Convergent synthesis of isomeric pairs of metallocene-containing polyether macrocycles

Article abstract of DOI:10.1021/jo0259212

A general method for rapid construction of metallocene-based hosts is described. Treatment of polyether bridged bis(diphenylacetylenes) with a source of cyclopentadienyl cobalt at high temperature leads, via macrocyclization and capture of the intermediat

Full text of DOI:10.1021/jo0259212

Con ver gen t Syn th esis of Isom er ic P a ir s of
Meta llocen e-Con ta in in g P olyeth er
Ma cr ocycles
Gina A. Virtue, Ninani E. Coyne, and
Darren G. Hamilton*
Department of Chemistry, Mount Holyoke College,
South Hadley, Massachusetts 01075
hamilton@mtholyoke.edu
Received May 9, 2002
Abstr a ct: A general method for rapid construction of
metallocene-based hosts is described. Treatment of polyether
bridged bis(diphenylacetylenes) with a source of cyclopenta-
dienyl cobalt at high temperature leads, via macrocyclization
and capture of the intermediate cyclobutadiene, to macro-
cyclic systems tethered to an integral metallocene platform.
Equal yields of the two possible structural isomers, the
products of parallel or antiparallel cycloaddition of the
precursor, are obtained. A combination of X-ray crystal-
lography and NMR is employed to assign structures to the
individual isomers and to assess the strength of lithium ion
binding by the individual macrocycles.
F IGURE 1. Reaction considerations: schematic retrosyn-
thetic disconnection of a polyether bridged cyclobutadiene-
metal complex and possible fates upon cyclization of this
precursor.
nection across the four-membered ring yields an acyclic
bis-acetylene. Forward reaction, i.e. cycloaddition (with
attendant macrocyclization) and capture by ML, can lead
to products with two substitution patterns. Antiparallel
cycloaddition and complexation of the resulting cyclo-
butadiene will yield 1,3-substituted systems; the corre-
sponding parallel cycloaddition pathway leads to 1,2-
substituted systems.
We selected diphenylacetylene as the alkyne source for
these reactions since this material is known to undergo
efficient cyclodimerization and capture by a source of
cyclopentadienyl cobalt and provides sites on which to
locate additional binding functionality in later generation
systems.4 Alkylation of poly(ethylene glycol) ditosylates
1 or 2 with 2-bromophenol gave the general bridged
polyether derivatives 3 and 4 in reasonable yield without
recourse to chromatography (Scheme 1). Palladium-
catalyzed coupling of these dibromides with phenyl
acetylene gave bridged bis(diphenylacetylene) derivatives
5 and 6, the immediate precursors to the target organo-
metallic systems. Macrocyclization of 5 or 6, via acetylene
dimerization, was conducted in refluxing p-xylene in the
presence of 5.0 equiv of CpCoCl2.5 Both reactions afforded
two principal cobalt-containing materials, identified as
the structural isomers 7 and 8, or 9 and 10, in essentially
identical yields of around 20% of each isomer (Scheme
2).6
Numerous metallocene-receptor conjugates, systems
designed to provide an electrochemical response to the
presence of a guest, have been reported over the past 20
years.1 The structures of these systems were largely
dictated by the employment of the metallocene as an
external scaffold on which to hang separate, discrete
macrocycles. We sought to circumvent the limitation
imposed by the use of distinct, preformed, binding and
metallocene units, and against this background we have
developed a short (three-step) sequence to metallocene-
based macrocyclic receptors that represents a derivation
of the elegant work of Brisbois2 and Gleiter.3 The ap-
proach is illustrated with the preparation of some simple
members of a potentially large class of macrocycles based
on this general approach.
A simple sketch of a polyether bridged cyclobutadiene,
captured by a suitable organometallic fragment (ML:
e.g., -CoCp, -Fe(CO)3), is shown in Figure 1. Starting
from a sketch of a 1,3-substituted target, retrosynthetic
breaking of the cyclobutadiene-metal bond and discon-
* To whom correspondence should be addressed. Visiting Scholar,
Massachusetts Institute of Technology (2002).
(1) Selected references: (a) Bell, A. P.; Hall, C. D. J . Chem. Soc.,
Chem. Commun. 1980, 163. (b) Hall, C. D.; Chu, S. Y. F. J . Organomet.
Chem. 1995, 498, 221. (c) Medina, J . C.; Goodnow, T. T.; Rojas, M. T.;
Atwood, J . L.; Lynn, B. C.; Kaifer, A. E.; Gokel, G. W. J . Am. Chem.
Soc. 1992, 114, 10583. (d) Beer, P. D.; Crane, C. G.; Danks, J . P.; Gale,
P. A.; McAleer, J . F. J . Organomet. Chem. 1995, 490, 143. (e) Beer, P.
D.; Chen, Z.; Drew, M. G. B.; J ohnson, A. O. M.; Smith, D. K.; Spencer,
P. Inorg. Chim. Acta 1996, 246, 143. (f) Plenio, H.; Eldesoky, H.;
Heinze, J . Chem. Ber. 1993, 126, 2403. (g) Plenio, H.; Yang, J . J .;
Diodone, R.; Heinze, J . Inorg. Chem. 1994, 33, 4098. (h) Plenio, H.;
Aberle, C. Organometallics 1997, 16, 5950. (i) Plenio, H.; Aberle, C.
Chem. Eur. J . 2001, 7, 4438. (j) Plenio, H.; Aberle, C.; Al Shihadeh,
Y.; Lloris, J . M.; Martinez-Manez, R.; Pardo, T.; Soto, J . Chem. Eur.
J . 2001, 7, 2848.
(3) A family of metallocenophanes has been prepared that contain
cyclopentadienyl cobalt capped cyclobutadiene or cyclopentadieneone
rings. These systems are derived from preformed, acetylene-containing,
carbocyclic rings: (a) Haberhauer, G.; Gleiter, R. Tetrahedron Lett.
1998, 39, 6695. (b) Roers, R.; Rominger, F.; Gleiter, R. Tetrahedron
Lett. 1999, 40, 3141. (c) Roers, R.; Rominger, F.; Nuber, B.; Gleiter, R.
Organometallics 2000, 19, 1578. Conceptually related work has
employed similar chemistry in syntheses of endohedral metallo-
cenophanes: Scholz, G.; Gleiter, R.; Rominger, F. Angew. Chem., Int.
Ed. 2001, 40, 2477.
(4) (a) Stevens, A. M.; Richards, C. J . Organometallics 1999, 18,
1346. (b) Waybright, S. M.; Singleton, C. P.; Wachter, K.; Murphy, C.
J .; Bunz U. H. F. J . Am. Chem. Soc. 2001, 123, 1828. (c) J ohannessen,
S. C.; Brisbois, R. G.; Fischer, J . P.; Grieco, P. A.; Counterman, A. E.;
Clemmer, D. E. J . Am. Chem. Soc. 2001, 123, 3818.
(2) Issues of isomer formation and regiocontrol in reactions of this
type have been discussed and illustrated: Brisbois, R. G.; Fogel, L.
E.; Nicaise, O. J .-C.; DeWeerd, P. J . J . Org. Chem. 1997, 62, 6708.
10.1021/jo0259212 CCC: $22.00 © 2002 American Chemical Society
Published on Web 08/22/2002
6856
J . Org. Chem. 2002, 67, 6856-6859

SCHEME 1. P r ep a r a tion of P olyeth er Br id ged
P r ecu r sor s
F IGURE 2. X-ray structure of the 1,3-bridged isomer 7 (one
conformation only is shown for a small disordered portion of
the polyether chain, with hydrogen atoms omitted).
SCHEME 2. Ma cr ocycliza tion a n d Com p lex
F or m a tion : Isom er ic 1,3- a n d 1,2-Su bstitu ted
Mon om er ic Hosts
ular weight material was a plausible product of this
reaction.
Isomers 7 and 8 both possess symmetry elements that
could reduce the number of aromatic carbon environ-
ments to 10, as long as rotation of the phenyl substituents
is fast with respect to the chemical shift time scale. Ten
peaks are indeed observed in the aromatic region of the
carbon spectrum of 1,2-isomer 8. However, 1,3-isomer 7
presents 13 peaks, indicating either slow rotation of the
phenyl substituents, or an unsymmetrical disposition of
the polyether strap with respect to the central plane of
the molecule. As there is no obvious reason to expect
phenyl group rotation to be retarded in 7 compared to 8,
we conclude that the key features of the solid-state
conformation of 7 persist in chloroform solutionsthe
noncentral disposition of the polyether strap breaking the
predicted symmetry and increasing the number of carbon
Simple symmetry arguments predict the same number
of resonances for each isomer (at least as drawn in
Scheme 2). Assignment of the correct structures to these
pairs of isomeric products was thus greatly aided by
obtaining X-ray quality crystals of the slower eluting
isomer from the macrocyclization reaction involving 5.
The structure reveals a 1,3-disposition of the polyether
bridge, confirming the identity of this material as 7
(Figure 2). Our bias toward drawing “open” host struc-
tures is also laid bare: the system adopts an essentially
self-organized structure, the metallocene unit having
rotated into the cavity defined by the polyether strap
where stabilizing C-H...O contacts are established be-
tween cyclopentadienyl protons and oxygen atoms in the
polyether chain.7 By elimination, the slightly faster
eluting fraction from this first reaction had to be the
corresponding 1,2-isomer 8 as no other identical molec-
1
environments.8 High field (600 MHz) H NMR spectra
for 7 and 8 support the assertion of the highest possible
symmetry for the latter (1,2-isomer), but reduced for the
former (1,3-isomer).9 The spectra obtained for the prod-
ucts, 9 and 10, of the reaction of 6 with CpCoCl2 match
the arguments presented for their analogues 7 and 8. No
crystals could be obtained for either of these materials,
and their structural assignments are therefore made by
analogy.10 These assignments are made with a very high
level of confidence as the number of observed aromatic
carbon environments is exactly mirrored from 8 to 10
(10), and 7 to 9 (13).11
(7) Geometry: O2‚‚‚H5 ) 2.62 Å, O2‚‚‚H5-C5 ) 165°; O4‚‚‚H6 )
2.43 Å, O4‚‚‚H6-C6 ) 172°. These distances and angles place the
observed contacts in well-established ranges for interactions of this
character: Steiner, T. Chem. Commun. 1997, 727.
(8) Near coincidence of two resonances reduces the predicted 14
features to the observed 13.
(5) The concentration of bis(diphenylacetylene) precursors was
maintained at around 5-10 mM to attempt to bias monomer formation.
During separation small amounts of a number of other cobalt-
containing materials were noted, though the amounts were too small
for unambiguous characterization. Dimers and perhaps higher order
oligomers are likely to be represented among these. The choice of
CpCoCl2 as the cobalt source was motivated by a wish to avoid any
products deriving from CO insertion into the cyclobutadiene ring, i.e.,
cyclopentadienones, in reactions that would necessarily only afford
moderate yields of monomeric macrocyclic products.
(6) Reaction of CpCo(CO)2 with diphenylacetylene yields 50% of the
corresponding sandwich complex (Rausch, M. D., Genetti, R. A. J . Org.
Chem. 1970, 35, 3888); our total yields of 7 and 8, or 9 and 10, are
comparable.
(9) Both the aromatic and aliphatic regions of the 1H NMR spectra
are far more complex for 7 than for 8. Substantial peak overlap (even
at 600 MHz) renders the picture less clear than with the corresponding
carbon spectra, though it is possible to locate 10 distinct aromatic
resonances for 7, the number expected from the persistence in solution
of the conformation revealed in the solid-state structure of this
material. Similarly, despite a mirror plane bisecting the polyether
chain, eight distinct resonances are observed for the methylene groups
as these are diastereotopic in the favored conformation.
(10) The only metallocene-macrocycle in this series that we could
obtain as a crystalline solid was 7. Crudely, this observation suggests
that only 7 has a well-defined structure, while 8-10 possess the
conformational flexibility typical of large ring systems.
(11) The parallel extends to coincidence of two resonances (ref 8).
J . Org. Chem, Vol. 67, No. 19, 2002 6857

Addition of a 5-fold excess of LiClO4 to a solution of 7
(in d3-MeCN) did not produce measurable shifts in any
of the observed resonances. The corresponding experi-
ment for 8 revealed shifts of several spectroscopically
uncluttered resonances and confirmed, as planned, that
the cyclopentadienyl singlet, the only noncoupled reso-
nance in the spectrum, offers a useful probe of binding
events. Accordingly, NMR titrations were performed to
determine binding constants (K)12 for hosts 8-10 with
lithium perchlorate.13 The choice of lithium was relatively
arbitrary and based on the suspicion that the available
binding cavities of these first generation hosts would be
rather constrained and perhaps best suited to a cation
of small diameter.14
are thus defined as electrochemically quasireversible. The
rather weak binding of lithium ion by these macrocycles
did not suggest that meaningful data might be obtained
from attempting to measure cation-induced shifts of the
redox process.18 However, the identification of an acces-
sible redox process in these systems at least suggests the
possibility of electrochemical cation detection in strongly
binding systems based on this model.
In conclusion, the employment of poly(ethylene glycol)
linkers in macrocycles 7-10 has allowed the establish-
ment of a concise and potentially readily modifiable
synthesis of metallocene-containing hosts. Several op-
tions for modification of the binding pocket of these
prototype systems suggest the possibility of improving
binding, and imparting selectivity. Options include al-
teration of donor atom number and type in the strap, the
position of substitution of the phenyl groups bearing the
strap, and the inclusion of inward facing coordinating
groups on the remaining phenyl rings. It is likely that
the position of attachment of the strap plays a significant
role in defining the nature of the binding site, and such
options may be readily explored by simple changes in the
substitution pattern of the starting materials.
In contrast to 1,3-isomer 7,15 1,2-isomer 8 binds weakly
to lithium ion with a binding constant (K) of 20 ( 10 M-1
.
This weak binding capacity suggests the presence of a
rather constrained binding cavity; accordingly, we an-
ticipated stronger interaction with the expanded systems
9 and 10. Association constants with these hosts were 1
and 2 orders of magnitude greater for 1,2-isomer 10 (K
) 190 ( 50 M-1) and 1,3-isomer 9 (K ) 1400 ( 300 M-1),
respectively. These values suggest that, at least for ortho
strap attachment, the longer chain affords a better
lithium ion binding site and that the “bridged” 1,3-isomer
represents the better disposition of the polyether strap.16
Cyclic voltammetric analysis of macrocycles 7-10
reveals the expected one electron oxidation feature at
potentials of +620 to +680 mV.17 The observed peak
separations of 70-110 mV are greater than those exhib-
ited by fully reversible systems and were found to vary
with sweep rate; the oxidation features of these systems
Exp er im en ta l Section
Gen er a l. All starting materials were obtained from com-
mercial sources with the exception of hexa(ethylene glycol)
ditosylate, which was prepared according to a literature proce-
dure.19 NMR spectra (300 or 600 MHz) were recorded in CDCl3
or CD2Cl2 and referenced to residual solvent. Electrochemical
measurements, at 1 mM concentrations of macrocycle, were
made in acetonitrile solution containing 0.1 M nBu4PF6 as the
supporting electrolyte with Pt auxiliary and working electrodes
and an Ag/AgCl reference electrode. Twenty cycles of 0 to +1.0
V at a sweep rate of 200 mV s-1 were made and runs 5-15
averaged to afford the quoted potentials.
(12) K represents the 1:1 binding constant between lithium cation
and the respective macrocycle and is quoted to two significant figures
to take account of errors involved in the determination.
1,13-Di(2-b r om op h en yl)-1,4,7,10,13-p en t a oxa p en t a d e-
ca n e (3). A mixture of 2-bromophenol (4.35 g, 25 mmol), tetra-
(ethylene glycol) ditosylate 1 (5.00 g, 10 mmol), anhydrous K2CO3
(6.90 g, 50 mmol), and anhydrous MeCN (80 mL) was heated at
reflux in an argon atmosphere for 16 h. After cooling and
filtering the solvent was removed under vacuum. The residue
was dissolved in CH2Cl2 (100 mL) and extracted with 10% aq.
NaOH (3 × 100 mL) and water (100 mL). The organic phase
was separated, dried over anhydrous MgSO4, and evaporated
to yield an amber oil that was further dried under high vacuum
to afford 3 as an oil (1.70 g, 34%): C20H24O5Br2 requires C 47.64,
H 4.80, found C 47.80, H 4.76; HR ESI-MS (positive ion)
C20H24O5Br2Na ([M + Na]+) requires 524.9888, found 524.9880;
13C NMR (CDCl3, 67 MHz) δ 155.38, 133.48, 128.55, 122.20,
113.74, 112.48, 71.33, 70.91, 69.69, 69.20 ppm; 1H NMR (CDCl3,
250 MHz) δ 7.52 (d, J ) 8 Hz, 2H), 7.23 (m, 2H), 6.90 (d, J ) 8
Hz, 2H), 6.82 (m, 2H), 4.17 (t, J ) 5 Hz, 4H), 3.91 (t, J ) 5 Hz,
4H), 3.79 (m, 4H), 3.70 (m, 4H) ppm.
(13) All titrations were performed in d3-MeCN at 298 K. Hosts were
purified by preparative TLC and rigorously dried under high vacuum
prior to use; lithium perchlorate was recrystallized from ethanol and
dried at 160 °C for 48 h. Deuterated acetonitrile was taken from fresh
ampules, and every reasonable precaution made to minimize exposure
to the atmosphere during solution preparation and aliquot addition.
In a typical titration 0.5 mL of a 5 mM solution of host was titrated
with up to 0.7 mL of a 100-500 mM solution of alkali metal
perchlorate, the concentration of guest solution being matched to an
estimate of the binding constant obtained from a rough first titration
to ensure good coverage of the 1:1 binding isotherm. Data were
analyzed using Macintosh routines custom written within the research
group of Professor Christopher A. Hunter (available at http://www-
.shef.ac.uk/uni/projects/smc/software.html). See: Bisson, A. P.; Hunter,
C. A.; Morales, J . C.; Young, K. Chem. Eur. J . 1998, 4, 845.
(14) The hosts presented here were designed as vehicles for estab-
lishing a general synthetic method and not as specific ionophores. The
NMR binding experiments were conducted in a similar spirit, the aim
being to identify general structural characteristics of use in informing
the design of second generation systems.
1,19-Di(2-b r om op h en yl)-1,4,7,10,13,16,19-h ep t a oxa h ep -
ta d eca n e (4). This material was prepared in a manner analo-
gous to that described for 3 from 2-bromophenol (3.70 g, 21
mmol), hexa(ethylene glycol) ditosylate 2 (5.00 g, 8 mmol), and
(15) Self-organization, rather than cation interaction, is preferred
by 7; no measurable responses to the presence of cations were observed.
The NMR analysis supports the retention in solution of the structure
revealed in the solid state (Figure 2) and the energetic stabilization
provided by the formation of C-H...O contacts must, despite providing
only
a small energetic benefit, successfully compete with cation
coordination. The entropic cost of the establishment of these hydrogen
bonding contacts can be expected to be minimal; all that is required is
for the cobalt complex to rotate about its 1,3-substitution axis. It seems
that by accidental design an ideal length strap was included to prime
7 for efficient self-organization.
(16) Some experiments with potassium perchlorate were also per-
formed, but the data obtained from these NMR titrations could only
be fitted to a 2:1 host:guest binding model. Potassium ion’s propensity
to form “sandwich”-type complexes with crowns is well-known, and it
is likely that the binding processes were accurately modeled by a 2:1
curve fitting analysis (e.g., yielding K ≈ 2100 M-1 for the dimeric (10)2‚
K+ complex).
(17) The specific oxidation potentials for the individual macrocycles
are 7, +620 mV; 8, +660 mV; 9, +680 mV; 10, +645 mV.
(18) The huge excesses of lithium ion that would have been required
to achieve substantial occupancy of the macrocyclic binding sites of 8
and 10 argued against attempts to measure possible shifts in oxidation
potential in the presence of this cation. Such an experiment was
attempted for the best binding macrocycle 9, but no measurable change
in the peak oxidation potential was noted and general interpretation
of the trace was hampered by the irreversibility of oxidation in this
case.
(19) Ouchi, M.; Yoshihisa, I.; Yu, L.; Nagamune, S.; Nakamura, S.;
Wada, K.; Tadao, H. Bull. Chem. Soc. J pn. 1990, 63, 1260.
6858 J . Org. Chem., Vol. 67, No. 19, 2002

anhydrous K2CO3 (5.85 g, 42 mmol), in anhydrous MeCN (50
mL). We found that more exhaustive base extraction was
required to remove excess phenol from this preparation, but
relatively pure material could still be obtained without recourse
to chromatography. Compound 4 was obtained as an oil (1.85 g,
37%): C24H32O7Br2 requires C 46.83, H 5.00, found C 46.53, H
5.07; HR ESI-MS (positive ion) C24H32O7Br2Na ([M + Na]+)
requires 613.0412, found 613.0399; 13C NMR (CDCl3, 67 MHz)
δ 155.38, 133.45, 128.52, 122.20, 113.78, 112.48, 71.30, 70.88,
70.77 (two unresolved peaks), 69.69, 69.20 ppm; 1H NMR (CDCl3,
250 MHz) δ 7.50 (d, J ) 8 Hz, 2H), 7.23 (m, 2H), 6.90 (d, J ) 8
Hz, 2H), 6.85 (m, 2H), 4.17 (t, J ) 5 Hz, 4H), 3.90 (t, J ) 5 Hz,
4H), 3.78 (m, 4H), 3.69-3.65 (m, 12H) ppm.
1,13-Di[2-(eth yn ylp h en yl)p h en yl)]-1,4,7,10,13-p en ta oxa -
p en ta d eca n e (5). A mixture of dibromide 3 (1.00 g, 2 mmol),
bis(diphenylphosphino)palladium(II) dichloride (70 mg, 1 mmol),
triphenylphosphine (50 mg, 0.2 mmol), copper(I) iodide (20 mg,
0.1 mmol), and anhydrous triethylamine (75 mL) was heated at
reflux in an argon atmosphere for 40 h. After cooling the entire
reaction mixture was added dropwise to a stirred mixture of 10%
aqueous H2SO4 (500 mL) and CHCl3 (200 mL). The organic layer
was separated, washed with 10% aqueous H2SO4 (2 × 100 mL)
and water (2 × 100 mL), dried (MgSO4), and evaporated to afford
a dark red oily residue. Chromatography (SiO2; 1:3, EtOAc/
hexanes) gave the desired bis(diphenylacetylene) 5 (Rf ) 0.22
under given conditions) as a pale yellow oil (0.62 g, 56%):
C36H34O5 requires C 79.10, H 6.27, found C 79.22, H 6.17; HR
ESI-MS (positive ion) C36H34O5Na ([M + Na]+) requires 569.2304,
found 569.2309; 13C NMR (CDCl3, 67 MHz) δ 159.47, 133.45,
131.63, 129.78, 128.42, 128.17, 123.84, 120.94, 113.22, 112.52,
93.58, 86.11, 71.33, 70.88, 69.79, 68.96 ppm; 1H NMR (CDCl3,
250 MHz) δ 7.53-7.49 (m, 4H), 7.32-7.23 (m, 10H), 6.95-6.86
(m, 4H), 4.18 (t, J ) 5 Hz, 4H), 3.90 (t, J ) 5 Hz, 4H), 3.78 (m,
4H), 3.60 (m, 4H) ppm.
Data for 7: mp 217-219 °C; C41H39O5Co requires C 73.42, H
5.86, found C 73.62, H 5.90; HR ESI-MS (positive ion) C41H39O5-
CoNa ([M + Na]+) requires 693.2022, found 693.2024; 13C NMR
(CDCl3, 75 MHz) δ 156.07, 140.14, 138.96, 132.77, 129.83,
127.96, 127.69, 126.94, 125.32, 125.00, 124.62, 119.96, 111.32,
83.79, 73.65 (second cyclobutadienyl resonance obscured by
solvent peaks), 71.60, 70.99, 69.14, 67.67 ppm; 1H NMR (300
MHz, CDCl3) δ 7.63-7.58 (m, 3H), 7.56-7.52 (m, 3H), 7.28-
7.22 (m, 2H), 7.16-7.09 (m, 3H), 7.03-7.01 (m, 3H), 6.89 (d, J
) 7 Hz, 2H), 6.83 (d, J ) 7 Hz, 2H), 4.96 (s, 5H), 4.29-4.23 (m,
2H), 3.92-3.81 (m, 4H), 3.72-3.55 (m, 8H), 3.26-3.21 (m, 2H).
Coba lt-Con ta in in g Ma cr ocycles fr om 6. 1,2-Isom er 10
a n d 1,3-Isom er 9. These materials were prepared in a manner
analogous to that described for 7 and 8 from bis(diphenylacet-
ylene) 6 (0.31 g, 0.49 mmol), in anhydrous m-xylene (100 mL),
and CpCoCl2 (0.5 mL). Initial purification on a short column
(SiO2; DCM/MeOH, 98:2) gave two mixed orange fractions, which
were separated by preparative TLC (SiO2; DCM/MeOH, 98:2)
to yield two orange-yellow oils. The faster eluting fraction (Rf )
0.71; CHCl3/MeOH, 95:5) was the 1,2-isomer 10 (70 mg, 19%),
and the slower (Rf ) 0.68; CHCl3/MeOH, 95:5) the 1,3-isomer 9
(67 mg, 18%). Data for 10: C45H47O7Co requires C 74.37, H 6.52,
found C 74.30, H 6.52; HR ESI-MS (positive ion) C45H47O7CoNa
([M + Na]+) requires 781.2546, found 781.2546; 13C NMR
(CDCl3, 75 MHz) δ 156.19, 138.33, 130.90, 128.98, 127.59,
127.45, 126.53, 125.67, 119.84, 111.54, 83.33, 74.81 (second
cyclobutadienyl resonance obscured by solvent peaks), 70.94,
70.84 (two coincident peaks), 70.53, 69.19, 67.28 ppm; 1H NMR
(300 MHz, CDCl3) δ 7.40-7.37 (m, 4H), 7.28-7.16 (m, 10H), 6.83
(d, J ) 7 Hz, 2H), 6.74 (t, J ) 7 Hz, 2H), 4.78 (s, 5H), 3.84 (br
m, 4H), 3.70-3.63 (m, 4H), 3.61-3.53 (m, 8H), 3.46-3.41 (m,
4H), 2.94-2.89 (m, 2H), 2.83 (br m, 2H) ppm. Data for 9:
C45H47O7Co requires C 74.37, H 6.52, found C 74.31, H 6.50;
HR ESI-MS (positive ion) C45H47O7CoNa ([M + Na]+) requires
781.2546, found 781.2552; 13C NMR (CD2Cl2, 75 MHz) δ 157.31,
140.15, 134.75, 132.55, 132.42, 129.13, 128.92, 128.11, 127.14,
125.84, 125.23, 120.94, 113.20, 83.44, 76.69, 74.62, 71.25, 71.03,
70.92, 70.83, 69.26, 68.57 ppm; 1H NMR (300 MHz, CDCl3) δ
8.01 (dd, J ) 7 Hz, 2 Hz, 2H), 7.72-7.65 (m, 4H), 7.58-7.53 (m,
2H), 7.50-7.47 (m, 2H), 7.36-7.31 (m, 2H), 7.13-7.07 (m, 4H),
6.86 (d, J ) 8 Hz, 2H), 4.69 (s, 5H), 3.78 (t, J ) 5 Hz, 4H), 3.57-
3.52 (m, 8H), 3.45-3.36 (m, 8H), 2.92 (t, J ) 5 Hz, 4H) ppm.
1,19-Di[2-(eth yn ylp h en yl)p h en yl)]-1,4,7,10,13,16,19-h ep -
ta oxa h ep ta d eca n e (6). This material was prepared in
a
manner analogous to that described for 5 from dibromide 4 (1.85
g, 3.1 mmol), bis(diphenylphosphino)palladium(II) dichloride
(110 mg, 0.16 mmol), triphenylphosphine (82 mg, 0.31 mmol),
and copper(I) iodide (30 mg, 0.16 mmol), in anhydrous triethyl-
amine (75 mL). Workup and isolation as described above (SiO2;
1:1, EtOAc/hexanes) gave 6 (Rf ≈ 0.4 under given conditions) as
a pale yellow oil (0.83 g, 42%): C40H42O7 requires C 75.69, H
6.67, found C 75.45, H 6.80; HR ESI-MS (positive ion) C40H42O7-
Na ([M + Na]+) requires 657.2828, found 657.2840; 13C NMR
(CDCl3, 67 MHz) δ 159.37, 133.27, 131.45, 129.64, 128.28,
128.03, 123.73, 120.80, 113.08, 112.38, 93.41, 86.00, 71.23, 70.74
(two unresolved peaks), 70.60, 69.69, 68.89 ppm; 1H NMR
(CDCl3, 250 MHz) δ 7.51-7.45 (m, 4H), 7.32-7.25 (m, 10H),
6.92-6.87 (m, 4H), 4.20 (t, J ) 5 Hz, 4H), 3.91 (t, J ) 5 Hz,
4H), 3.80 (m, 4H), 3.62-3.57 (m, 12H) ppm.
Ack n ow led gm en t. We thank the donors of the
Petroleum Research Fund, administered by the Ameri-
can Chemical Society, for support of this research (grant
35605-GB1) and the X-ray Structural Characterization
Laboratory at the Chemistry Department, University
of Massachusetts, Amherst supported by the University
and National Science Foundation (grant CHE-9974648).
We also thank Dr. A. Chandrasekaran for the X-ray
structural work. General Electric (Faculty for the
Future Program) and the Balfour Program for under-
represented minorities in the sciences provided summer
research stipends to N.E.C. and G.A.V., respectively.
The contributions of S. T. Ghebremicael (MHC) and Dr,
L. C. Dickinson (UMass, Amherst) are gratefully noted.
D.G.H. thanks Profs. J . K. M. Sanders (Cambridge,
U.K.) and C. A. Hunter (Sheffield, U.K.) for helpful
advice and Prof. D. G. Nocera (MIT) for providing
generous access to facilities.
Coba lt-Con ta in in g Ma cr ocycles fr om 5. 1,2-Isom er 8
a n d 1,3-Isom er 7. A solution of bis(diphenylacetylene) 5 (0.29
g, 0.52 mmol) in anhydrous m-xylene (100 mL) was prepared in
a flask equipped with a reflux condenser, a Pasteur pipet
through which a slow, continuous stream of nitrogen was passed,
and a mineral oil bubbler. CpCoCl2 (0.5 mL) was added via
syringe to the stirred solution, and the mixture heated to reflux
for 40 h. After cooling the solvent was removed to leave a dark
residue, which was subjected to an initial purification on a short
column (SiO2; DCM/MeOH, 99.5:0.5), and the orange fractions
were collected. These materials were separated by preparative
TLC (SiO2; DCM/MeOH, 99.5:0.5) to yield two orange materials.
The faster eluting fraction (Rf ) 0.45) was identified as the 1,2-
isomer 8, a sticky low-melting solid (80 mg, 23%), and the slower
(Rf ) 0.40) as 1,3-isomer 7, an orange crystalline material (70
mg, 20%). Data for 8: C41H39O5Co requires C 73.42, H 5.86,
found C 73.67, H 5.89; HR ESI-MS (positive ion) C41H39O5CoNa
([M + Na]+) requires 693.2022, found 693.2017; 13C NMR
(CDCl3, 75 MHz) δ 156.25, 138.59, 132.19, 128.77, 127.68,
127.52, 125.99, 125.67, 119.85, 111.76, 83.52, 74.99 (second
cyclobutadienyl resonance obscured by solvent peaks), 71.20,
Su p p or tin g In for m a tion Ava ila ble: Details of the data
collection, structure solution, and refinement of 7. A crystal-
lographic data file in CIF format is also available. This
material is available free of charge via the Internet at
http://pubs.acs.org.
1
70.79, 69.27, 67.45 ppm; H NMR (300 MHz, CDCl3) δ 7.48 (d,
J ) 8 Hz, 2H), 7.43 (m, 4H), 7.24-7.18 (m, 8H), 6.84 (d, J ) 8
Hz, 2H), 6.78 (d, J ) 8 Hz, 2H), 4.72 (s, 5H), 3.76-3.69 (m, 4H),
3.54-3.43 (m, 8H), 3.21-3.14 (m, 2H), 3.00-2.93 (m, 2H) ppm.
J O0259212
J . Org. Chem, Vol. 67, No. 19, 2002 6859