New Zirconocene-Coupling Route to Large, Functionalized Macrocycles

Article abstract of DOI:10.1021/ja0020310

The development of a new macrocyclization reagent, Rosenthal's Cp₂Zr(pyr)(Me₃SiC≡CSiMe₃), has allowed the preparation of two large, functionalized macrocycles. The diyne 5,5′-bis(4-trimethylsilylphenyl)-2,2′-bipyridine was

Full text of DOI:10.1021/ja0020310

J. Am. Chem. Soc. 2000, 122, 10345-10352
10345
New Zirconocene-Coupling Route to Large, Functionalized
Macrocycles
Jonathan R. Nitschke, Stefan Zu1rcher, and T. Don Tilley*
Contribution from the Department of Chemistry, UniVersity of California, Berkeley,
Berkeley, California 94720-1460 and the Chemical Sciences DiVision, Lawrence
Berkeley National Laboratory, 1 Cyclotron Road, Berkeley, California 94720
ReceiVed June 7, 2000
Abstract: The development of a new macrocyclization reagent, Rosenthal’s Cp2Zr(pyr)(Me3SiCtCSiMe3),
has allowed the preparation of two large, functionalized macrocycles. The diyne 5,5′-bis(4-trimethylsilylphenyl)-
2,2′-bipyridine was trimerized to the bipyridine-containing macrocycle 5, and 1,4-bis(4-(trimethylsilyl)ethynyl-
phenyl)-2,3-diphenyl-1,4-diazabuta-1,3-diene was dimerized to macrocycle 6. Both zirconocene-containing
macrocycles were characterized by single-crystal X-ray diffraction. Macrocycle 5, which is cleanly converted
to air-stable cyclophane 7 with benzoic acid, has a cavity of van der Waals radius 6.03 Å. In general, the
Cp2Zr(pyr)(Me3SiCtCSiMe3) reagent gives higher yields of macrocycles via the coupling of diynes than does
the reagent generated by addition of n-BuLi to Cp2ZrCl2.
We recently demonstrated the utility of zirconocene-coupling
methods for the high-yield synthesis of macrocycles from readily
available diynes.1-6 This synthetic approach to macrocycles
offers a number of advantages over more traditional methods,
which typically provide relatively low yields and often require
conditions of high dilution and lengthy separations from
oligomeric byproducts.7,8 More recently, we attempted to expand
the scope of zirconocene-based macrocyclization methods by
developing synthetic modifications which lead to macrocycles
and cages of various shapes, sizes, and functionalization.2-6 The
use of zirconocene coupling in the synthesis of functionalized
macrocycles has proven particularly challenging because deriva-
tives of zirconocene are highly reactive toward a wide range of
functional groups, especially those containing heteroatoms.9,10
Macrocycles containing pyridine or bipyridine units are of
interest as ligands for transition metals and as supramolecular
building blocks. We recently reported the use of Negishi’s in
situ-generated zirconocene reagent synthesized from Cp2ZrCl2
and n-BuLi, believed to be Cp2Zr(THF)-(1-butene),11 in the
syntheses of macrocycles 1 and 2.5 These syntheses were
possible only after allowing the zirconocene reagent to form
initially at low temperature, before addition of the diyne. A
notable feature of 1 is its extremely rigid structure, with the
pyridyl rings locked into an orthogonal conformation with
respect to the overall plane of the macrocycle by steric
constraints at the triangle’s “corners”.
In this contribution, we describe modifications of the zir-
conocene-coupling procedure which allow (1) the synthesis of
larger, bipyridine-containing macrocycles, which greatly expand
the potential for host-guest chemistry and supramolecular
chemistry for macrocycles obtained by such coupling methods,
and (2) the synthesis of less rigid diimine-containing macro-
cycles. The trimeric macrocycle 5 (Scheme 1) possesses a cavity
which is significantly larger than those of 1 and 2, such that
host-guest interactions involving molecular guests should be
possible. In addition, 5 would be expected to have a highly rigid
triangular geometry, but with bipyridyl groups which are free
to rotate about the pyridyl-phenylene bonds; the phenyl-phenyl
bond in biphenyl has a barrier to rotation of only 1.4 kcal/mol.12
This rotational freedom should allow (for example) the linking
of macrocycles via coordination chemistry. In the case of
macrocycle 6, the incorporation of diimine ligands into a
macrocyclic framework is expected to create novel ligand
environments for transition metal diimine complexes, which
have recently been associated with interesting transition-metal-
based transformations including the efficient polymerization of
olefins.13,14
(1) Mao, S. S. H.; Tilley, T. D. J. Am. Chem. Soc. 1995, 117, 7031.
(2) Liu, F. Q.; Harder, G.; Tilley, T. D. J. Am. Chem. Soc. 1998, 120,
3271.
(3) Mao, S. S. H.; Tilley, T. D. J. Am. Chem. Soc. 1995, 117, 5365.
(4) Mao, S. S. H.; Liu, F. Q.; Tilley, T. D. J. Am. Chem. Soc. 1998,
120, 1193.
(5) Nitschke, J. R.; Tilley, T. D. J. Org. Chem. 1998, 63, 3673.
(6) Mao, S. S. H.; Nitschke, J. R.; Liu, F.-Q.; Clouston, L. C.; Harder,
G. H. manuscript in preparation.
(7) Dietrich, B.; Viout, P.; Lehn, J.-M. Macrocyclic Chemistry: Aspects
of Organic and Inorganic Supramolecular Chemistry; VCH: New York,
1993.
In our previous research on macrocycle syntheses via the
coupling of diynes, we exclusively employed the Negishi
zirconocene reagent, which gives good results for phenylene-
bridged diynes1 and (under specific reaction conditions) in the
syntheses of 1 and 2.5 However, we have found that this reagent
fails to produce macrocycles from many functionalized diynes.
(8) Parker, D. Macrocycle Synthesis: a Practical Approach; Oxford
University Press: Oxford, 1996.
(9) Fagan, P. J.; Nugent, W. A.; Calabrese, J. C. J. Am. Chem. Soc. 1994,
116, 1880.
(12) Almenningen, A.; Bastiansen, O.; Fernholt, L. J. Mol. Struct. 1985,
128, 59.
(13) Johnson, L. K.; Killian, C. M.; Brookhart, M. J. Am. Chem. Soc.
1995, 117, 6414.
(10) Fisher, R. A.; Buchwald, S. L. Organometallics 1990, 9, 871.
(11) Negishi, E. I.; Cederbaum, F. E.; Takahashi, T. Tetrahedron Lett.
1986, 27, 2829.
(14) Britovsek, G. J. P.; Gibson, V. C.; Kimberley, B. S.; Maddox, P.
J.; McTavish, S. J.; Solan, G. A.; White, A. J. P.; Williams, D. J. Chem.
Commun. 1998, 849.
10.1021/ja0020310 CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/29/2000

10346 J. Am. Chem. Soc., Vol. 122, No. 42, 2000
Nitschke et al.
As reported here, these problems can be overcome through the
use of a different zirconocene macrocyclization reagent, the
alkyne complex Cp2Zr(pyr)(Me3SiCtCSiMe3) that was reported
by Rosenthal.15
palladium metal before reaction of the deactivated C-Br bond
at the 5-position of the pyridyl ring17 could take place. The
desired product 5,5′-bis(4-trimethylsilylphenyl)-2,2′-bipyridine
was obtained in 93% isolated yield using the catalyst Pd-
(AsPh3)2Cl2 in DMF, which has been shown to be more stable
and more active for Stille couplings.18 The following reaction,
involving cleavage of the C-Si bonds with ICl, was ac-
complished in 82% yield. The resulting diiodide was then
converted to the desired compound 3 (79% isolated yield) via
reaction with trimethylsilylacetylene under Sonogashira condi-
tions.19
Results and Discussion
Syntheses of Diynes. The precursor diyne 5,5′-bis[4-(tri-
methylsilylethynyl)phenyl]-2,2′-bipyridine (3) was synthesized
according to the procedure outlined in Scheme 1. The broadly
useful synthon p-Me3SnC6H4SiMe3 has been described previ-
ously,16 but we obtained this compound by a simplified
procedure, directly from 1,4-dibromobenzene in 95% isolated
yield. Initial attempts to employ Stille coupling conditions in
the reaction of this reagent with 5,5′-dibromo-2,2′-bipyridine,
using Pd(PPh3)4 catalyst at 110 °C, were unsuccessful. This
failure appeared to result from decomposition of the catalyst to
The diyne 1,4-bis[4-(trimethylsilyl)ethynylphenyl]-2,3-di-
phenyl-1,4-diazabuta-1,3-diene (4) was obtained in 74% yield
by the condensation of 4-(trimethylsilyl)ethynyl-aniline20 with
benzil, using titanium tetrachloride/triethylamine as the dehy-
drating agent (eq 1).
Scheme 1

Zirconocene-Coupling Route
J. Am. Chem. Soc., Vol. 122, No. 42, 2000 10347
Syntheses of Diimine-Containing Macrocycles. Attempts
to synthesize macrocycle 5 (Scheme 1) via reaction of diyne 3
with Cp2ZrCl2/n-BuLi, using our previously described proce-
dure,5 produced only complex reaction mixtures. Heating such
mixtures to 80 °C for 12 h increases the abundance of one
product, as indicated by the development of a purple color and
the increase of a resonance in the 1H NMR spectrum at δ 5.37.
This shift is nearly identical to that exhibited by the Cp ligands
of deep-purple Cp2Zr(2,2′-bipyridine),21 which suggests that
formation of a bipyridine complex competes with generation
of the desired macrocycle. Similarly, macrocycle 6 (eq 3) was
not formed upon reaction of Cp2ZrCl2/n-BuLi with 4 in
tetrahydrofuran. Instead, this reaction produced a complex
Figure 1. CHARON diagram of 5 showing three benzene guests and
van der Waals surfaces.
1
mixture which did not contain the desired product 6 (by H
NMR spectroscopy).
Scheme 2
The zirconocene reagent generated from Cp2ZrCl2 and n-BuLi
presents several practical difficulties in the coupling of func-
tionalized diynes. We have observed variations in the yields of
alkyne-coupling reactions that seem to reflect changes in the
quality of commercial n-BuLi. The highly reactive nature of
the Negishi zirconocene reagent is reflected both in its decom-
position at room temperature and in its indiscriminate reactivity
toward multifunctional molecules such as 3 and 4. To overcome
these difficulties, a more stable and selective zirconocene reagent
was sought. Several stable Cp2Zr(II) derivatives are known, for
example Cp2Zr[Ph2P(CH2)3CHdCH2]23 and Cp2Zr(CO)2,24 but
these have not been developed into generally useful alkyne-
coupling reagents. A more attractive candidate appeared to be
Rosenthal’s recently reported Cp2Zr(pyr)(Me3SiCtCSiMe3).
This compound is stable and isolable, and was reported to
exhibit certain reactivity patterns analogous to those for the
putative Cp2Zr(THF)(1-butene).11 We speculated that the greater
stability of Cp2Zr(pyr)(Me3SiCtCSiMe3) might reflect a greater
selectivity in its reactions toward functionalized diynes, allowing
the synthesis of macrocycles containing chelating diimine units.
presence of Me3SiCtCSiMe3, followed by the in situ displace-
ment of the THF ligand by pyridine.
When equimolar amounts of Cp2Zr(pyr)(Me3SiCtCSiMe3)
and 3 are mixed in benzene, crystals of 5 are observed to
precipitate within hours. Heating to 40 °C enables nearly
quantitative formation (91% isolated yield) of the macrocycle
within 24 h. Thus, the efficient conversion to 5 may be assisted
by its low solubility in benzene. The greater stability of Cp2-
Zr(pyr)(Me3SiCtCSiMe3) seems to result in a higher selectivity
for this reagent, as the deep purple zirconocene bipyridine
complex observed with use of the Cp2ZrCl2/n-BuLi reagent does
not form (visually or by NMR spectroscopy) under the new
protocol. However, heating a toluene solution of 5 in a sealed
NMR tube to 100 °C for 2 h resulted in the development of a
purple color, and a 1H NMR resonance at δ 5.37 appeared. The
selective formation of zirconacyclopentadiene functionalities
with use of the Cp2Zr(pyr)(Me3SiCtCSiMe3) reagent is,
therefore, under kinetic control.
The reported synthesis of Cp2Zr(pyr)(Me3SiCtCSiMe3)
involves the generation and purification of Cp2Zr(THF)(Me3-
SiCtCSiMe3), followed by reaction with pyridine.15 The
intermediate THF adduct is somewhat unstable; it loses THF
under vacuum or in pentane solution to yield a product in which
C-H activation of the cyclopentadienyl rings has occurred.22
We, therefore, developed a simplified synthetic procedure for
this compound (Scheme 2) which avoids isolation of the THF
adduct and gives an improved yield (85% from Cp2ZrCl2, as
compared to the literature yield of 49%15,22). This synthesis
involves generation of the Negishi zirconocene reagent in the
From a crystal grown during the course of the reaction, the
X-ray crystal structure of 5 was determined (Figure 1). The
(15) Rosenthal, U.; Ohff, A.; Baumann, W.; Tillack, A. Z. Anorg. Allg.
Chem. 1995, 621, 77.
(16) Cook, M. A.; Eaborn, C.; Walton, D. R. M. J. Organomet. Chem.
1970, 23, 85.
(17) Mitchell, T. N. Synthesis 1992, 803.
(18) Farina, V.; Krishnan, B. J. Am. Chem. Soc. 1991, 113, 9585.
(19) Ziessel, R.; Suffert, J.; Youinou, M.-T. J. Org. Chem. 1996, 61,
6535.
(21) Bishop, L. A.; Turner, M. A.; Kool, L. B. J. Organomet. Chem.
1998, 553, 53.
(22) Rosenthal, U.; Ohff, A.; Michalik, M.; Goerls, H.; Burlakov, V.
V.; Shur, V. B. Angew. Chem., Int. Ed. Engl. 1993, 32, 1193.
(23) Yamazaki, A.; Nishihara, Y.; Nakajima, K.; Hara, R.; Takahashi,
T. Organometallics 1999, 18, 3015.
(20) Takahashi, S.; Kuroyama, Y.; Sonogashira, K.; Hagihara, N.
Synthesis 1980, 1980, 627.
(24) Demerseman, B.; Bouquet, G.; Bigorgne, M. J. Organomet. Chem.
1976, 107, C19.

10348 J. Am. Chem. Soc., Vol. 122, No. 42, 2000
Nitschke et al.
structure consists of independent macrocycles, each of which
contains three molecules of benzene as guests in the macrocycle
cavity. The largest sphere that could be accommodated within
the central cavity of the macrocycle would have a van der Waals
radius of 6.03 Å. This is the length of the shortest vector that
may be drawn from the center of the molecular cavity to the
van der Waals surface of the bipyridyl groups. The average
torsion angle between the phenyl and pyridyl rings is 37°, which
is close to the literature value for biphenyl (44°),25 and
substantially greater than the average pyridyl-pyridyl torsion
angle of 6°, which is close to the experimental value of 0° for
2,2′-bipyridine.26
Macrocycle 5 readily demetalates upon treatment with benzoic
acid in benzene solution, yielding colorless cyclophane 7, as
shown in eq 2. Attempts to form porous coordination networks
by linking molecules of 7 into a solid-state structure have so
far proven unsuccessful. Mixtures of 7 with CuPF6, Cu(O3SCF3),
AgPF6, Ag(O3SCF3), AgNO3, and Pd(BF4)2 in nitromethane,
acetonitrile, acetone, dimethylformamide, and pyridine yielded
clear solutions of the same color as M(bipy)2. Vapor or liquid
diffusion of Et2O or C6H6 into these solutions, or gels made
from these solutions with poly(ethyleneoxide),27 yield precipi-
tated materials of poor crystallinity. Ongoing efforts along these
lines are targeting derivatives of 7 that feature alternative
structures that result from zirconocene-transfer reactions.
The zirconocene reagent Cp2Zr(pyr)(Me3SiCtCSiMe3) also
allows the clean macrocyclization of diazadiene 4, but the less
rigid character of this diyne results in a dimerization, rather than
a trimerization (eq 3). When this reaction was carried out in
hexane, the yield was slightly improved to 66% (vs 63% in
benzene solution). Highly soluble side products, which were
presumed to be a mixture of oligomers and the diazadiene
chelate complex of zirconocene, remained in the hexane solution
after crystallization of the product. The broad peaks of the
macrocycle in the proton NMR spectrum at room temperature
in benzene-d6 suggest a conformational equilibrium which is
slow on the NMR time scale. A better spectrum with signifi-
cantly improved resolution was obtained in dichloromethane-
d2 at room temperature.
Figure 2. (a) ORTEP diagram of one of the two crystallographically
independent molecules of 6, viewed down the pseudo-C2 axis. Ellipsoids
drawn at 50% probability. (b) Line drawing of 6 perpendicular to the
pseudo-C2 axis.
were grown by slow diffusion of pentane into a diethyl ether
solution at -10 °C. An ORTEP view of 6 is given in Figure 2.
The asymmetric unit contains two crystallographically inde-
pendent molecules. Both are pseudo-C2-symmetric and folded
about an axis that passes roughly through the two diimine groups
(Figure 2b). This folding may be described by the angles
between the two ZrN2 planes, which are 85 and 88° for the two
molecules. It is associated with a twisting of the diimine units,
which may be described by N-C-C-N torsion angles ranging
from 89 to 103°. In this conformation, the CdN double bonds
are nearly coplanar with the exocyclic phenyl rings. Two of
the four nitrogen atoms point toward the center of the macro-
cycle and are separated by a distance of 3.6 Å, whereas the
other two nitrogen atoms point away from the macrocycle
center. The crystal of 6 that was used for data collection also
contained substantial regions of electron density, which are
undoubtedly due to incorporated solvent (about 32 molecules
of diethyl ether or pentane per unit cell). This disordered solvent
resulted in a rather poor resolution of the structure (R ) 11%).
Attempts to demetalate macrocycle 6 with benzoic acid
resulted in a complex reaction mixture which contained only a
minor amount of the desired macrocyclic product (<40%
isolated by column chromatography). A side product isolated
from the same reaction was the diazadiene 4, which indicates
that benzoic acid acts to cleave the macrocycle.
Single crystals of macrocycle 6 suitable for an X-ray analysis
(26) Almenningen, A.; Bastiansen, O.; Gundersen, S.; Samdal, S. Acta
Chem. Scand. 1989, 43, 932.
(27) Yaghi, O. M.; Li, G.; Li, H. Chem. Mater. 1997, 9, 1074.
(25) Almenningen, A.; Bastiansen, O.; Fernholt, L. J. Mol. Struct. 1985,
128, 59.

Zirconocene-Coupling Route
J. Am. Chem. Soc., Vol. 122, No. 42, 2000 10349
Other Coupling Reactions with Cp2Zr(pyr)(Me3SiCt
CSiMe3). To gain more insight into the observed selectivity
exhibited by Cp2Zr(pyr)(Me3SiCtCSiMe3) for alkyne-coupling
over bipyridine coordination, its direct reaction with 2,2′-
bipyridine was examined. The reaction of 10 mg of the
zirconocene reagent with 1 equiv of 2,2′-bipyridine in benzene-
d6 required ca. 24 h at 25 °C and gave a quantitative yield of
the known complex Cp2Zr(2,2′-bipyridine).21
The zirconocene reagent Cp2Zr(pyr)(Me3SiCtCSiMe3) reacts
with 1 equiv of 4,4′-bis(trimethylsilylethynyl)biphenyl in ben-
zene to yield a near-quantitative yield of the previously reported
macrocycle 8 (eq 4). This observation suggests that the Cp2-
Zr(pyr)(Me3SiCtCSiMe3) reagent is superior to Cp2ZrCl2/n-
BuLi, which yields macrocycle 8 in 90% isolated yield. One
problem associated with the use of the latter reagent, which is
avoided with the new protocol, is separation of the poorly
soluble 8 from the Cp2ZrCl2/n-BuLi coproduct LiCl.
procedure introduced here, Cp2Zr(pyr)(Me3SiCtCSiMe3) is
readily prepared in large quantities directly from commercially
available Cp2ZrCl2. Unlike the zirconocene reagent that is
generated by adding n-BuLi to Cp2ZrCl2, Cp2Zr(pyr)(Me3SiCt
CSiMe3) is stable at room temperature, and can be stored
indefinitely under an inert atmosphere. This allows much more
precise control over the stoichiometry of reactions, especially
because Cp2Zr(THF)(1-butene) is unstable and may not form
quantitatively. The stability of Cp2Zr(pyr)(Me3SiCtCSiMe3)
also allows for coupling reactions that are slower and require
elevated temperatures to drive the reactions to completion. The
side products from coupling reactions (pyridine and Me3SiCt
CSiMe3) are soluble and volatile and, thus, much easier to
remove than the LiCl side-product associated with Cp2ZrCl2/
n-BuLi. A further advantage of this reagent derives from its
solubility properties, which allow coupling reactions in a variety
of solvents and, in particular, nonpolar solvents such as pentane.
In contrast, use of the Cp2ZrCl2/n-BuLi reagent is generally
limited to tetrahydrofuran solvent. Finally, Cp2Zr(pyr)(Me3SiCt
CSiMe3) exhibits a useful selectivity in the coupling of alkynes
in the presence of certain functional groups, such as diimine
and pyridine groups. This latter factor has proven beneficial in
providing synthetic routes to the functionalized macrocycles
reported here.
In earlier reports, we described the use of zirconocene
coupling in the high-yield syntheses of macrocycles having
various shapes and sizes.1-6 It is remarkable that in some cases,
this method allows convenient, high-yield syntheses of macro-
cycles having very large molecular dimensions. Macrocycle 5,
which results from the cyclization of a diyne containing a spacer
group of four aromatic rings, is similar in size to the macrocycle
previously formed by the cyclization of Me3SiCtC(C6H4)4Ct
CSiMe3.6,29 In extending the utility of this method, we have
sought to produce functionalized macrocycles that can act as
ligands or participate in host-guest interactions. Whereas
smaller rings containing bipyridine units have been reported
previously, 5 represents the first macrocycle with the potential
to display a wide range of binding properties due to its larger
cavity. The formation of macrocycle 6 is interesting in that it
represents a different kind of structure based on diyne dimer-
ization, which results from the less rigid structure of the spacer
group. Future investigations will address not only manipulation
of macrocycle structures available via zirconocene-coupling
routes, but the use of such derivatized macrocycles in supramo-
lecular chemistry.
Use of Cp2Zr(pyr)(Me3SiCtCSiMe3) in the coupling of 5,5′-
bis(trimethylsilylalkynyl)-2,2′-bipyridine was also investigated.
The coupling of this diyne to macrocycle 9, in 76% yield with
Cp2ZrCl2/n-BuLi, was reported previously.5 In benzene-d6, the
reaction of eq 5 proceeded to a 1:3 mixture of the desired
macrocycle 9 and what is believed to be the bipyridine complex
1
10 (by H NMR spectroscopy). However, when pentane was
used as the reaction solvent, a 92% isolated yield of macrocycle
9 was obtained. In this case, it appears that the precipitation of
9 from pentane favors formation of the product. Note that the
use of the Cp2ZrCl2/n-BuLi reagent in pentane under similar
conditions is severely limited by the negligible solubility of Cp2-
ZrCl2 in nonpolar solvents at -78 °C.28
Concluding Remarks
The diyne macrocyclization chemistry described here was
made possible by the identification of a new reagent for alkyne
coupling, Rosenthal’s Cp2Zr(pyr)(Me3SiCtCSiMe3). This re-
agent offers a number of compelling advantages over the widely
used Cp2ZrCl2/n-BuLi and should significantly expand the scope
of zirconocene-based coupling reactions. With the synthetic
Experimental Section
General. All reactions involving air-sensitive compounds were
carried out under nitrogen or argon using standard Schlenk techniques
and dry, oxygen-free solvents. Pentane, diethyl ether, benzene, and
tetrahydrofuran were distilled under nitrogen from sodium benzophen-
(28) Takahashi, T.; Kotora, M.; Fischer, R.; Nishihara, Y.; Nakajima,
K. J. Am. Chem. Soc. 1995, 117, 11039.
(29) Mao, S. S. H. Ph.D. Thesis, the University of California, Berkeley,
1996.

10350 J. Am. Chem. Soc., Vol. 122, No. 42, 2000
Nitschke et al.
one ketyl. Toluene, toluene-d8, and benzene-d6 were distilled under
nitrogen from Na/K. Methylene chloride-d2 was distilled under vacuum
from CaH2. n-BuLi was used as a 1.6 M solution in hexanes.
4-(Trimethylsilyl)ethynyl-aniline,20 5,5′-dibromobipyridine,30 and 5,5′-
bis(trimethylsilylalkynyl)-2,2′-bipyridine31 were prepared according to
literature methods.
All NMR spectra were recorded at room temperature. Chemical shifts
are referenced to the residual proton or carbon resonance of the
deuterated solvent. In some cases, DEPT32 was used to assign carbon
resonances. All mass spectra were obtained at the Mass Spectrometry
Facility of the University of California, Berkeley. Combustion analyses
were provided by Desert Analytics.
was charged with trimethylsilyl chloride (12.7 mL, 100 mmol), which
was then added dropwise at -78 °C. This caused the reaction mixture
to change from orange to colorless. On completion of the addition, the
cold bath was removed, and the reaction mixture was allowed to warm
to room temperature. Stirring at room temperature was continued for
30 min. The reaction vessel was cooled to -78 °C again, and the
addition funnel was again charged with n-BuLi (62.5 mL, 100 mmol),
which was added dropwise with stirring, which caused the reaction
mixture again to turn orange. The dry ice bath was removed after the
addition was complete, and the reaction mixture was allowed to warm
with stirring to 5 °C. The reaction mixture was then cooled to -78 °C
again, and the addition funnel was charged with trimethylstannyl
chloride (100 mL, 100 mmol, as supplied by Aldrich in THF solution),
which was then added dropwise, which caused the orange suspension
to bleach to a white suspension. The cold bath was then removed, and
the reaction mixture was allowed to warm to room temperature and
stirring was continued for 1 h. The clear, colorless reaction mixture,
containing a copious amount of precipitated LiCl, was then washed
with 200 mL of H2O and dried over MgSO4. It was then passed through
a plug of 50 mL of silica, and solvents were removed under dynamic
An Improved Synthesis of Cp2Zr(pyr)(Me3SiCtCSiMe3). Cp2-
ZrCl2 (5.00 g, 17.10 mmol) and Me3SiCtCSiMe3 (2.91 g, 17.1 mmol)
were loaded into a 400-mL Schlenk flask in the drybox and dissolved
in dry THF (100 mL) on the Schlenk line. The clear, colorless solution
was chilled to -78 °C in a dry ice/acetone bath, and n-BuLi (21.4 mL,
34.2 mmol) was added dropwise via syringe, causing a slow color
change to light yellow. The mixture was stirred for 10 min at -78 °C,
then the cold bath was removed and the reaction mixture allowed to
warm to room temperature over 2 h, over which time it turned deep
crimson. Pyridine (1.38 mL, 17.1 mmol) was then added via syringe,
which caused the reaction mixture to immediately turn opaque purple.
The solvent was then removed under dynamic vacuum until only 20
mL remained, at which point the reaction mixture was triturated with
dry pentane (250 mL). The pentane solution was stirred for 30 min,
then cannula-filtered into an Ar-purged 500-mL Schlenk flask. The
volume of this solution was reduced by half under dynamic vacuum,
and the flask was placed in a -80 °C freezer for 24 h, which resulted
in the formation of opaque purple crystals. The supernatant was cannula-
filtered off and the crystals dried under dynamic vacuum for 1 h. The
isolated yield was 6.844 g, 85%. The 1H and 13C spectra of this
compound in toluene-d8 match the literature values.15 1H NMR (500
MHz, benzene-d6) δ 0.32 (br s, 18 H, Si(CH3)3), 5.46 (s, 10 H,
cyclopentadienyl), 6.40 (m, 2 H, pyridine), 6.79 (m, 1 H, pyridine),
8.85 (m, 2 H, pyridine). 13C{1H} NMR (125 MHz, benzene-d6) δ 2.86
(Si(CH3)3), 106.71 (cyclopentadienyl), 123.43 (pyridine), 136.85 (py-
ridine), 154.41 (pyridine).
1-Trimethylsilyl-4-trimethylstannylbenzene. This compound has
been reported in the literature.16 The synthesis described herein requires
only one reaction vessel and proceeds in higher yield. Into a 1-L Schlenk
flask fitted with a 100 mL pressure-equalizing addition funnel was
loaded 1,4-dibromobenzene (23.6 g, 100 mmol) in the drybox. Dry
ether (300 mL) was added via cannula. The mixture was stirred until
a clear solution was obtained, at which point the solution was chilled
to -78 °C, using a dry ice/acetone bath, to give a white suspension.
The addition funnel was then charged with n-BuLi (62.5 mL, 100
mmol), which was added dropwise with stirring. On completion of the
addition, the cold bath was removed, and the light orange slurry was
allowed to warm to 5 °C with stirring. The reaction was then once
again chilled to -78 °C using the cold bath, and the addition funnel
1
vacuum. The yield of colorless crystals was 29.78 g, 95%. H NMR
(500 MHz, benzene-d6) δ 0.23 (s, 9H, Si(CH3)3), 0.25 (s, 9H, Sn(CH3)3),
7.52 (d, J ) 7.5 Hz, 2H, phenylene), 7.49 (d, J ) 7.5 Hz, 2H,
phenylene); 13C{1H} NMR (125 MHz, benzene-d6) δ -9.52 (Sn(CH3)3),
-0.75 (Si(CH3)3), 133.67, 136.03, 140.51, 143.28; GC-MS m/z 312
(M+), 297 (M+-CH3), 239 (M+-Si(CH3)3), 149 (M+-Sn(CH3)3). Anal.
Calcd for C12H22SiSn: C, 46.03; H, 7.08. Found: C, 46.43; H, 7.27.
5,5′-Bis(4-trimethylsilylphenyl)-2,2′-bipyridine. In the drybox,
1-trimethylsilyl-4-trimethylstannylbenzene (8.79 g, 28.1 mmol), 5,5′-
dibromo-2,2′-bipyridine (4.19 g, 13.4 mmol), palladium dichloride (78.4
mg, 0.668 mmol), triphenylarsine (409 mg, 1.34 mmol), and lithium
chloride (1.13 g, 26.7 mmol) were loaded into a 100-mL Schlenk flask
with a Teflon stopper. Dry DMF (50 mL) was then added to the flask
on the Schlenk line. The sealed flask was heated to 110 °C with
vigorous stirring behind a shield. Over the next several hours, white
product was observed to precipitate out of the green-brown homoge-
neous solution. The reaction was allowed to proceed for 96 h at 110
°C. The thick white slurry was treated with NaOH (1.12 g, 28 mmol)
in H2O (100 mL) to neutralize the toxic trimethylstannyl chloride
byproduct and then collected on a medium frit. The product was then
washed with H2O (200 mL) and pentane (200 mL), and dried under
dynamic vacuum. The yield of white powder was 5.60 g, 93%. 1H NMR
(500 MHz, benzene-d6) δ 0.25 (s, 18H, Si(CH3)3), 7.42 (d, J ) 8 Hz,
4H, phenylene), 7.48 (d, J ) 8 Hz, 4H, phenylene), 7.68 (dd, J ) 2.5,
8Hz, 2 H, pyridyl), 8.94 (dd, J ) 1, 8 Hz, 2H, pyridyl), 9.12 (dd, J )
1, 2.5 Hz, 2H, pyridyl); 13C{1H} NMR (125 MHz, benzene-d6) δ -0.75
(Si(CH3)3), 121.73, 127.10, 134.71, 135.57, 137.11, 138.92, 140.44,
148.47, 155.92. EI-MS m/z 452 (M+), 437 (M+-CH3). Anal. Calcd
for C28H32N2Si2: C, 74.28; H, 7.12; N, 6.19. Found: C, 73.97; H, 7.27;
N, 5.94.
5,5′-Bis(4-iodophenyl)-2,2′-bipyridine. A 500-mL, thick-walled
Schlenk flask with a Teflon stopper was loaded with 5,5′-bis(4-
trimethylsilylphenyl)-2,2′-bipyridine (1.80 g, 3.98 mmol) and purged
with N2. On the Schlenk line, dry CH2Cl2 (200 mL) was added, and
the mixture was stirred until a clear, colorless solution was obtained.
(30) Romero, F. M.; Ziessel, R. Tetrahedron Lett. 1995, 36, 6471.
(31) Romero, F. M.; Ziessel, R. Tetrahedron Lett. 1994, 35, 9203.
(32) Derome, A. E. Modern NMR Techniques for Chemistry Research,
1st ed.; Pergamon: Oxford, 1987.

Zirconocene-Coupling Route
J. Am. Chem. Soc., Vol. 122, No. 42, 2000 10351
Iodine monochloride was melted by immersing the reagent bottle in a
warm water bath (40 °C), and ICl (0.811 mL, 15.9 mmol) was added
as a liquid under N2 counterflow using a warm pipet. Reaction ensued
immediately, with copious formation of a yellow precipitate from the
burgundy solution. The stopper of the flask was closed, and the reaction
was heated behind a shield to 45 °C (Caution! CH2Cl2 boils at 40 °C
at 1 atm pressure) for 48 h. Once the reaction mixture had cooled to
room temperature, a solution of NaOH (0.80 g, 19.8 mmol) and Na2S2O3
(2.14 g, 15.9 mmol) in H2O (50 mL) was added, and the flask was
resealed. The reaction mixture was vigorously stirred for 12 h until
the yellow precipitate had completely converted to a light tan precipitate.
The organic phase, containing much precipitate, was then separated,
washed with 2 N NH4OH (50 mL), and reduced in volume to 100 mL
under dynamic vacuum. The precipitate was then collected on a fine
frit; was washed with H2O (50 mL), CH2Cl2 (30 mL), and pentane
(100 mL); and dried under dynamic vacuum. The yield was 1.82 g,
82%. 1H NMR (500 MHz, dichloromethane-d2) δ 7.45 (d, J ) 8.5 Hz,
4H, phenylene), 7.86 (d, J ) 8.5 Hz, 4H, phenylene), 8.03 (dd, J ) 3,
8 Hz, 2H, pyridyl), 8.54 (dd, J ) 1, 8 Hz, 2H, pyridyl), 8.91 (dd, J )
1, 3 Hz, 2H, pyridyl); EI-MS m/z 560 (M+), 433 (M+-I). Anal. Calcd
for C22H14I2N2: C, 47.17; H, 2.52; N, 5.00. Found: C, 47.95; 2.48; N,
4.74.
5,5′-Bis(4-(trimethylsilylethynyl)phenyl)-2,2′-bipyridine (3). In the
drybox, a 500-mL thick-walled Schlenk flask with a Teflon stopper
was loaded with 5,5′-bis(4-iodophenyl)-2,2′-bipyridine (7.19 g, 12.8
mmol), Pd(PPh3)4 (297 mg, 0.257 mmol), and CuI (98 mg, 0.51 mmol).
On the Schlenk line, dry toluene (250 mL) and diisopropylamine (50
mL) were added, and the mixture was stirred for 15 min, yielding a
pink suspension. Trimethylsilylacetylene (5.44 mL, 38.5 mmol) was
then added via syringe, and the flask was closed. The reaction mixture,
which turned yellow over several hours, was stirred vigorously for 24
h. The solvents were then removed under dynamic vacuum, and the
product (solubility in CH2Cl2 ) 1.5 g/L) was taken up in 4 L of CH2-
Cl2. This suspension (diisopropylammonium iodide has very low
solubility) was filtered through a plug of silica (500 mL), and the CH2-
Cl2 was removed under dynamic vacuum. The orange product was
added to 500 mL of toluene in a 1-L flask, which was then stirred and
brought to a boil and then allowed to cool to room temperature. The
product was isolated on a medium frit, as white flakes, and dried under
dynamic vacuum. The yield of pure product was 5.06 g, 79%. 1H NMR
(500 MHz, benzene-d6) δ 0.28 (s, 18H, Si(CH3)3), 7.06 (d, J ) 8.5
Hz, 4H, phenylene), 7.42 (dd, J ) 1, 8 Hz, 2H, pyridyl), 7.49 (d, J )
8.5 Hz, 4H, phenylene), 8.83 (d, J ) 8 Hz, 2H, pyridyl), 8.88 (d, J )
1 Hz, 2H, pyridyl); 13C{1H} NMR (125 MHz, benzene-d6) δ 0.41 (Si-
(CH3)3), 96.11 (alkynyl), 106.06 (alkynyl), 121.59, 123.72, 127.03,
127.85, 133.23, 135.43, 136.03, 138.33, 148.27. EI-MS m/z 500 (M+),
485 (M+-CH3). Anal. Calcd for C32H32N2Si2: C, 76.75; H, 6.44; N,
5.59. Found: C, 76.78; H, 6.44; N, 5.59.
heated to 40 °C for 24 h. Brown crystals of the product, some of which
were of X-ray quality, precipitated from solution over this time. The
solvent was removed by cannula filtration, and the product was dried
under dynamic vacuum. The isolated yield was 1.97 g, 91%. 1H NMR
(500 MHz, benzene-d6) δ -0.12 (s, 18 H, Si(CH3)3), 6.21 (s, 10H,
cyclopentadienyl), 6.79 (d, J ) 8 Hz, 4H, phenylene), 7.16 (d, obscured
by solvent, phenylene), 7.36 (dd, J ) 2, 8 Hz, 2H, pyridyl), 8.61 (d, J
) 8 Hz, 2H, pyridyl), 8.85 (d, J ) 2 Hz, 2H, pyridyl). Anal. Calcd for
C126H126N6Si6Zr3: C, 69.85; H, 5.86; N, 3.88. Found: C, 69.62; H,
5.85; N, 3.62.
Macrocycle 6. Diyne 4 (1.13 g, 1.80 mmol) and Cp2Zr(pyr)(Me3-
SiCtCSiMe3) (0.89 g, 1.80 mmol) were dissolved in benzene (50 mL)
and stirred at 55 °C for 24 h. The resulting orange solution was filtered
and evaporated to dryness. The residue was washed twice with hexane
(20 mL). Yield: 0.88 g (63%) of yellow microcrystals. Single crystals
suitable for an X-ray analysis of the macrocycle could be grown by
1
slow diffusion of pentane into a diethyl ether solution at -10 °C. H
NMR (400 MHz, benzene-d6): δ -0.30 (br s, 36H, Si(CH3)3), 5.7-
6.0 (br), 6.05 (br s, 20H, cyclopentadienyl), 6.5-6.7 (br), 6.8-7.1 (br),
1
7.8-8.2 (br); H NMR (500 MHz, dichloromethane-d2): δ -0.42 (s,
36H, Si(CH3)3), 5.62 (m, 4H, phenylene), 6.02 (m, 4H, phenylene),
6.29 (s, 20H, cyclopentadienyl), 6.61 (m, 8H, phenyl), 7.37 (m, 12H,
phenyl), 7.84 (m, 8H, phenyl). FAB-MS (NPOE): m/z 1547 (M+), 1327
(M+-Cp2Zr), 1107 (M+-2 Cp2Zr). Anal. Calcd for C100H112N4Si4O2-
Zr2: C, 70.79; H, 6.65; N, 3.30. Found: C, 71.04; H, 6.23; N, 3.28.
Macrocycle 7. Macrocycle 5 (1.97 g, 0.907 mmol) and benzoic acid
(1.00 g, 8.19 mmol) were loaded into a 100-mL flask in the drybox.
On the Schlenk line, dry benzene (50 mL) was added, and the reaction
was stirred at room temperature for 12 h until the yellow color of the
zirconacycle had faded. The solvent was then removed under dynamic
vacuum, and the product mixture was suspended in CH2Cl2 (300 mL).
The translucent suspension was washed with concentrated NH4OH (100
mL) and H2O (2 × 100 mL washes), dried over Na2CO3, and filtered
through a plug of silica (100 mL). The CH2Cl2 was then removed under
dynamic vacuum, and the product was recrystallized from boiling
1
hexane/toluene. The isolated yield was 1.20 g, 88%. H NMR (500
MHz, dichloromethane-d2) δ -0.02 (s, 18H, Si(CH3)3), 6.31 (s, 2H,
CdCH), 7.19 (d, J ) 8 Hz, 4H, phenylene), 7.49 (d, J ) 8 Hz, 4H,
phenylene), 7.96 (dd, J ) 2, 8 Hz, 2H, pyridyl), 8.41 (d, J ) 8 Hz,
2H, pyridyl), 8.83 (d, J ) 2 Hz, 2H, pyridyl); DEPT 13C{1H} NMR
(125 MHz, dichloromethane-d2) δ 0.43 (Si(CH3)3), 121.07 (CdCH),
126.23 (CH), 130.62 (CH), 131.78 (CH), 135.04 (C), 135.91 (CH),
136.22 (C), 142.02 (C), 147.75 (CH), 154.97 (C), 162.13 (C). FAB-
MS m/z 1508 (M+), 1473 (M+-SiMe3). Anal. Calcd for C96H102N6Si6:
C, 76.44; H, 6.82; N, 5.57. Found: C, 76.52; H, 6.74; N, 5.62.
Reaction of Cp2Zr(pyr)(Me3SiCtCSiMe3) with 4,4′-Bis(trimeth-
ylsilylalkynyl)-biphenyl. Cp2Zr(pyr)(Me3SiCtCSiMe3) (8.29 g, 17.6
mmol) and 4,4′-bis(trimethylsilylalkynyl)biphenyl (6.16 g, 17.8 mmol)
were loaded in the drybox into a 100-mL Schlenk flask with a Teflon
stopper, and benzene (90 mL) was added. The dark purple color of the
zirconocene reagent was seen to fade within minutes, passing through
a deep amber color, which faded to a light yellow-orange within 2 h at
25 °C. The reaction was stirred for 48 h at 60 °C, and the orange
supernatant was cannula-filtered off at room temperature. The isolated
1,4-Bis(4-(trimethylsilyl)ethynyl-phenyl)-2,3-diphenyl-1,4-diazabuta-
1,3-diene (4). Titanium tetrachloride (1.31 mL, 2.27 g; 12.0 mmol)
was dissolved in diethyl ether (125 mL) and the resulting solution was
added to a solution of 4-(trimethylsilyl)ethynyl-aniline (5.18 g, 27.3
mmol), benzil (2.05 g, 9.70 mmol), and triethylamine (20 mL) in diethyl
ether (300 mL). The mixture was stirred at 25 °C for 12 h and then
filtered through diatomaceous earth and evaporated to dryness under
dynamic vacuum. The crude product was recrystallized from diethyl
1
yield of orange crystals was 9.91 g, 99%. By H and 13C NMR, the
product consisted solely of pure 8. 1H NMR (500 MHz, benzene-d6) δ
-0.26 (s, 18H, Si(CH3)3), 6.13 (s, 10H, cyclopentadienyl), 6.71 (d, J
) 8.4 Hz, 2H, phenylene), 7.16 (d, obscured by solvent, phenylene).
Reaction of Cp2Zr(pyr)(Me3SiCtCSiMe3) with 5,5′-bis(trimeth-
ylsilylalkynyl)-2,2′-bipyridine in benzene-d6. Cp2Zr(pyr)(Me3SiCt
CSiMe3) (13.5 mg, 28.8 µmol) and 5,5′-bis(trimethylsilylalkynyl)-2,2′-
bipyridine (10.0 mg, 28.8 µmol) were loaded in the drybox into a
J-Young NMR tube, and benzene-d6 (0.5 mL) was added. After 24 h
at 25 °C, by 1H NMR, only resonances corresponding to the two
reported5 rotamers of macrocycle 1, pyridine, bis(trimethylsilyl)-
acetylene, and those assigned to chelate 10 were present. 1H NMR (500
MHz, benzene-d6) bis(trimethylsilyl)acetylene δ 0.15 (s, 18H); pyridine
δ 8.52 (m, 2H), 6.97 (m, 1H), 6.66 (m, 2H); macrocycle 1 δ -0.36,
-0.37 (overlapping s, 31.5H, Si(CH3)3), 5.98, 5.99, 6.00 (overlapping
s, 17.5H, cyclopentadienyl), 6.59 (dd, J ) 2, 8 Hz, 1H, pyridyl), 6.78
(dd, J ) 2, 6.5 Hz, 1H, pyridyl), 6.79 (dd, J ) 2.5, 7 Hz, 1H, pyridyl),
1
ether. The yield was 4.55 g (74%) of yellow crystals. H NMR (400
MHz, chloroform-d) δ 0.22 (s, 18H, Si(CH3)3), 1.20 (t, J ) 7.0, 6H,
CH3 of Et2O), 3.47 (q, J ) 7, 4H, CH2 of Et2O), 6.41 (d, J ) 8.6 Hz,
4H, phenylene), 7.19 (d, J ) 8.6 Hz, 4H, phenylene), 7.3-7.5 (m, 6H,
phenyl), 7.84 (d, J ) 7.1 Hz, 4H, phenyl); 13C NMR (100 MHz,
chloroform-d) δ 0.24 (Si(CH3)3), 15.4 (CH3 of Et2O), 66.01 (CH2 of
Et2O), 93.76 (alkynyl), 105.13 (alkynyl), 119.35, 119.96, 128.31,
128.82, 131.51, 132.23, 137.18, 149.32, 164.15 (CdN). EI-MS: M/z
552 (M+), 372 (M+-C11H13NSi), 276 (M+/2).
Macrocycle 5. Diyne 3 (1.50 g, 3.00 mmol) and Cp2Zr(pyr)(Me3-
SiCtCSiMe3) (1.41 g, 3.00 mmol) were loaded into a 100-mL Schlenk
flask equipped with a Teflon stopper, which was then charged with
dry benzene (90 mL). The flask was sealed, and the reaction mixture
was sonicated at room temperature for 30 min, until it had become a
deep brown homogeneous solution. The reaction mixture was then

10352 J. Am. Chem. Soc., Vol. 122, No. 42, 2000
Nitschke et al.
6.92 (dd, J ) 2.5, 8 Hz, 0.5H, pyridyl), 7.95 (d, J ) 8 Hz, 1H, pyridyl),
8.18 (d, J ) 2.5 Hz, 0.5H, pyridyl), 8.25 (d, J ) 7.5 Hz, 1H, pyridyl),
8.25 (s, 1H, pyridyl), 8.35 (s, 0.5H, pyridyl), 8.37 (d, J ) 1 Hz, 1H,
pyridyl), 8.43 (d, J ) 1.5 Hz, 1H, pyridyl), 8.55 (d, J ) 8.5 Hz, 1H,
pyridyl); putative chelate 8 δ 0.32 (s, 18H, Si(CH3)3), 5.16 (s, 10H,
cyclopentadienyl), 6.43 (dd, J ) 1, 8 Hz, 2H, pyridyl), 6.45 (dd, J )
2, 8 Hz, 2H, pyridyl), 7.75 (dd, J ) 1, 2 Hz, 2H, pyridyl).
corresponding to Cp2Zr(2,2′-bipyridine), bis(trimethylsilyl)acetylene,
and pyridine were observed after 24 h at 25 °C. The resonances assigned
to Cp2Zr(2,2′-bipyridine) match those reported in the literature.25 1H
NMR (500 MHz, benzene-d6) bis(trimethylsilyl)acetylene δ 0.15 (s,
18H); pyridine δ 8.52 (m, 2H), 6.97 (m, 1H), 6.66 (m, 2H); Cp2Zr-
(bipyridine) δ 5.37 (m, 2H, pyridyl), 5.39 (s, 10H, cyclopentadienyl),
6.30 (m, 2H, pyridyl), 6.92 (m, 2H, pyridyl), 7.41 (m, 2H, pyridyl).
An Improved Synthesis of Bipyridine-Containing Macrocycle 1.
Cp2Zr(pyr)(Me3SiCtCSiMe3) (3.14 g, 6.66 mmol) and 5,5′-bis-
(trimethylsilylalkynyl)-2,2′-bipyridine (2.32 g, 6.66 mmol) were loaded
in the drybox into a 400-mL Schlenk flask. On the Schlenk line, dry
pentane (200 mL) was added via cannula. The reaction mixture changed
color from opaque purple to opaque green over several minutes as the
reagents dissolved. The reaction mixture was stirred for 48 h, during
which time solid 1 precipitated from solution. The opaque green
supernatant was removed by cannula filtration, and the dark green solid
was washed with pentane (100 mL). The yellow, microcrystalline
product was dried under dynamic vacuum for 2 h at 25 °C. The isolated
yield was 3.49 g, 92%. NMR data match those previously reported.4
Reaction of Cp2Zr(pyr)(Me3SiCtCSiMe3) with 2,2′-Bipyridine.
Cp2Zr(pyr)(Me3SiCtCSiMe3) (15.1 mg, 32.0 µmol) and 2,2′-bipyridine
(5.0 mg, 32.0 µmol) were loaded into a J-Young NMR tube, and
benzene-d6 (0.5 mL) was added. By 1H and 13C NMR, only resonances
Acknowledgments. This work was supported by the Director,
Office of Basic Energy Sciences, Chemical Sciences Division,
of the U.S. Department of Energy under Contract No. DE-AC03-
76SF00098. S.Z. is grateful to the Swiss National Science
Foundation and the Novartis Stiftung, vormals Ciba-Geigy-
Jubila¨ums-Stiftung for financial support.
Supporting Information Available: Tables of crystal, data
collection, and refinement parameters; bond distances and
angles; and anisotropic displacement parameters for 5 and 6
(PDF). This material is available free of charge via the Internet
at http://pubs.acs.org.
JA0020310