study were purchased from com-
mercial sources (Aldrich, Strem, Eastman, Fisher, or Mallinck-
rodt) and were used without further purification. Acetonitrile
and triethylamine were distilled over calcium hydride under
nitrogen. Analytical thin-layer chromatography (TLC) was run
on 0.25 mm Merck F-254 silica gel glass plates. Visualization
was achieved using UV illumination and/or phosphomolybdic
acid (PMA). Flash chromatography was performed according
to the method reported in the literature33 with Woelm silica
gel (0.040-0.063 mm) packing. Compound 3 was prepared
as described in the literature.10c Diazocyclopentadiene was
always used as a 1.73 M pentane solution prepared accord-
ing to Reimer and Shaver.9b For safety reasons, diazocyclo-
pentadiene should always be kept in solution and never used
neat.9b 1-Nonenylboronic acid (7e) was prepared according
to a general procedure reported by Brown and co-workers.34
A literature procedure18 for the synthesis of estradiol deriva-
tive
9 was followed. Characterization data of halogen-
substituted complexes 6a and 6b were in good agreement with
previously reported data.9a,13a,35 Purification conditions and
characterization data for some acyloxy- (6c-f)11 and carbon-
substituted (8a -e, 10)12 CpRe(CO)3 complexes have already
been reported.
Melting points were determined using a Thomas-Hoover
melting point apparatus and are uncorrected. Infrared spectra
were obtained with a Mattson Galaxy Series FT-IR 3000 using
NaCl plates or KBr pellets. NMR spectra were recorded on a
U400 or U500 Varian FT-NMR spectrometer. Chemical shifts
(δ) are reported in parts per million (ppm) downfield from
internal tetramethylsilane or from proton resonances resulting
from incomplete deuteration of the NMR solvent. Low-resolu-
tion electron impact (EI) mass spectra were obtained on a
Finnigan MAT CH5 spectrometer. High-resolution EI mass
spectra were obtained on a Finnigan MAT 731 spectrometer.
The cyclic voltammogram for 6g was recorded on a Bioana-
lytical System BAS-CV 50W electrochemical analyzer with a
platinum working electrode. The measurement was performed
at ambient temperatures under a nitrogen atmosphere in a
0.1 mM acetonitrile solution of [n-Bu4N]+[PF6]- vs Ag/AgCl
as the reference electrode. The E1/2 value is reported vs Ag/
AgCl; the couple is described as reversible (r), as verified from
the ratio of the anodic to cathodic current (ia/ic), which was
within 10% of unity.
Typ ica l Exp er im en ta l P r oced u r e for th e Syn th esis of
Acyloxy-Su bstitu ted Cp Re(CO)3 Com p lexes (6c-g, 12a -
d ). (Et4N)2[ReBr3(CO)3] (3) (3 39 mg, 0.050 mmol) was dis-
solved in anhydrous CH3CN (1.5 mL) and treated with 40 mg
(0.15 mmol) of AgOTf. AgBr was removed by filtration, and
the supernatant was added to a solution containing diazocy-
clopentadiene (35 µL of a 1.73 M pentane solution, 0.060
mmol), the carboxylic acid (5c, 5e-g, 11a -d ) (0.10 mmol), and
triethylamine (0.20 mmol) in CH3CN (1 mL). The mixture was
heated at 80 °C for 45 min and then concentrated under
vacuum. In the case of 5d , the commercially available sodium
salt was used (0.10 mmol); therefore no triethylamine was
added to the reaction mixture. The crude reaction product was
purified by flash chromatography.
Con clu sion
+
The reaction of diazocyclopentadiene, a fac-Re(CO)3
precursor, and a nucleophile has proved to be an
efficient synthetic method for preparing substituted η5-
CpRe(CO)3 complexes. The reaction shows a great
tolerance toward reactive functional groups. The main
limitation of this reaction is related to the strength of
the nucleophiles used: only moderate nucleophilic spe-
cies, such as halides and carboxylates among heteroa-
tom nucleophiles and boronic acids among carbon
nucleophiles, give good yields. The reaction also proved
to be rather sensitive to steric factors. These limitations
and further structure-reactivity relationships estab-
lished experimentally are in good agreement with the
mechanism herein proposed, which involves an initial
rapid association between the nucleophilic species and
the tricarbonyl-metal precursor. This association pro-
cess produces up to four intermediary species in equi-
librium, as shown by NMR experiments, although their
exact constitution proved difficult to ascertain. Accord-
ing to our proposed mechanism, a preassociated uni-
dentate metal-nucleophile precursor then reacts, in the
rate-determining step, with diazocyclopentadiene through
a concerted SN2-like transition state to produce a short-
lived η1-complex, which rapidly undergoes a ring slip-
page through a η3-species to the final η5-complex. In
these last steps, from the Re-nucleophile intermediates
to the final complex, no further intermediates could be
detected. This general mechanism should be equally
valid for the formation of different complexes such as
halogen-, acyloxy-, and carbon-substituted ones, and it
accounts for structural features that prevent certain
nucleophiles from reacting effectively. The experimental
results and the theoretical speculations herein reported
constitute a solid basis for starting more rigorous
mechanistic studies devoted to accurately establish
kinetic parameters (reaction orders, rate constants), as
well as for designing experiments to more precisely
identify and characterize the reaction intermediates
involved in this synthetically useful transformation.
[(F e r r o c e n y lo x y )c y c lo p e n t a d i e n y l]t r i c a r b o n y l-
r h en iu m (6g). Yield: 75%, orange solid. Mp: 155 °C. Rf (silica
(33) Still, W. C.; Kahn, M.; Mitra, A. J . Org. Chem. 1978, 43, 2923-
2925.
(34) (a) Brown, H. C.; Gupta, S. K. J . Am. Chem. Soc. 1975, 97,
5249-5255. (b) For characterization data of 7e see: Matteson, D. S.;
J esthi, P. K. J . Organomet. Chem. 1976, 110, 25-37.
(35) (a) Nesmeyanov, A. N.; Kolobova, N. E.; Anisimov, K. N.;
Makarov, Yu. V. Izv. Akad. Nauk SSSR, Ser. Khim. 1969, 2, 357-
359. (b) Nesmeyanov, A. N.; Kolobova, N. E.; Makarov, Yu. V.;
Anisimov, K. N. Izv. Akad. Nauk SSSR, Ser. Khim. 1969, 2, 1992-
1996.

η5-Cyclopentadienyltricarbonylrhenium Complexes
Organometallics, Vol. 18, No. 13, 1999 2529
gel, Hex/Et2O 9:1): 0.21. IR (NaCl): cm-1 2021, 1914, 1737,
1445, 1267, 1233, 1092. 1H NMR (500 MHz, CDCl3): δ 5.63
(pseudo t, 2H, J ) 2.4 Hz), 5.19 (pseudo t, 2H, J ) 2.4 Hz),
4.84 (pseudo t, 2H, J ) 2.0 Hz), 4.51 (pseudo t, 2H, J ) 2.0
Hz), 4.27 (s, 5H). 13C NMR (126 MHz, CDCl3): δ 193.69,
168.81, 129.65, 79.37, 74.45, 72.43, 70.52, 70.11, 68.20. MS
(EI, 70 eV): m/z 564 (M, 187Re, 39), 562 (M, 185Re, 25), 358 (M
- 3CO - (C5H5)Fe - H, 187Re, 10), 356 (M - 3CO - (C5H5)Fe
- H, 185Re, 6), 213 ((C5H5)Fe(C5H4CO), 100), 185 ((C5H5)
Fe(C5H4), 39). HRMS (EI, 70 eV): Calcd for C19H13O5
C15H8NO7185Re 498.9831. Found: 498.9826. Calcd for C15H8-
NO7187Re: 500.9858. Found: 500.9846.
Typ ica l Exp er im en ta l P r oced u r e for th e Syn th esis of
Ca r bon -Su bstitu ted Cp Re(CO)3 Com p lexes (8a -f). Con -
d ition s A. Compound 3 (39 mg, 0.050 mmol) was dissolved
in anhydrous CH3CN (1.5 mL) and treated with 40 mg (0.15
mmol) of AgOTf. AgBr was removed by filtration, and the
supernatant was added to a solution containing diazocyclo-
pentadiene (35 µL of a 1.73 M pentane solution, 0.060 mmol),
the boronic acid (7a -f) (0.5 mmol, 10 eq), and triethylamine
(1.0 mmol, 20 equiv) in CH3CN (1 mL). The mixture was
heated at 80 °C for 45 min and then concentrated under
vacuum. The crude reaction product was purified by flash
chromatography. Con d ition s B. Compound 3 (39 mg, 0.050
mmol) was dissolved in anhydrous CH3CN (1.5 mL) and
treated with 40 mg (0.15 mmol) of AgOTf. AgBr was removed
by filtration, and the supernatant was added to a solution
containing diazocyclopentadiene (35 µL of a 1.73 M pentane
solution, 0.060 mmol), the boronic acid (7a -f) (0.10 mmol, 2
equiv), and triethylamine (0.20 mmol, 4 equiv) in CH3CN (1
mL). The mixture was heated at 80 °C for 14 h and then
concentrated under vacuum. The crude reaction product was
purified by flash chromatography.
Fe185Re 561.9642. Found: 561.9651. Calcd for C19H13O5Fe187
-
Re: 563.9670. Found: 563.9657. E1/2 (FeII/FeIII): +747 (r) mV.
[(4-Me t h o x y b e n zo y lo x y )c y c lo p e n t a d ie n y l]t r ic a r -
bon ylr h en iu m (12a ). Yield: 78%, colorless solid. Mp: 101
°C. Rf (silica gel, Hex/Et2O 9:1): 0.12. IR (NaCl): cm-1 2022,
1916, 1742, 1606, 1512, 1466, 1258, 1231, 1168, 1059. 1H NMR
(500 MHz, CDCl3): δ 8.00 (AA′XX′, 2H, J AX ) 9.0 Hz, J AA′/ XX′
) 2.4 Hz), 6.95 (AA′XX′, 2H, J AX ) 9.0 Hz, J AA′/ XX′ ) 2.4 Hz),
5.65 (pseudo t, 2H, J ) 2.4 Hz), 5.21 (pseudo t, 2H, J ) 2.4
Hz), 3.88 (s, 3H). 13C NMR (101 MHz, CDCl3): δ 193.55,
164.40, 163.12, 132.40, 129.71, 120.09, 114.04, 79.63, 74.70,
55.56. MS (EI, 70 eV): m/z 486 (M, 187Re, 12), 484 (M, 185Re,
7), 458 (M - CO, 187Re, 9), 456 (M - CO, 185Re, 5), 402 (M -
3CO, 187Re, 5), 400 (M - 3CO, 185Re, 3), 135 (CH3OC6H4CO,
100). HRMS (EI, 70 eV): Calcd for C16H11O6185Re 484.0085.
Found: 484.0078. Calcd for C16H11O6187Re: 486.0113. Found:
486.0102.
(3-Th ien ylcyclop en ta d ien yl)tr ica r bon ylr h en iu m (8f).36
Yield: 72% (conditions A). HRMS (EI, 70 eV): Calcd for
C
12H7O3S185Re 415.9646. Found: 415.9641. Calcd for C12H7O3-
S187Re: 417.9674. Found: 417.9672.
NMR Exp er im en ts Rep or ted in F igu r e 4. Spectrum a
(δ ) 172.72 ppm): commercially available (Aldrich) [1-13C]-
acetic acid (5.9 mg, 0.097 mmol) was dissolved in CD3CN (1.1
mL) in a NMR tube. Spectrum b (δ ) 176.10 ppm): triethy-
lamine (28 µL, 0.20 mmol) was added. Spectrum c (δ ) 176.78,
177.27, 177.49, 180.67 ppm): the solution derived from
spectrum b was added to a vial containing 3 (78 mg, 0.10
mmol) and AgOTf (84 mg, 0.33 mmol); AgBr was removed by
filtration, and the supernatant was stirred at room tempera-
ture for 45 min, then returned to the NMR tube and analyzed.
[(Ben zoyloxy)cyclop en t a d ien yl]t r ica r b on ylr h en iu m
(12b). Yield: 79%. Colorless solid. Mp: 99 °C. Rf (silica gel,
Hex/Et2O 9:1): 0.17. IR (NaCl): cm-1 2028, 1908, 1745, 1462,
1267, 1062. 1H NMR (500 MHz, CDCl3): δ 8.07-8.05 (m, 2H),
7.64 (t, 1H, J ) 7.5 Hz), 7.49 (t, 2H, J ) 7.9 Hz), 5.68 (pseudo
t, 2H, J ) 2.5 Hz), 5.23 (pseudo t, 2H, J ) 2.4 Hz). 13C NMR
(101 MHz, CDCl3): δ 193.43, 163.42, 134.29, 130.18, 129.33,
128.76, 127.91, 79.69, 74.78. MS (EI, 70 eV): m/z 456 (M,
187Re, 16), 454 (M, 185Re, 10), 428 (M - CO, 187Re, 24), 426 (M
- CO, 185Re, 14), 372 (M - 3CO, 187Re, 12), 370 (M - 3CO,
185Re, 7), 105 (C6H5CO, 100), 77 (C6H5, 46). HRMS (EI, 70
eV): Calcd for C15H9O5185Re 453.9980. Found: 453.9966. Calcd
for C15H9O5187Re: 456.0008. Found: 455.9994.
Spectrum d (δ ) 168.69, 177.19, 177.63, 180.63 ppm):
a
pentane solution (1.73 M) of diazocyclopentadiene (0.3 mL, 0.5
mmol) was added, and the mixture was stirred for 5 min at
room temperature. Spectrum e (δ ) 169.69, 180.63 ppm): the
solution derived from spectrum d was stirred at room tem-
perature for 4.5 h. Spectrum f (δ ) 169.90 ppm): the solution
derived from spectrum e was heated to 70 °C (outside the
spectrometer!) for 1 h; after cooling to room temperature, the
resulting dark suspension was analyzed by NMR.
[(4-Acet ylb en zoyloxy)cyclop en t a d ien yl]t r ica r b on yl-
r h en iu m (12c). Yield: 68%, colorless solid. Mp: 106 °C. Rf
(silica gel, Hex/Et2O 7:3): 0.15. IR (NaCl): cm-1 2023, 1916,
1
1749, 1466, 1256, 1232, 1079. H NMR (500 MHz, CDCl3): δ
8.14 (AA′XX′, 2H, J AX ) 8.6 Hz, J AA′/ XX′ ) 1.8 Hz), 8.05 (AA′XX′,
2H, J AX ) 8.7 Hz, J AA′/ XX′ ) 1.8 Hz), 5.70 (pseudo t, 2H, J )
2.4 Hz), 5.24 (pseudo t, 2H, J ) 2.4 Hz), 2.66 (s, 3H). 13C NMR
(126 MHz, CDCl3): δ 197.24, 193.30, 162.58, 141.17, 131.59,
130.45, 128.83, 128.48, 79.78, 74.81, 26.91. MS (EI, 70 eV):
m/z 498 (M, 187Re, 17), 496 (M, 185Re, 10), 470 (M - CO, 187Re,
21), 468 (M - CO, 185Re, 12), 414 (M - 3CO, 187Re, 14), 412
(M - 3CO, 185Re, 9), 147 (CH3C(O)C6H4CO, 100), 119 (CH3C-
(O)C6H4, 12). HRMS (EI, 70 eV): Calcd for C17H11O6185Re
496.0085. Found: 496.0071. Calcd for C17H11O6187Re: 498.0113.
Found: 498.0095.
Com petition Reaction P r ocedu r e for Differ en tly pa r a -
Su bstitu ted Ben zoic Acid s 11a -d . Rela tive Kin etic Con -
sta n t Va lu es (kX/kH) Deter m in a tion . In one experiment, 39
mg (0.050 mmol) of 3 was dissolved in anhydrous CH3CN (1.5
mL) and treated with 40 mg (0.15 mmol) of AgOTf. AgBr was
removed by filtration, and the supernatant was added to a
solution containing diazocyclopentadiene (35 µL of a 1.73 M
pentane solution, 0.060 mmol), 11a (15 mg, 0.10 mmol), 11b
(12 mg, 0.10 mmol), 11c (16 mg, 0.10 mmol), and triethylamine
(0.60 mmol). The mixture was heated at 80 °C for 45 min and
then concentrated under vacuum. The product ratio in the
crude reaction mixture was determined by 1H NMR integration
of independently and unambiguously assigned peaks, which
[(4-Nit r ob en zoyloxy)cyclop en t a d ien yl]t r ica r b on yl-
r h en iu m (12d ). Yield: 62%, pale yellow solid. Mp: 122 °C.
Rf (silica gel, Hex/Et2O 8:2): 0.14. IR (NaCl): cm-1 2022, 1921,
1758, 1527, 1457, 1346, 1260, 1234, 1075. 1H NMR (500 MHz,
CDCl3): δ 8.34 (AA′XX′, 2H, J AX ) 9.0 Hz, J AA′/ XX′ ) 2.1 Hz),
8.24 (AA′XX′, 2H, J AX ) 9.0 Hz, J AA′/ XX′ ) 2.1 Hz), 5.71 (pseudo
t, 2H, J ) 2.5 Hz), 5.25 (pseudo t, 2H, J ) 2.4 Hz). 13C NMR
(126 MHz, CDCl3): δ 193.13, 161.60, 151.21, 133.27, 131.33,
128.30, 123.90, 79.87, 74.86. MS (EI, 70 eV): m/z 501 (M,
187Re, 57), 499 (M, 185Re, 35), 473 (M - CO, 187Re, 100), 471
(M - CO, 185Re, 69), 445 (M - 2CO, 187Re, 16), 443 (M - 2CO,
185Re, 12), 417 (M - 3CO, 187Re, 12), 415 (M - 3CO, 185Re, 7),
387 (M - 3CO - NO, 187Re, 22), 385 (M - 3CO - NO, 185Re,
12), 371 (M - 3CO - NO2, 187Re, 8), 369 (M - 3CO - NO2,
185Re, 4),150 (O2NC6H4CO, 88). HRMS (EI, 70 eV): Calcd for
gave the following product distribution: 12a (41%, kMeO/kH
)
1.14), 12b (36%, kH/kH ) 1.00), 12c (23%, kAc/kH ) 0.64). In a
second experiment, 39 mg (0.050 mmol) of 3 were dissolved in
anhydrous CH3CN (1.5 mL) and treated with 40 mg (0.15
mmol) of AgOTf. AgBr was removed by filtration, and the
supernatant was added to a solution containing diazocyclo-
pentadiene (35 µL of a 1.73 M pentane solution, 0.060 mmol),
11c (16 mg, 0.10 mmol), 11d (17 mg, 0.10 mmol), and
(36) Purification conditions and most characterization data of 8f
have already been reported in: Minutolo, F.; Katzenellenbogen, J . A.
Angew. Chem., in press.

2530 Organometallics, Vol. 18, No. 13, 1999
Minutolo and Katzenellenbogen
triethylamine (0.40 mmol). The mixture was heated at 80 °C
for 45 min and then concentrated under vacuum. As in the
first experiment, 1H NMR integration analysis gave the
following product distribution: 12c (67%, kAc/kH ) 0.64
mentation was provided in part from the W. M. Keck
Foundation, the National Institutes of Health (PHS 1
S10 RR104444-01), and the National Science Founda-
tion (NSF CHE 96-10502). Mass spectra were obtained
on instruments supported by grants from the National
Institute of General Medical Sciences (GM 27029), the
National Institutes of Health (RR 01575), and the
National Science Foundation (PCM 8121494). We thank
Gregory S. Girolami and Thomas B. Rauchfuss for
helpful discussions.
determined in the first experiment), 12d (33%, kNO /kH ) 0.31).
2
Ack n ow led gm en t. We are grateful for support of
this research through grants from the National Insti-
tutes of Health (5R01 CA25836) and the Department
of Energy (DE FG02 86ER60401). NMR spectra were
obtained in the Varian Oxford Instrument Center for
Excellence in NMR Laboratory. Funding for this instru-
OM990128T