Mechanistic investigation of cycloreversion/cycloaddition reactions between zirconocene metallacycle complexes and unsaturated organic substrates

Article abstract of DOI:10.1021/om010091o

Treatment of the diazametallacycle Cp₂Zr(N(t-Bu)C=N(SiMe₃)N(SiMe₃)) (4a) with diphenylacetylene resulted in the formation of the azametallacyclobutene Cp₂Zr(N(t-Bu)C(Ph)=C(Ph)) (6a) and Me₃SiN=C=NSiMe₃ in high yield. A kinetic study using UV-vis spectroscopy was carried out on the transformation. Saturation kinetic behavior was observed for the system, which is supportive of a mechanism that involves a reversible formal [2 + 2] retrocycloaddition of 4a to generate the transient imido species Cp₂Zr=N-t-Bu (7a) and Me₃SiN=C=NSiMe₃. Trapping 7a with diphenylacetylene in an overall [2 + 2] cycloaddition reaction affords zirconacycle 6a. The study of cycloreversion/cycloaddition reactions between diazametallacycle complexes and diphenylacetylene was extended to other zirconocene systems. Detailed kinetic studies were performed for the exchange reactions between the diazametallacycle complexes Cp₂Zr(N(2,6-Me₂Ph)C=N(SiMe₃)N (SiMe₃)) (8a) and Cp₂Zr(N(2,6-Me₂Ph) C=N(t-Bu)N(t-Bu)) (8b) with diphenylacetylene (5a) to give the corresponding azametallacyclobutene complex Cp₂Zr(N(2,6-Me₂Ph)C(Ph)=C(Ph)) (6c) and extruded carbodiimides (Me₃SiN=C=NSiMe₃ for 8a and (t-Bu)N=C=N(t-Bu) for 8b). For both systems, the reactions were found to be first order in metallacycle and zero order in alkyne. Treatment of the diazametallacycle complexes Cp₂Zr(N(2,6-ii-Pr₂Ph)C=N(Cyc)N(Cyc)) (9a) and Cp₂Zr(N(2,6-i-Pr₂Ph)C=N(i-Pr₂) N(i-Pr₂)) (9b) with alkyne 5a resulted in the formation of the six-membered zirconacylces 10a,b, respectively, upon heating at 75°C. The products 10a,b are generated from the overall insertion of alkyne 5a into the nitrogen-carbon bond of the zirconium-containing diazacyclobutane. Complex 10a has been characterized by an X-ray crystallographic study. When the azacyclobutene Cp₂Zr(N(2,6-i-Pr₂Ph)C(Ph)) (6e) was treated with CycN=C=NCyc or (i-Pr)N=C=N(i-Pr), the same six-membered zirconacycle complexes 10a,b were obtained. Kinetic analysis of the reaction of 6e and (i-Pr)N=C=N(i-Pr) to yield 10b supports an associative process wherein alkyne 5a directly inserts into the zirconium-carbon bond of 6e. The diazametallacycle complex 4a underwent a stoichiometric metathetical exchange with symmetrical carbodiimides RN=C=NR (R = p=Tol, m-Tok, i-Pr, Cyc) to generate new cyclic zirconocene complexes and Me₃SiN=C=NSiMe₃. Kinetic studies were carried out on the exchange reaction between 4a and (m-Tol)N=C=N(m-Tol) to form 4e and Me₃SiN=C=NSiMe₃. The experimental rate data obtained are consistent with a dissociative mechanism. Additionally, the saturation rate constant derived for this system from the data is the same (within experimental error) as the saturation rate constant obtained from the kinetic study of 4a and diphenylacetylene to form 6a and Me₃SiN=C=NSiMe₃. These findings provide additional support for a dissociative mechanistic pathway in the exchange reactions, since the rate constant in the formal [2 + 2] retrocycloaddition reaction to generate imidozirconocene species Cp₂Zr=N-t-Bu (7a) and Me₃SiN=C-NSiMe₃ should be the same for both reactions.

Full text of DOI:10.1021/om010091o

1792
Organometallics 2001, 20, 1792-1807
Mech a n istic In vestiga tion of Cyclor ever sion /
Cycloa d d ition Rea ction s betw een Zir con ocen e
Meta lla cycle Com p lexes a n d Un sa tu r a ted Or ga n ic
Su bstr a tes†
Rebecca L. Zuckerman and Robert G. Bergman*
Department of Chemistry and Center for New Directions in Organic Synthesis (CNDOS),
University of California, Berkeley, California 94720-1460
Received February 7, 2001
Treatment of the diazametallacycle Cp2Zr(N(t-Bu)CdN(SiMe3)N(SiMe3)) (4a ) with diphe-
nylacetylene resulted in the formation of the azametallacyclobutene Cp2Zr(N(t-Bu)C(Ph)d
C(Ph)) (6a ) and Me3SiNdCdNSiMe3 in high yield. A kinetic study using UV-vis spectroscopy
was carried out on the transformation. Saturation kinetic behavior was observed for the
system, which is supportive of a mechanism that involves a reversible formal [2 + 2]
retrocycloaddition of 4a to generate the transient imido species Cp2ZrdN-t-Bu (7a ) and Me3-
SiNdCdNSiMe3. Trapping 7a with diphenylacetylene in an overall [2 + 2] cycloaddition
reaction affords zirconacycle 6a . The study of cycloreversion/cycloaddition reactions between
diazametallacycle complexes and diphenylacetylene was extended to other zirconocene
systems. Detailed kinetic studies were performed for the exchange reactions between the
diazametallacycle complexes Cp2Zr(N(2,6-Me2Ph)CdN(SiMe3)N(SiMe3)) (8a ) and Cp2Zr-
(N(2,6-Me2Ph)CdN(t-Bu)N(t-Bu)) (8b) with diphenylacetylene (5a ) to give the corresponding
azametallacyclobutene complex Cp2Zr(N(2,6-Me2Ph)C(Ph)dC(Ph)) (6c) and extruded car-
bodiimides (Me3SiNdCdNSiMe3 for 8a and (t-Bu)NdCdN(t-Bu) for 8b). For both systems,
the reactions were found to be first order in metallacycle and zero order in alkyne. Treatment
of the diazametallacycle complexes Cp2Zr(N(2,6-i-Pr2Ph)CdN(Cyc)N(Cyc)) (9a ) and Cp2Zr-
(N(2,6-i-Pr2Ph)CdN(i-Pr2)N(i-Pr2)) (9b) with alkyne 5a resulted in the formation of the six-
membered zirconacycles 10a ,b, respectively, upon heating at 75 °C. The products 10a ,b are
generated from the overall insertion of alkyne 5a into the nitrogen-carbon bond of the
zirconium-containing diazacyclobutane. Complex 10a has been characterized by an X-ray
crystallographic study. When the azacyclobutene Cp2Zr(N(2,6-i-Pr2Ph)C(Ph)dC(Ph)) (6e) was
treated with CycNdCdNCyc or (i-Pr)NdCdN(i-Pr), the same six-membered zirconacycle
complexes 10a ,b were obtained. Kinetic analysis of the reaction of 6e and (i-Pr)NdCdN(i-
Pr) to yield 10b supports an associative process wherein alkyne 5a directly inserts into the
zirconium-carbon bond of 6e. The diazametallacycle complex 4a underwent a stoichiometric
metathetical exchange with symmetrical carbodiimides RNdCdNR (R ) p-Tol, m-Tol, i-Pr,
Cyc) to generate new cyclic zirconocene complexes and Me3SiNdCdNSiMe3. Kinetic studies
were carried out on the exchange reaction between 4a and (m-Tol)NdCdN(m-Tol) to form
4e and Me3SiNdCdNSiMe3. The experimental rate data obtained are consistent with a
dissociative mechanism. Additionally, the saturation rate constant derived for this system
from the data is the same (within experimental error) as the saturation rate constant obtained
from the kinetic study of 4a and diphenylacetylene to form 6a and Me3SiNdCdNSiMe3.
These findings provide additional support for a dissociative mechanistic pathway in the
exchange reactions, since the rate constant in the formal [2 + 2] retrocycloaddition reaction
to generate imidozirconocene species Cp2ZrdN-t-Bu (7a ) and Me3SiNdCdNSiMe3 should
be the same for both reactions.
In tr od u ction
Monomeric metal imido (MdNR) complexes are now
known throughout the transition-metal series. The
imide linkage can be either reactive or unreactive,
depending on the metal and the remaining ligand set
of the complex. Unreactive organoimido ligands have
served as ancillary groups in a number of organometallic
compounds, such as the tungsten carbene complex 1
depicted in Figure 1, discovered by Schrock and co-
F igu r e 1. Homogeneous olefin metathesis catalyst 1.
workers to be an effective homogeneous catalyst for
olefin metathesis.1 Other early-metal alkylidenes with
spectator imide ligands have been applied as catalysts
for olefin metathesis as well as polymerization reac-
tions.2,3
*To whom correspondence should be addressed. E-mail: bergman@
cchem.berkeley.edu.
† This paper is dedicated to Prof. Ronald Breslow on the occasion of
his 70th birthday.
10.1021/om010091o CCC: $20.00 © 2001 American Chemical Society
Publication on Web 04/04/2001

Reactions between Zr Metallacycles and Organic Substrates
Organometallics, Vol. 20, No. 9, 2001 1793
Sch em e 1
In addition to the large number of organometallic
structures known in which the organoimido ligand is
unreactive, there are a growing number of systems in
which the imide linkage undergoes chemical transfor-
mation. In early-transition-metal imido complexes, the
nucleophilicity of the nitrogen atom combined with the
high electrophilicity of the metal center can generate
high reactivity with small, unsaturated organic mol-
ecules. Some multiply bonded metal-nitrogen com-
plexes are known for their synthetic applications as
stoichiometric nitrene sources in heteroatom transfer
reactions.4,5 Additionally, metal imido complexes have
been proposed as the catalytic agents in many organic
transformations such as hydroamination of alkynes6-8
and allenes,9,10 hydrodenitrogenation,11,12 nitrile reduc-
tion,13-15 and the metathetical exchange between or-
ganic imines.16-21
Many of these reactions involve an initial overall [2
+ 2] cycloaddition between the metal-nitrogen double
bond and the double bond of the unsaturated organic
substrate to generate a four-membered metallacycle (eq
1). The intermediacy of a metallacycle is supported by
plexes.22 More recently, Birdwhistell and co-workers
have published results of carbodiimide metathesis reac-
tions that are catalyzed by vanadium imido complexes.23
In both cases, the researchers speculate that the car-
bodiimides are metathesized by imido exchange through
a diazametallacycle intermediate (e.g., the Weiss et al.
mechanism in Scheme 1).
Since cycloreversion/cycloaddition reactions of transi-
tion-metal imido complexes with heterocumulenes are
established or proposed fundamental steps in many
processes, we have been interested in enhancing our
mechanistic understanding of these reactions. Mono-
meric zirconocene imido complexes with the general
structure Cp2ZrdNR react with unsaturated organic
substrates (alkynes,10,24,25 allenes,10 imines,18,19 azides,19
carbodiimides,26 etc.) to afford formal [2 + 2] cycload-
dition products. Because of the thermal stability of many
of these complexes, isolation and direct examination of
their reactivity with unsaturated substrates can be
carried out. Reported in this paper is a study of
cycloreversion/cycloaddition reactions between zir-
conocene imido complexes and unsaturated organic
molecules (alkynes and carbodiimides). Data from the
kinetic analysis of several of these systems are pre-
sented.
mechanistic and kinetic studies for several systems.10,18-21
However, there remains a paucity of mechanistic evi-
dence in many systems for which metallacycles have
been proposed as critical intermediates.
One specific case in which metallacycles have been
postulated as critical intermediates is metal-catalyzed
carbodiimide metathesis reactions. For example, Weiss
and co-workers previously reported carbodiimide me-
tathesis reactions catalyzed by tungsten imido com-
(1) Schaverien, C. J .; Dewan, J . C.; Schrock, R. R. J . Am. Chem.
Soc. 1986, 108, 2771.
(2) Ivin, K. I. Olefin Metathesis; Academic Press: London, 1983.
(3) Hursthouse, M. B.; Motevalli, M.; Sullivan, A. C.; Wilkinson, G.
J . J . Chem. Soc., Chem. Commun. 1986, 1398.
(4) (a) Wang, W.; Espenson, J . H. Organometallics 1999, 18, 5170.
(b) McInnes, J . M.; Mountford, P. Chem. Commun. 1998, 1669. (c)
McInnes, J . M.; Blake, A. J .; Mountford, P. J . Chem. Soc., Dalton Trans.
1998, 3623.
(5) Lee, S. Y.; Bergman, R. G. J . Am. Chem. Soc. 1996, 118, 6396.
(6) Haak, E.; Bytschkov, I.; Doye, S. Angew. Chem., Int. Ed. Engl.
1999, 38, 3389.
(7) McGrane, P. L.; J ensen, M.; Livinghouse, T. J . Am. Chem. Soc.
1992, 114, 5460.
(8) McGrane, P. L.; Livinghouse, T. J . Am. Chem. Soc. 1993, 115,
11485.
Resu lts
Exch a n ge Rea ction s of Dia za m eta lla cycle Com -
p lexes 4a -d w ith Alk yn es. The zirconocene diaza-
metallacycle complexes utilized throughout this study
were prepared from Cp2(THF)ZrdNR (R ) t-Bu, 2,6-
Me2C6H3, 2,6-i-Pr2C6H3) and 1 equiv of a symmetrical
carbodiimide (R′NdCdNR′).26 The highly colored zir-
(9) Walsh, P. J .; Baranger, A. M.; Bergman, R. G. J . Am. Chem.
Soc. 1992, 114, 1708.
(10) Baranger, A. M.; Walsh, P. J .; Bergman, R. G. J . Am. Chem.
Soc. 1993, 115, 2753.
(11) Gray, S. D.; Smith, D. P.; Bruck, M. A.; Wigley, D. E. J . Am.
Chem. Soc. 1992, 114, 5462.
(12) Perot, G. Catal. Today 1991, 10, 447.
(13) Bakir, M.; Fanwick, P. E.; Walton, R. A. Inorg. Chem. 1988,
27, 2016.
(14) Rhodes, L. F.; Venanzi, L. M. Inorg. Chem. 1987, 26, 2692.
(15) Han, S. H.; Geoffroy, G. L. Polyhedron 1988, 7, 2331.
(16) Cantrell, G. K.; Meyer, T. Y. Organometallics 1997, 16, 5381.
(17) Cantrell, G. K.; Meyer, T. Y. J . Am. Chem. Soc. 1998, 120, 8035.
(18) Meyer, K. E.; Walsh, P. J .; Bergman, R. G. J . Am. Chem. Soc.
1994, 116, 2669.
(22) Meisel, I.; Hertel, G.; Weiss, K. J . Mol. Catal. 1986, 36, 159.
(23) Birdwhistell, K. R.; Lanza, J .; Pasos, J . J . Organomet. Chem.
1999, 584, 200.
(24) Walsh, P. J .; Hollander, F. J .; Bergman, R. G. J . Am. Chem.
Soc. 1988, 110, 8729.
(25) (a) Walsh, P. J .; Hollander, F. J .; Bergman, R. G. Organome-
tallics 1993, 12, 3705. (b) Molina, P.; Alajarin, M.; Saez, J . Synth.
Commun. 1982, 12, 573.
(19) Meyer, K. E.; Walsh, P. J .; Bergman, R. G. J . Am. Chem. Soc.
1995, 117, 974.
(20) Krska, S. W.; Zuckerman, R. L.; Bergman, R. G. J . Am. Chem.
Soc. 1998, 120, 11828.
(21) Zuckerman, R. L.; Krska, S. W.; Bergman, R. G. J . Am. Chem.
Soc. 2000, 122, 751.
(26) The details of the study for overall [2 + 2] cycloaddition of Cp2-
ZrdNR species with carbodiimides has been published indepen-
dently: Zuckerman, R. L.; Bergman, R. G. Organometallics 2000, 19,
4795.

1794 Organometallics, Vol. 20, No. 9, 2001
Zuckerman and Bergman
Ta ble 1. Selected Bon d Len gth s (Å) a n d Bon d
An gles (d eg) for Com p ou n d 4a
Bond Lengths
Zr1-N1
Zr1-Cp1
Si1-N2
Si1-C7
Si2-N3
Si2-C10
Si3-N1
N1-C2
N3-C1
2.117(3)
2.2361(3)
1.676(3)
1.779(4)
1.674(2)
1.868(4)
1.669(3)
1.429(5)
1.278(4)
Zr1-N2
Zr1-Cp2
C1-C2
Si1-C6
Si2-C9
Si2-C11
NI-C1
2.1222(2)
2.242631)
1.435(3)
1.789(4)
1.884(4)
1.877(4)
1.393(4)
1.408(4)
1.680(5)
N2-C1
C2-C4
Bond Angles
N1-Zr1-N2
N1-Zr1-Cp1
N2-Zr1-C1
N2-Zr1-Cp2
N2-Si1-C6
N2-Si1-C8
C6-Si1-C8
N3-Si2-C9
N3-Si2-C11
Zr1-N1-Si3
Zr1-N1-C2
C1-N1-C2
Zr1-N2-C1
Si2-N3-C1
NI-C -N3
64.17(9)
111.55(7)
32.39(9)
110.3(7)
105.5(2)
112.8(2)
106.8(2)
113.8(1)
110.3(1)
138.7(1)
138.6(2)
126.3(3)
93.7(2)
N1-Zr1-C1
N1-Zr1-Cp2
N2-Zr1-C101
C101-Zr1-Cp1
Zr1-N1-C1
N2-Si1-C7
C6-Si1-C7
N3-Si2-C10
Si3-N1-C1
Zr1-N2-Si1
Si1-N2-C1
N1-C1-N2
N2-C1-N3
N1-C2-C4
C3-C2-C5
31.97(9)
111.53(7)
112.78(7)
128.65(1)
159.1(2)
112.1(2)
108.6(2)
111.8(1)
126.8(2)
140.2(2)
125.1(2)
107.0(3)
124.8(3)
110.4(3)
107.9(3)
F igu r e 2. ORTEP diagram of 4a . The thermal ellipsoids
are scaled to represent the 50% probability surface.
conacycles generated from the reaction can be isolated
in high yields after crystallization. For example, treat-
ment of Cp2(THF)ZrdN-t-Bu (2a ) with 1 equiv of Me3-
SiNdCdNSiMe3 (3a ) in C6H6 at 25 °C gave clean
conversion to the diazametallacycle complex 4a (eq 2).
154.7(2)
128.1(3)
111.5(3)
N1-C2-C3
Ta ble 2. Cr ysta llogr a p h ic Da ta for Com p ou n d s 4a
a n d 10a
4a
10a
formula
fw
C
21H37N3Si2Zr
C49H58N3Zr
780.24
478.93
color
orange, plate
monoclinic
primitive
143
P21/n (No. 14)
9.3306(1)
14.3003(2)
18.5893(1)
90.00
91.839(1)
90.00
2479.10(4)
4
orange, prism
monoclinic
primitive
149
P21/n (No. 14)
11.9802(2)
20.9083(4)
19.0025(4)
90.00
98.175(1)
90.00
4711.5(1)
4
cryst syst
lattice type
temp (K)
space group
a, Å
b, Å
c, Å
R, deg
â, deg
γ, deg
V, Å3
Z
µ(Mo KR) cm-1
5.50
1.283
2.65
1.1
D
calcd, g/cm3
total no. of data
residuals: R; Rw
GOF
11 855
0.034; 0.046
1.73
22 126
0.046; 0.047
1.40
radiation
Mo KR (λ ) 0.710 69 Å),
graphite monochromated
Treatment of diazametallacycle 4a with 1 equiv of
diphenylacetylene (5a ) (Table 3) at 25 °C in C6D6
resulted in an immediate color change from reddish
orange to forest green. Azametallacyclobutene 6a was
the organometallic product, produced in 92% yield based
on an internal standard. The identity of 6a was con-
Red X-ray-quality crystals were obtained by slow evapo-
ration of a solution of 4a in hexanes (45% yield). An
ORTEP diagram of 4a is shown in Figure 2. The
metallacyclobutane ring of 4a is planar; the dihedral
angle between the plane formed by Zr, N1, and N2 and
the plane formed by C1, N1, and N2 is 177°. The Zr-
N1 and Zr-N2 bond distances are 2.117(3) and 2.1222(2)
Å, respectively, and are close to those seen in several
other structurally similar zirconocene amido com-
plexes.25,27,28 The bond lengths for N1-Cl and N2-C1
are virtually the same, 1.393(4) and 1.408(4) Å, respec-
tively (Table 1; see Table 2 for crystallographic data).
The shorter N3-Cl bond length of 1.278(4) Å indicates
greater electron density in the exocyclic C-N bond of
the four-membered zirconacycle.
1
firmed by comparison of its H NMR spectroscopic data
with literature values.24 Spectroscopic data revealed
that the heterocumulene Me3SiNdCdNSiMe3 (3a ) had
also formed during the course of the reaction in 98%
yield.
When the analogous reaction was carried out in the
presence of 2-butyne (5b) (Table 3), there was a color
(27) Cardin, D. J .; Lappert, M. F.; Raston, C. L. Chemistry of
Organo-Zirconium and -Hafnium Compounds; Ellis Horwood: New
York, 1986.
(28) Arney, D. J .; Bruck, M. A.; Huber, S. R.; Wigley, D. E. Inorg.
Chem. 1992, 31, 3749.

Reactions between Zr Metallacycles and Organic Substrates
Organometallics, Vol. 20, No. 9, 2001 1795
Ta ble 3. Exch a n ge Rea ction of Dia za m eta lla cycle
Com p lexes 4a ,b w ith Alk yn es
Sch em e 2
diaza-
metallacycle
reactant
metalla-
cyclobutene diimide
product
carbo-
product (1H NMR)
yield, %
R′
alkyne
4a
4a
4b
SiMe3
SiMe3
i-Pr
5a
5
5a
6a
6b
6
3a
3a
3c
92
81
20
change from reddish orange to blue. Similar to the
previous reaction, formation of metallacyclobutene 6b
and carbodiimide 3a was observed by 1H NMR spec-
troscopy in 81% and 97% yields, respectively, based on
an internal standard.
This alkyne/carbodiimide exchange reaction was dif-
ficult to generalize. When zirconacycle 4b was treated
with alkyne 5a in C6D6, no reaction was observed at 25
°C (Table 3). Heating at 80 °C for 66 h did result in the
complete consumption of zirconacycle complex 4b, but
only ca. 20% of metallacyclobutene 6a formed, based on
an internal standard, with unidentifiable impurities
comprising the remaining organometallic material.
When zirconacycle 4c was heated at 80 °C in the
presence of diphenylacetylene (5a ) or 2-butyne (2b), only
decomposition products were observed by 1H NMR
spectroscopy. Diazametallacycle complex 4d was found
to be more thermally robust when heated with diphe-
nylacetylene (5a ) or 2-butyne (5b) than either 4b or 4c
and did not begin to decompose until it was heated at
135 °C for 24 h. There was no observable azametalla-
cyclobutene 6a or 6b formation under the reaction
conditions employed.
Kin etic An a lysis of th e Rea ction betw een 4a a n d
Alk yn e 5a . The proposed mechanism of the exchange
reaction between complex 4a and diphenylacetylene (5a )
is outlined in Scheme 2. To verify that this reaction
occurs by a dissociative process, kinetic studies were
carried out by monitoring the formation of metallacy-
clobutene complex 6a . Applying the steady-state ap-
proximation to the concentration of intermediate 7a
predicts the rate law shown in eq 3. According to this
the reactant zirconacycle 4a in C6H6 to ensure pseudo-
first-order conditions (see Table S-11 in the Supporting
Information). The absorbance of product metallacy-
clobutene 6a was monitored (λ ) 420 nm) by UV-vis
spectroscopy at 15.4 ( 0.1 °C and plotted vs time (see
Figure S-3 in the Supporting Information for a sample
plot).
Excellent pseudo-first-order behavior was observed
under the experimental conditions employed. Plots of
the increase in [6a ] vs time gave exponential growth
consistent with the reaction being first order in [4a ].
Rates were measured using variable excess concentra-
tions of alkyne 5a and carbodiimide 3a . At lower
concentrations, the rates of reaction were accelerated
with added 5a and slowed by addition of carbodiimide
3a . At higher concentrations of 5a , saturation kinetics
were observed, as illustrated in the plot of kobs vs [3a ]/
[5a ] (Figure 3). A plot of 1/kobs versus [5a ]/[3a ] (Figure
3, inset) shows a linear dependence of the ratio of alkyne
to carbodiimide. The data are congruent with the rate
law derived in eq 3, and the values of k1 ) (3.3 ( 0.1) ×
104 s-1 and k3/k2 ) 0.4 at 15.4 ( 0.1 °C are obtained
from these data.
rate ) k2k3[4a ][5a ]/(k2[3a ] + k3[5a ])
(3)
rate law, when k2[3a ] . k3[5a ], the reaction will be first
order in [4a ] and [5a ] and inverse first order in [3a ] (eq
4). At high concentration of [5a ], the reaction should
We carried out a competition experiment to indepen-
dently confirm the value of k3/k2. The coordinatively
unsaturated imidozirconocene compound 7a is not an
isolable species, but complex 2a functions as a conve-
nient precursor to this transient intermediate (Scheme
3). A solution of zirconocene imido complex 2a in the
presence of 10 equiv of 3a and 10 equiv of 5a in d8-
toluene at -78 °C was prepared. The reaction between
2a and the unsaturated organic complexes 3a and 5a
to generate the corresponding diazametallacycle com-
plexes 4a and 6a was monitored at low temperature by
1H NMR spectroscopy. The metallacycle ratio was
measured from the integration of the cyclopentadienyl
signals for the two corresponding organometallic com-
rate ) k1k3[4a ][5a ]/(k2[3a ])
(4)
exhibit saturation kinetics wherein the rate will be first
order in [4a ] and zero order in [3a ] and [5a ] (eq 5).
rate ) k1[4a ]
(5)
Conditions for the kinetic experiments are discussed
in the Experimental Section. Each run was performed
with excess concentrations of carbodiimide 3a and
diphenylacetylene (5a ) relative to the concentration of

1796 Organometallics, Vol. 20, No. 9, 2001
Zuckerman and Bergman
Ta ble 4. Exch a n ge Rea ction of Dia za m eta lla cycle
Com p lexes 8a ,b w ith Alk yn es
diaza-
metallacycle
reactant
metalla-
cyclobutene diimide
product
carbo-
product (1H NMR)
yield, %
R′
alkyne
8a
8a
8b
SiMe3
SiMe3
t-Bu
5a
5b
5a
6c
6d
6c
3a
3a
3b
100
93
98
between 4 and alkynes, we wanted to determine whether
the substituent on the imido nitrogen would have a
similarly strong effect. We therefore prepared the aryl-
substituted diazametallacycle complex 8a with a pro-
cedure analogous to that used for the generation and
isolation of diazametallacycle complexes 4a -e (eq 6) and
F igu r e 3. Plot of kobs vs [5a ]/[3a ] for the appearance of
6a at 15.4 ( 0.1 °C and the double-reciprocal plot of 1/kobs
vs [3a ]/[5a ] (inset). A linear fit gave the correlation coef-
ficient γ ) 0.99.
Sch em e 3
pounds (δ 5.96 for 6a and δ 6.02 for 4a ) at -40 °C. To
ensure the product ratio was being measured in the
kinetic regime, monitoring of the reaction was continued
while the probe of the NMR spectrometer was warmed
to 25 °C. The ratio began to shift at -15 °C (6a :4a 0.67:l
from 6a :4a 0.4:1 at -40 °C), demonstrating the revers-
ibility of the reaction. Complete conversion to the
thermodynamic product, metallacyclobutene 6a , was
obtained by the time the probe was warmed to 25 °C.
Measuring the ratio of metallacycle complexes 4a and
6a generated under kinetic conditions and dividing out
the concentrations gave a k3/k2 value of 0.4 ( 0.04 at
-40 °C. This is in good agreement with the experimen-
tally derived value of k3/k2 ) 0.4 ( 0.01 at 15.4 ( 0.1
°C from the kinetic analysis.
examined the reaction of 8a with diphenylacetylene
(5a ). Treatment of zirconacycle 8a in C6D6 with 1 equiv
of 5a (Table 4) resulted in an immediate color change
to green from red. The 1H NMR spectroscopic data
showed the disappearance of free diphenylacetylene and
a shift in the resonance for the cyclopentadienyl ligands
of 8a from δ 6.96 to 5.83. There was virtually no change
in the chemical shift for the methyl groups on the
aromatic ring (δ 2.19), but the two Me3Si resonances of
zirconacycle 8a (δ 0.32 and 0.06) were no longer present.
A resonance corresponding to the methyl groups of free
Me3SiNdCdNSiMe3 (3a ) at δ 0.1 was observed. The
organometallic product was confirmed to be diazamet-
allacyclobutene 6c by comparison of spectroscopic data
with literature values for 6c.24 Similar to the reaction
of 4a and 2-butyne to generate metallacyclobutene 6b,
metallacycle 8a also reacted with 2-butyne to generate
azametallacyclobutene 6d (Table 4) with the concurrent
release of carbodiimide 3a .
Exch a n ge Rea ction s of Dia za m eta lla cycle Com -
p lexes 8a -d w ith Alk yn es. Since the N-R substitu-
ents from the parent carbodiimide fragment of zir-
conocene diazametallacycle complexes 4a -d were found
to have a substantial influence on the rate of exchange

Reactions between Zr Metallacycles and Organic Substrates
Organometallics, Vol. 20, No. 9, 2001 1797
Sch em e 4
Supporting Information). The appearance of product
metallacyclobutene 6c was monitored at 5.0 ( 0.1 °C
by UV-vis spectroscopy, and rate data were collected
over four half-lives at λ ) 550 nm.
A plot of [6c] vs time gave exponential growth
indicative of a pseudo-first-order reaction. The reaction
was first order in [8a ] but zero order in [5a ], even at
the lowest concentrations of this reactant that we were
conveniently able to reach while still maintaining
pseudo-first-order conditions in [8a ]. Two runs were
carried out with no added 3a while the concentration
of added 5a was varied to ensure that we achieved
saturation in the reactions with added 3a . The kobs
values for both of these reactions were virtually the
same. These data (see Table S-12 in the Supporting
Information) are consistent with the rate law described
in eq 7, with k4 ) (2.3 ( 0.02) × 10-3 s-1. For this
rate ) k4k8[8a ][5a ]/(k6[3a ] + k8[5a ])
(7)
system, k8/k6 could not be obtained from the kinetic
study due to the inaccessibility of the region in which
k6[3a ] . k8[5a ].
Kin etic An a lysis of th e Rea ction betw een Di-
a za m eta lla cycle Com p lex 8b a n d Alk yn e 5a . We
extended the above study by examining the mechanism
of the alkyne exchange reaction between metallacycle
8b and alkyne 5a (Scheme 4). Similar results were
obtained when kinetic studies were performed on this
analogous system, again utilizing a large excess of added
carbodiimide 3b and 5a relative to the concentration of
metallacycle 8b. The rate and concentration data are
listed in Table S-13 in the Supporting Information. The
data obtained for this system were consistent with the
rate law given in eq 8, but once again only the regime
Analogous exchange reactivity was observed between
diazametallacycle 8b and diphenylacetylene to cleanly
form metallacyclobutene 6c and (t-Bu)NdCdN(t-Bu)
(3b) (Table 4). Reaction of 8b with 2-butyne (5b) gave
an immediate reaction at 25 °C, demonstrated by an
instantaneous color change from purple to blue upon
mixing the two reactants. The organometallic product
was identified as 6d by comparison of spectroscopic data
with literature values for 6d .24 When zirconacycles
8c-e were treated with either diphenylacetylene or
2-butyne, there was no observable exchange reaction.
Upon further heating at elevated temperatures, only
thermal decomposition products were observed.
rate ) k5k8[8b][5a ]/(k7[3b] + k8[5a ])
(8)
in which k8[5a ] . k7[3a ] could be accessed (i.e., the
reaction was zero-order in [5a ]). The experimentally
derived value of k5 ) (2.00 ( 0.02) × 10-3 s-1 was
obtained. Unfortunately, when we increased the con-
centration of added carbodiimide 3b in order to access
the region in which k7[3a ] . k8[5a ], side reactions
occurred that prevented the collection of reliable kinetic
data.
In ser tion Rea ction s of Dia za m eta lla cycle Com -
p lexes 9a -d w it h Dip h en yla cet ylen e. Different
chemistry is observed when alkynes are allowed to react
with metallacycles that contain bulkier ortho substit-
uents on the phenyl ring. From previous studies of the
formal [2 + 2] cycloaddition reactions between zir-
conocene imido complexes with carbodiimides, we knew
that 2c does not react with Me3SiNdCdNSiMe3 or (t-
Bu)NdCdN(t-Bu), even at elevated temperatures. How-
ever, metallacycle complexes 9a ,b can be isolated in
high yields from the overall cycloaddition between 2c
and CycNdCdNCyc (3d ) and (i-Pr)NdCdN(i-Pr) (3c),
respectively (Scheme 5).26 Diazametallacycle 9a was
treated with 1 equiv of alkyne 5a at 25 °C in C6D6. No
observable reaction took place until the mixture was
heated to 75 °C, upon which a color change from purple
to orange occurred. Analysis of the reaction mixture by
1H NMR spectroscopy revealed a new organometallic
product had cleanly formed that incorporated 1 equiv
Kin etic An a lysis of th e Rea ction betw een 8a a n d
Alk yn e 5a . Comparing the reactions of the two diaza-
metallacycle complexes 4a and 8a in the exchange with
alkyne 5a , we noted qualitatively an increase in the rate
of metallacyclobutene formation with the reactant 8a .
For this reason, we decided to determine whether the
mechanistic pathway for the exchange reaction of 8a
with diphenylacetylene was a dissociative process, as
in the case of 4a and 5a (Scheme 2). If this system
follows a similar mechanism (Scheme 4), we would be
able to study the effect of the N-alkyl group on reaction
rate. This information should facilitate our understand-
ing of substituent effects in the cycloreversion/cycload-
dition of zirconocene imido complexes and alkyne 5a .
Accordingly, metallacycle 8a was treated with 10
equiv of 3a and 5a at 25 °C in C6D6, and the reaction
1
was monitored by H NMR spectroscopy. Under these
conditions, azametallacyclobutene complex 6c was cleanly
formed with no observable impurities. Using UV-vis
spectroscopy, several kinetic runs were performed in
C6H6 with excess concentrations of Me3SiNdCdNSiMe3
(3a ) and diphenylacetylene (5a ) relative to the concen-
tration of the limiting reactant zirconacycle 8a to ensure
pseudo-first-order conditions (see Table S-12 in the

1798 Organometallics, Vol. 20, No. 9, 2001
Zuckerman and Bergman
Sch em e 5
Ta ble 5. Selected Bon d Len gth s (Å) a n d Bon d
An gles (d eg) for Com p ou n d 10a
Bond Lengths
Zr1-N1
2.117(4)
2.2949(6)
1.438(8)
1.299(6)
1.495(6)
1.363(7)
1.503(7)
1.409(7)
1.378(8)
1.395(7)
1.528(9)
Zr1-N3
2.126(4)
2.1437(6)
1.453(6)
1.475(6)
1.481(7)
1.500(7)
1.515(7)
1.509(8)
1.375(8)
1.554(7)
Zr1-Cp1
N1-C11
N2-C13
N3-C13
C11-C12
C12-C13
C14-C15
C16-C17
C18-C19
C20-C21
Zr1-Cp2
N1-C14
N2-C38
N3-C44
C11-C26
C12-C32
C15-C20
C17-C18
C23-C25
Bond Angles
N1-Zr1-N3
N1-Zr1-Cp2
Cp1-Zr1-Cp2
Zr1-Ni-C14
N1-Zr1-C8
C13-N2-C38
Zr1-N3-C44
N3-Zr1-C2
C11-C12-C13
N2-C13-C12
C15-C14-C19
C14-C19-C23
N2-C38-C39
91.6(2)
103.3(1)
126.60(2)
136.7(3)
103.0(2)
123.1(5)
125.3(3)
81.5(2)
127.0(4)
124.7(4)
120.4(5)
122.8(5)
106.7(4)
N1-Zr1-Cp1
N3-Zr1-Cp1
Zr1-N1-C11
N1-C10-C5
C11-N1-C14
Zr1-N3-C13
C13-N3-C44
N1-C11-C26
C13-C12-C32
N1-C14-C15
C17-C18-C19
C15-C20-C21
N3-C44-C45
115.3(1)
106.9(1)
108.3(3)
116.7(2)
114.8(4)
117.1(3)
117.2(4)
116.5(4)
113.2(4)
121.0(5)
121.6(6)
112.2(5)
113.7(4)
F igu r e 4. ORTEP diagram of 10a . The thermal ellipsoids
are scaled to represent the 50% probability surface.
of 5a . Unlike the exchange reactivity seen between
zirconocene diazametallacycle complexes 4a ,b and 8a ,b
and alkynes, there was no evidence of free carbodiimide
being formed during the course of the reaction. Ad-
ditionally, the product exhibited two resonances in the
cyclopentadienyl region of the 1H NMR spectrum (δ 6.43
and 5.02), each integrating to five hydrogens. Repeating
the reaction on a preparative scale resulted in the
isolation of an orange solid; crystallization from hexanes
provided pure material in 65% yield. An ORTEP dia-
gram obtained from an X-ray diffraction study (Figure
4) showed the complex to have structure 10a , formed
by overall insertion of diphenylacetylene into the ArN-C
bond of diazametallacycle complex 9a . The metallacycle
ring is in a twist-boat conformation, as is expected for
an unsaturated six-membered ring with bulky substit-
uents. The N3-C13 bond distance of 1.405(6) Å is
significantly longer than the N2-C13 bond distance of
1.299(6) Å, which clearly shows substantial double-bond
character between the atoms N2 and C13 (Table 5; see
Table 2 for crystallographic data on 10a ). Double-bond
character is also observed between the atoms C11 and
C12 with a bond distance of 1.363(7) Å.
Treatment of zirconacycle 9b with 1 equiv of diphe-
nylacetylene yielded similar results. There was no
reaction until the temperature reached 75 °C, which
caused a color change from purple to yellow. Spectro-
scopic analysis by 1H NMR showed the clean conversion
to one new organometallic product that contained di-
astereotopic Cp resonances at δ 6.35 and 5.92. The
reaction was carried out on a preparative scale in C6H6
to give a yellow, viscous oil. This material gave yellow
crystals in 71% yield from Et2O at -35 °C and has been
fully characterized as the six-membered zirconacycle
10b.
We obtained an interesting result when we reversed
the order of addition of the reagents. Imido complex 2c
was treated with a stoichiometric amount of dipheny-

Reactions between Zr Metallacycles and Organic Substrates
Organometallics, Vol. 20, No. 9, 2001 1799
lacetylene to afford an immediate color change from
yellow to forest green at 25 °C (Scheme 5). Removal of
the solvent followed by crystallization from Et2O at -35
°C gave green crystals (77% yield). This complex has
been fully characterized as azametallacyclobutene 6e.
When 6e was treated with the symmetrical carbodiim-
ide CycNdCdNCyc (3d ) in C6H6, the solution turned
from green to orange over the course of 1 h at 25 °C.
Analysis of the 1H NMR spectrum revealed that the
product formed under these conditions was 10a , identi-
cal with the complex generated from the reaction of 9a
with diphenylacetylene at 75 °C.
This reaction was found to be general for the carbo-
diimide substrates tested (Scheme 5). Reaction of 6e
with 1 equiv of RNdCdNR (R ) i-Pr, p-Tol) resulted in
clean conversion to the corresponding six-membered-
ring complexes 10b,c in high yields.
Kin etic An a lysis of th e Rea ction betw een 6e a n d
1,3-Diisop r op ylca r bod iim id e (3c). Because of its
rapid rate, we suspected an associative mechanism for
the above class of reactions, involving direct insertion
of an unsaturated nitrogen-carbon moiety of the car-
bodiimide into the zirconium-carbon bond (eq 9). If this
F igu r e 5. Plot of kobs vs [3c] for the disappearance of 6e
at 15.4 ( 0.1 °C. A linear fit gave the correlation coefficient
γ ) 0.99.
reactivities observed with diazametallacycle complexes
9 and diphenylacetylene to give zirconium-containing
six-membered rings vs the reactivity examined with
zirconacycles 4a and 8a ,b with diphenylacetylene (5a )
and 2-butyne (5b). In the latter cases, only exchanges
of the alkyne with the symmetrical carbodiimide com-
ponent of the metallacycles to generate azametallacy-
clobutene complexes 6a (Table 3) and 6c (Table 4) were
observed. Additionally, there was no evidence for further
insertion of the released carbodiimides into the zirco-
nium-carbon bond of the product metallacyclobutene
complexes, unlike the reaction of 6e with symmetrical
carbodiimides (Scheme 5). To further probe the nature
of the insertion reaction, we treated azametallacy-
clobutene complex 6a with 1 equiv of (p-Tol)NdCdN(p-
Tol) (3e) in the presence of Me3SiNdCdNSiMe3 at 25
°C in C6D6 (eq 11). The solution turned from green to
mechanistic pathway is operative for the formation of
zirconacycle 10b, the rate should be second order
overall, first order in the concentration of 6e, and first
order in the concentration of carbodiimide 3c (eq 10).
rate ) k9[6e][3c]
(10)
The reaction was run under pseudo-first-order condi-
tions, with concentrations of (i-Pr)NdCdN(i-Pr) (3c)
maintained in 10-fold excess or greater relative to the
concentration of metallacyclobutene 6e (see Table S-14
in the Supporting Information for rate and concentra-
tion data for individual kinetic runs). The rate of
disappearance of [6e] vs time showed exponential decay
for 6e. Data were taken for over 11 half-lives. A plot of
kobs vs [3c] was obtained (Figure 5) and shows a linear
dependence on carbodiimide concentration 3c, with no
sign of saturation at high concentration. These data are
consistent with an associative mechanism in which
there is a first-order rate dependence on the concentra-
tion of metallacycle 6e and carbodiimide 3c. A value of
k9 ) (1.9 ( 0.1) × 10-1 M-1 s-1 at 15.4 ( 0.1 °C was
derived from the experimental data.
orange over 30 min. Analysis of the reaction mixture
by 1H NMR spectroscopy revealed two new distinct
cyclopentadienyl signals, each integrating to five hy-
drogens, and the incorporation of (p-Tol)NdCdN(p-Tol)
into the organometallic product. The reaction was
repeated on a preparative scale to give zirconacycle 11
as an orange powder in 97% yield.
F u r th er In ser tion Rea ction s of Aza m eta lla cy-
clobu ten e Com p lexes 6a ,c w ith (p-Tol)NdCdN(p-
Tol). We were curious about the cause of the different

1800 Organometallics, Vol. 20, No. 9, 2001
Zuckerman and Bergman
Sch em e 6
Ta ble 6. Exch a n ge Rea ction of Dia za m eta lla cycle
Com p lexes 4a a n d 8a w ith Ca r bod iim id es
reactant
diaza-
reactant
carbo-
product
diaza-
yield,
% (1H
metallacycle
R
R′
diimide metallacycle NMR)
4a
4a
4a
4a
8a
8a
8a
t-Bu
t-Bu
t-Bu
t-Bu
i-Pr
Cyc
p-Tol
m-Tol
3c
3d
3e
3f
3c
3d
3e
4c
4d
4e
4f
8c
8d
8e
92
95
97
100
97
100
87
2,6-Me2Ph i-Pr
2,6-Me2Ph Cyc
2,6-Me2Ph p-Tol
We next extended our investigation of carbodiimide
insertion into zirconacyclobutene complexes to 6c, the
aryl analogue of 6a . Treatment of 6c with 1 equiv of
(p-Tol)NdCdN(p-Tol) (3e) in the presence of Me3SiNd
CdNSiMe3 (3a ) gave results similar to those obtained
when 6a was treated with (p-Tol)NdCdN(p-Tol) (eq 11).
An orange color developed over time, and spectroscopic
analysis showed the split Cp resonances characteristic
of all the carbodiimide-inserted products prepared in
this study. Organometallic complex 12 was formed in
90% NMR yield based on an internal standard. Mass
spectroscopic analysis of the mixture showed a molec-
ular ion corresponding to the molecular weight of 12.
Treatment of 6c with (p-Tol)NdCdN(p-Tol) in the
presence of (t-Bu)NdCdN(t-Bu) (3b) also gave only
complex 12 (1H NMR spectroscopy).
observed when the reactant carbodiimide was 3c,d . The
solution turned from red to purple during both of these
reactions, and an evaluation of the resulting mixtures
by 1H NMR spectroscopy confirmed clean conversion to
the zirconium-containing diazacyclobutane complexes
8c,d , respectively.
Kin etic An a lysis of th e Rea ction betw een 4a a n d
th e Sym m etr ica l Ar yl Ca r bod iim id es 3e,f. To better
understand the mechanistic pathway involved in these
reactions, we undertook kinetic studies of the reaction
between metallacycle 4a and carbodiimide 3e using
UV-vis spectroscopy. Outlined in Scheme 6 is the
mechanistic model used for our kinetic analysis; the
corresponding rate expression is given in eq 12. Condi-
Exch a n ge Rea ction s of Dia za m eta lla cycle Com -
p lexes 4a a n d 8a w ith Sym m etr ica l Ca r bod iim -
id es. We pursued the study of exchange reactions
between diazametallacycle complexes and other sym-
metrical carbodiimides (Table 6) in order to gain further
mechanistic insight into the factors that govern cyclo-
reversion/cycloaddition transformations of these com-
plexes with unsaturated organic substrates. Treatment
of diazametallacycle 4a with 1 equiv of 1,3-di-p-tolyl-
carbodiimide (3e) resulted in a color change from
reddish orange to blue within several hours at 25 °C.
Spectroscopic analysis (1H and 13C NMR) confirmed that
the resulting products formed from this reaction were
the previously characterized zirconacycle 4d 26 and Me3-
SiNdCdNSiMe3 (3a ). Similar results were obtained
with carbodiimides 3c,d (Table 6). In both of these cases,
the reaction mixture turned from reddish orange to
purple within hours to generate diazametallacycle
complexes 4c,d with the concomitant formation of
carbodiimide 3a .
rate ) k1k10[4a ][3e]/(k2[3a ] + k10[3e])
(12)
tions for the kinetic experiments are discussed in the
Experimental Section. We initially attempted to hold
the concentrations of added Me3SiNdCdNSiMe3 (3a )
and (p-Tol)NdCdN(p-Tol) (3e) at 10-fold excess or
greater relative to the concentration of 4a to ensure
pseudo-first-order conditions in the disappearance of 4a .
However, this was frustrated by the lack of solubility
of carbodiimide 3e in C6H6. We therefore decided to
modify the study by employing carbodiimide 3e as the
limiting reactant and using concentrations of diazamet-
allacycle 4a and 3a in a minimum of 10-fold excess
relative to the concentration of 3e. This gives the rate
expression that is depicted in eqs 13 and 14. Table S-15
rate ) kobs[3e]
(13)
(14)
kobs ) k1k10[4a ]/(k2[3a ] + k10[3e])
Extending this study, treatment of 8a with 1 equiv
of diarylcarbodiimide 3e resulted in the release of
carbodiimide 3a with the simultaneous formation of 8e
in 87% 1H NMR yield (Table 6). It is interesting to note
that the reaction proceeded more quickly than the
reaction of 4a and 3e. Carbodiimide exchange was also
in the Supporting Information contains the concentra-
tion and rate data for the individual kinetic runs.
Excellent pseudo-first-order behavior was observed for
this system.

Reactions between Zr Metallacycles and Organic Substrates
Organometallics, Vol. 20, No. 9, 2001 1801
F igu r e 6. Plot of 1/kobs vs [3a ]/[4a ]. A linear fit gave the
correlation coefficient γ ) 0.99.
F igu r e 7. Plot of kobs vs [3f]/[3a ] for the appearance of 4e
at 15.4 ( 0.1 °C and the double-reciprocal plot of 1/kobs vs
[3a ]/[3f] (inset). A linear fit gave the correlation coefficient
γ ) 0.99.
Plots of the increase in [4d ] vs time gave exponential
growth consistent with the reaction being first order in
[3e]. Rates measured using variable concentrations of
both 4a and 3a demonstrated that, under conditions
where k2[3a ] . k10[3e], the rate was accelerated by
added 4a and inhibited by added carbodiimide 3a (eq
15). A plot of 1/kobs versus [3a ]/[4a ] shows a linear
the transient imidozirconocene complexes 7 with the
release of free carbodiimide might help to determine
why only certain N-R imido compounds are formed. For
example, the exchange reactions observed with diaza-
metallacycle complexes 8a ,b with alkynes resulted in
azametallacyclobutene complex 6c,d , depending on the
alkyne (Table 4). Treatment of Cp2(THF)ZrdN(t-Bu)
(2a ) with 1 equiv of diazametallacycle 8a in C6D6 at 25
°C resulted in the quantitative formation of Cp2(THF)-
ZrdN(2,6-Me2Ph) (2b) and diazametallacycle complex
4a (eq 17).
kobs ) k1k10[4a ]/(k2[3a ])
(15)
dependence on the carbodiimide/metallacycle ratio (Fig-
ure 6). The experimentally derived value for the rate
constant ratio k2/k10k1 is (8.1 ( 0.1) × 103 s-1 (Scheme
6).
We next replaced carbodiimide 3e with an analogue,
1,3-di-m-tolylcarbodiimide (3f) (Scheme 6); the rate
expression for this system is outlined in eq 16. In this
rate ) k1k11[4a ][3f]/(k2[3a ] + k11[3f])
(16)
case, there were no observable solubility problems
during our analysis by UV-vis spectroscopy, even at
higher concentrations of 3f. We were able to conve-
niently use excess concentrations of added 3a ,f relative
to the concentration of metallacycle complex 4a . Indi-
vidual kinetic runs were carried out at 15.4 ( 0.1 °C,
and rate data were taken for over six half-lives. Rate
and concentration data for the kinetic runs carried out
during the study are listed in Table S-16 in the Sup-
porting Information. A plot of kobs versus [3f]/[3a ] was
obtained (Figure 7). An inverse plot of l/kobs vs [3a ]/[3f]
(Figure 7, inset) shows a linear dependence on the
carbodiimide concentration ratio.
These data are once again consistent with a dissocia-
tive mechanism (Scheme 6) and the rate law given in
eq 16. Values of k1 ) (4.82 ( 0.02) × 10-4 s-1 and k2/k11
) 0.8 were derived from the inverse first-order plot. The
back-reaction corresponding to the trapping of zir-
conocene imido complex 7a by carbodiimide 3a is faster
than the reaction with 1,3-di-m-tolylcarbodiimide (3f)
to generate the thermodynamically favored product 4e.
Im id o Gr ou p Tr a n sfer . We wanted to examine the
relative thermodynamic stability of zirconocene imido
complexes and the corresponding diazametallacycle
complexes. Greater insight into the significant factors
influencing the retrocycloaddition reaction to generate
Discu ssion
Exch a n ge Rea ction s of Dia za m eta lla cycle Com -
p lexes 4a a n d 8a ,b w ith Alk yn es. The results re-
ported here provide mechanistic insight into the cyclo-
reversion/cycloaddition processes of Cp2ZrdNR species
with heterocumulenes, transformations critical for sev-
eral imidozirconocene-mediated stoichiometric and cata-
lytic processes.5,10,18-21,29,30 As shown in Table 3, when
zirconacycle 4a was treated with alkyne 5a , metalla-
cyclobutene complex 6a formed cleanly with the concur-
rent release of carbodiimide 3a . We considered two
possible pathways for this reaction. The first is an
associative process wherein diphenylacetylene (5a ) ini-
(29) Lee, S. Y.; Bergman, R. G. J . Am. Chem. Soc. 1995, 117, 5877.
(30) Lee, S. Y.; Bergman, R. G. Tetrahedron 1995, 51, 4255.

1802 Organometallics, Vol. 20, No. 9, 2001
Zuckerman and Bergman
Sch em e 7
tially reacts with 4a . This reaction would follow a rate
law that is overall second order: first order in [4a ] and
first order in [5a ]. The second is the dissociative process
outlined in Scheme 2. This begins with initial overall
[2 + 2] retrocycloaddition of 4a to generate the coordi-
natively unsaturated zirconocene imido complex 7a with
the release of the symmetrical carbodiimide 3a . Imi-
dozirconocene 7a can be trapped by either 3a to
regenerate the starting zirconacycle or by 5a to form
azametallocyclobutene complex 6a .
Applying the steady-state approximation to the con-
centration of imidozirconocene 7a gives the rate expres-
sion that is shown in eq 3. The experimental rate data
obtained fit this rate law and thus support the dissocia-
tive pathway depicted in Scheme 2. The rate constant
ratio k2/k1k3 ) (7.5 ( 0.2) × 103 s-1 was obtained from
the plot of 1/kobs vs [3a ]/[5a ] (Figure 3, inset). The
derived rate constant for the initial cycloreversion of 4a
to yield transient imidozirconocene 7a and the released
carbodiimide 3a was experimentally determined to be
k1 ) (3.3 ( 0.1) × 10-4 s-1 at 15.4 ( 0.1 °C. Once 7a
forms, the rate of trapping with 3a to regenerate 4a is
2.5 times faster than the rate of trapping 7a with 5a to
produce azametallacyclobutene 6a (k3/k2 ) 0.4).
Acquisition of rate data for the exchange reaction
between zirconacycle 8a and 5a (Table 4) allows for the
comparison with rate data obtained for the system of
4a and alkyne 5a to afford the corresponding azamet-
allacyclobutene complex 6a and symmetrical carbodi-
imide 3a (Table 3). The only structural difference
between these two systems is the parent imido sub-
stituent (t-Bu for 4a and 2,6-Me2Ph for 8a ). There were
several notable differences between these two systems.
For example, the reaction between 8a and 5a was too
fast to monitor at 15.4 °C by UV-vis spectrometry. The
enhanced rate for this system necessitated that the
study be carried out at 5.0 ( 0.1 °C in order to obtain
rate data for >4 half-lives. The reaction was determined
to be first order in [8a ] but zero order in [5a ] even at
the lowest concentrations of the latter reactant that we
were conveniently able to reach while still maintaining
pseudo-first-order conditions in [8a ]. These data (Table
S-12 in the Supporting Information) are consistent with
the mechanism in Scheme 6 and the rate law described
in eq 8. The rate constant for the cycloreversion of 8a
to give 7b and 3a (k4 ) (2.3 ( 0.02) × 10-3 s-1 at 5.0 (
0.1 °C) is an order of magnitude greater than the rate
constant for the unimolecular cycloreversion of 4a to
generate 7a and 3a (k1 ) (3.3 ( 0.1) × 10-4 s-1 at 15.4
( 0.1 °C). It is difficult to say whether this effect should
be attributed to a ground-state destabilization of the
metallacycle caused by replacing the t-Bu substituent
on the imido nitrogen with a di-ortho-substituted aryl
moiety or if the increased rate of reversion results from
stabilization of the transition state leading from 8a to
7b. Unfortunately, k8/k6 for this system could not be
obtained from the kinetic study due to the inaccessibility
of the kinetic region where k6[3a ] . k8[5a ].
the symmetrical carbodiimide (t-Bu)NdCdN(t-Bu) (Table
4) was consistent with the dissociative mechanism
outlined in Scheme 4. As in the above reactions, the
transformation was determined to be first order in [8b]
but zero order in [5a ]. Under all conditions employed,
the rate-determining transition state was found to be
that for retrocycloaddition of the reactant diazametal-
lacycle 8b to generate imidozirconocene complex 7b and
the released carbodiimide 3b. Because subsequent trap-
ping with diphenylacetylene to afford azametallacy-
clobutene complex 6c was fast, the relative rate constant
for this process was kinetically inaccessible in all the
successful runs of the study. The experimentally derived
value of the rate constant k5 for the unimolecular
cycloreversion of 8b at 5.0 ( 0.1 °C is (2.00 ( 0.02) ×
10-3 s-1
.
In ser tion Rea ction s of Dia za m eta lla cycle Com -
p lexes 9a ,b w ith Dip h en yla cetylen e. Unlike the
exchange reactions between zirconacycle compounds 4
and 8 with alkynes to generate azametallacyclobutene
complexes 6 and symmetrical carbodiimides, a different
type of reactivity was observed for diazametallacycle
complexes 9a ,b when they were treated with 1 equiv of
diphenylacetylene (5a ). Zirconium diazacyclobutane
species 9a ,b gave overall insertion products 10a ,b upon
heating in the presence of diphenylacetylene (Scheme
5). Outlined in Scheme 7 are two possible mechanistic
routes that lead to the generation of the six-membered
zirconacycle complexes 10. In the first pathway (a),
direct insertion of diphenylacetylene into the nitrogen-
carbon bond of diazametallacycle complex 9 occurs. To
our knowledge, a direct insertion of this type is unprec-
edented.
The second proposed pathway (b) takes place by initial
cycloreversion of diazametallacycle complex 9 to gener-
ate the transient zirconocene imido species 7c and the
organic carbodiimide RNdCdNR. Once the coordina-
tively unsaturated 16-electron imidozirconocene com-
plex 7c has formed, it can be trapped by an overall [2
+ 2] cycloaddition reaction with free carbodiimide to
We next examined a system in which the parent
carbodiimide component of the zirconacycle was varied
while retaining the nitrogen-containing aromatic sub-
stituent from the parent imidozirconocene complex 2b.
Kinetic analysis of the exchange reaction between 8b
and alkyne 5a to produce azametallacyclobutene 6c and

Reactions between Zr Metallacycles and Organic Substrates
Organometallics, Vol. 20, No. 9, 2001 1803
Sch em e 8
regenerate the starting material or by alkyne 5a to form
the corresponding diazametallacyclobutene complex 6.
The final six-membered zirconacycle complex 10 is then
formed by direct insertion of the N-C double bond in
free carbodiimide 3 into the zirconium-carbon bond of
6.
We favor pathway (b) for two reasons. First, carbo-
diimide insertion into a metal-carbon bond is well
precedented.31-33 For example, Gamboratta and co-
workers report the insertion of (p-Tol)NdCdN(p-Tol)
into the zirconium-carbon σ bond of dimethylzir-
conocene.34,35 Second, 10a -c can be prepared indepen-
dently by treatment of azametallacyclobutene complex
6e with carbodiimides RNdCdNR (R ) Cyc, i-Pr, p-Tol)
at 25 °C (Scheme 5). This reaction is very rapid
compared to the reaction of carbodiimide adduct 9 with
diphenylacetylene (5a ), supporting the conclusion that
6e is a kinetically competent intermediate for the
conversion of 9 to 10. The rapid rate of reaction of 6e
with (i-Pr)NdCdN(i-Pr) (3c) is consistent with our
kinetic study establishing an associative pathway that
is second order overall, first order in [6e] and first order
in [3c], for this reaction.
associative pathway in which the reactant carbodiimide
directly inserts into the zirconium-nitrogen bond of the
four-membered ring to generate six-membered zircona-
cycle 13. Extrusion of the symmetrical carbodiimide 3a
from 13 would give the new metallacyclobutane com-
plex. The second proposed pathway is a dissociative
process (Scheme 6), analogous to the mechanism estab-
lished for the exchange reaction of 4a with dipheny-
lacetylene (Scheme 2).
In ser tion Rea ction s of Dia za m eta lla cyclobu ten e
Com p lexes 6a ,c w ith (p-Tol)NdCdN(p-Tol) (3e). To
better understand the nature of the direct insertion of
free carbodiimide into the zirconium-carbon bond of
azametallacyclobutene complexes, we treated 6a with
(p-Tol)NdCdN(p-Tol) in the presence of Me3SiNdCd
NSiMe3. The six-membered zirconacycle 11 was formed
in high yield (eq 11). We propose a direct carbodiimide
insertion that is analogous to the mechanism estab-
lished for the reaction between azametallacyclobutene
complex 6e and (i-Pr)NdCdN(i-Pr) (3c) to generate 10b
(Scheme 7). The fact that there is no observable inser-
tion of free Me3SiNdCdNSiMe3 into the zirconium-
carbon bond of 6a under the reaction conditions em-
ployed suggests that there is a strong steric effect in
the insertion reaction. When aryl analogue 6c was
treated with (p-Tol)NdCdN(p-Tol) in the presence of
either Me3SiNdCdNSiMe3 or (t-Bu)NdCdN(t-Bu), com-
plex 12 was the only product (eq 11). These results
further support the steric control hypothesis.
Exch a n ge Rea ction s of Dia za m eta lla cycle Com -
p lex 4a w ith Sym m etr ica l Ca r bod iim id es. The
stoichiometric exchange between diazametallacycle com-
plex 4a and carbodiimides to afford a new zirconacycle
with the release of carbodiimide 3a is directly analogous
to the mechanistic pathway that Weiss and co-workers
postulated for tungsten imido-catalyzed metathesis
reactions between carbodiimides (Scheme 1). We saw
the opportunity to utilize the zirconium diazametalla-
cycle complexes in the exchange reaction with other
carbodiimides as a model system to gain mechanistic
insight into this type of reaction.
To distinguish these possibilities, kinetic analysis of
the carbodiimide exchange reaction was carried out
under pseudo-first-order conditions. Because of the
limited solubility of 1,3-di-p-tolylcarbodiimide (3e) under
the reaction conditions, the study was carried out with
excess concentrations of 4a and 3a , with carbodiimide
3e as limiting reactant. The reaction was found to be
first order in 3e, and the experimental data are consis-
tent with the derived rate law in eqs 13-15. The rate
constant ratio of k1k10/k2 ) (8.1 ( 0.1) × 103 s-1 was
obtained from the plot of 1/kobs vs [3a ]/[4a ] (Figure 6).
The kinetic data support the dissociative process out-
lined in Scheme 6.
Because we could not reach saturation conditions in
[4a ], we were unable to directly obtain the value for the
rate constant k1. To address this problem, we initiated
a kinetic analysis of a related system in which the
reactant carbodiimide (3e) was replaced with 1,3-di-m-
tolylcarbodiimide (3f) to overcome the solubility prob-
lems that we encountered with high concentrations of
3e. Pseudo-first-order conditions were maintained dur-
ing the course of the reaction by using excess concentra-
tions of meta-substituted aryl carbodiimide 3f and silyl-
substituted carbodiimide 3a . Saturation kinetics were
observed at high concentrations of 3f and the value for
k1 could be obtained ((4.82 ( 0.02) × 104 s-1). Addition-
ally, the ratio k2/k11 equals 0.8. Thus, trapping of
imidozirconocene 7a with silyl-substituted carbodiimide
3a occurs 1.25 times faster than the rate of trapping
7a with 3f.
Since k1 corresponds to the rate constant for the
unimolecular formal [2 + 2] retrocycloaddition of zir-
conacycle 4a to generate zirconocene imido complex 7a
and carbodiimide 3a , the value experimentally derived
for k1 in the exchange reaction of 4a with 3f should be
the same as that determined in the exchange reaction
between 4a and the para-substituted carbodiimide 3e.
Substituting the experimentally determined k1 value
obtained in the exchange reaction between 4a and 3f
((4.82 ( 0.02) × 10-4 s-1) into the experimentally
derived rate constant ratio of k1k10/k2 ) (8.1 ( 0.1) ×
There are two reasonable mechanistic pathways for
this process. One, depicted in Scheme 8, involves an
(31) Li, M.-D.; Chang, C. C.; Wang, Y.; Lee, G. H. Organometallics
1996, 15, 2571.
(32) Chang, C. C.; Hsiung, C.-S.; Su, H.-L.; Srinivas, B.; Chiang,
M. Y.; Lee, G.-H.; Wang, Y. Organometallics 1998, 17, 1595.
(33) Shapiro, P. J .; Henling, L. M.; Marsh, R. E.; Bercaw, J . E. Inorg.
Chem. 1990, 29, 4560.
(34) Gambarotta, S. S.; Floriani, C.; Chiesi-Villa, A.; Guastini, C.
Inorg. Chem 1985, 24, 654.
(35) Gambarotta, S. S.; Floriani, C.; Chiesi-Villa, A.; Guastini, C.
J . Am. Chem. Soc. 1985, 107, 6278.

1804 Organometallics, Vol. 20, No. 9, 2001
Zuckerman and Bergman
Sch em e 9
103 s-1 measured in the reaction between 4a and 3e,
we calculate k10/k2 ) 1.67. Thus, for this case, the
trapping of imidozirconocene 7a with para-substituted
aryl carbodiimide 3e is over 1.5 times faster than the
rate of trapping 7a with 3a . We attribute the difference
in trapping rates of imidozirconocene complex 7a with
3e,f to steric factors in the aryl carbodiimides.
Qualitatively, the relative rates of reaction with
carbodiimides were much faster for this aryl imido
species Cp2ZrdN-2,6-Me2Ph than the N-alkylated imido
complex Cp2ZrdN-t-Bu. Since we know the k1 value for
the formal [2 + 2] retrocycloaddition reaction of 8a to
generate 7b and alkyne 3a is an order of magnitude
greater than in the case of the N-t-Bu analogue 4a , we
can infer that zirconocene imido 7b is being generated
in the carbodiimide exchange reactions in an analogous
dissociative mechanism that has been established with
4a and carbodiimides 3e,f.
Im id o Gr ou p Tr a n sfer . Our experimental results
suggest that arylzirconocene imido complexes are ther-
modynamically favored over the corresponding N-alkyl
complexes. In all the cases examined in this study,
retrocycloaddition of diazametallacycle complexes 8
(Scheme 4) and 9 (Scheme 7) resulted in the generation
of zirconocene arylimido complexes 7b,c, respectively,
with release of symmetrical dialkylcarbodiimides. The
thermodynamic driving force could be due to either the
inherent stability of the arylimido complex, the stability
of the symmetrical carbodiimide, or both. However,
when diazametallacycle 8a was treated with an equimo-
lar amount of Cp2(THF)ZrdN(t-Bu) (2a ), quantitative
conversion to the zirconocene imido complex Cp2-
(THF)ZrdN(2,6-Me2Ph) (2b) and diazametallacycle com-
plex 4a was observed (eq 17). This result is consistent
with the postulate of greater thermodynamic stability
of the N-aryl imidozirconocene complexes over their
N-alkyl analogues.
During the course of this study, cycloreversion of 4a
consistently resulted in the generation of Cp2ZrdN(t-
Bu) (7a ) and the release of Me3SiNdCdNSiMe3. How-
ever, from previous studies,26 treatment of diazametal-
lacycle complex 4b with 1 equiv of (i-Pr)NdCdN(i-Pr)
resulted in generation of the symmetrical metallacycle
14 with the release of (t-Bu)NdCdN(i-Pr) (Scheme 9).
The proposed mechanism for this process invokes initial
cycloreversion of 4b to generate the zirconocene imido
complex 7d and the unsymmetrical carbodiimide. Once
complex 7d forms, it is rapidly trapped by (i-Pr)NdCd
N(i-Pr) to generate 14. In contrast to this observation,
we have never observed the release of an aryl/alkyl
carbodiimide from an appropriately substituted diaza-
metallacycle.
1. Kinetic studies have established that the heterocu-
mulene exchange takes place by a dissociative pathway
that involves initial cleavage of the four-membered
metallacycle and regeneration of the transient 16-
electron imidozirconocene intermediate Cp2ZrdNR.
2. The overall course and rate of these reactions are
quite sensitive to the presence of other unsaturated
molecules and to the structural and electronic properties
of both the metallacycle and heterocumulene substrates.
For example, when a zirconocene/heterocumulene ad-
duct metallacycle is thermally cleaved in the presence
of an alkyne, the 16-electron zirconocene imido complex
so generated is trapped by the alkyne to afford the
corresponding azametallacyclobutene complex. The ac-
tivation barrier for the initial metallacycle cleavage step
in the retrocycloaddition is lower when it is possible to
generate an imido complex with an aryl rather than an
alkyl N substituent (e.g., the cleavage of N-aryl metal-
lacycle 8a occurs more rapidly than the cleavage of
N-alkyl-substituted metallacycle 4a ). In general, we find
that N-arylimidozirconocene complexes form preferen-
tially over N-alkylimido complexes.
3. The metallacyclobutenes undergo further insertion
reactions with carbodiimides, leading to ring-expanded
six-membered heterometallacycles (e.g., 10a -c). This
requires relatively small N substitutents; when the
carbodiimide is sterically bulky, this reaction does not
occur.
Con clu sion s
In this study, we have established that the overall [2
+ 2] cycloadditions that take place between imidozir-
conocene complexes and heterocumulenes are reversible
and that this reversibility is general. Thus, the system
allows zirconocene complexes to mediate overall het-
erocumulene metathesis reactions. These reactions
provide a useful method for the generation of new
zirconium-containing metallacycles and small, unsatur-
ated organic substrates. The detailed mechanistic study
reported in this paper has allowed us to define the
following properties of this system.
4. The cycloaddition/ring expansion process can also
be carried out in the reverse sequence: that is, by
preparing the heterocumulene adduct first and then
treating it with an alkyne. Interestingly, this results in
formation of the same six-membered metallacycle (e.g.
10a ,b) as that formed when the metallacyclobutene is
treated with a carbodiimide. This implicates the met-
allacyclobutene (e.g., 6e, Scheme 7) as an intermediate
in both of these reactions.
5. Kinetic studies on the reaction of zirconacy-
clobutene 6e with (i-Pr)NdCdN(i-Pr) to generate 10b

Reactions between Zr Metallacycles and Organic Substrates
Organometallics, Vol. 20, No. 9, 2001 1805
149.2, 147.1, 141.1, 131.9, 131.8, 130.5, 127.1, 127.0, 124.1,
123.4, 122.4, 111.0. Anal. Calcd for C36H37NZr: C, 75.21; H,
6.49; N, 2.44. Found: C, 75.40; H, 6.68; N, 2.26.
support an associative pathway for the insertion of (i-
Pr)NdCdN(i-Pr) into the zirconium-carbon bond of 6e.
10a . Meth od A. Diphenylacetylene (55 mg, 0.30 mmol) and
diazametallacycle 9a (184 mg, 0.30 mmol) in benzene (5 mL)
were added to a glass reaction vessel that was equipped with
a Teflon stopcock and stir bar. The solution was heated to 75
°C, and an orange color developed slowly. After it was heated
for 15 h, the solvent was removed at reduced pressure to leave
an orange solid which was dissolved in hexanes (3 mL) and
stored at -35 °C for 45 days to afford orange prisms (152 mg,
65% yield). 1H NMR (C6D6): δ 7.22 (d, J ) 5 Hz, 2H, aryl),
6.85 (t, J ) 15 Hz, aryl), 6.72 (m, 5H, aryl), 6.42 (d, J ) 7 Hz,
2H, aryl), 6.36 (s, 5H, C5H5), 5.95 (s, 5H, C5H5), 3.72 (m, 2H,
CH(CH3)2), 3.36 (m, 1H, CH), 3.15 (m, 1H, CH), 2.38-0.68 (m,
20H, cyclohexyl CH2’s), 1.61 (d, J ) 6 Hz, 3H, CH(CH3)2), 1.53
(d, J ) 7 Hz, 3H, CH(CH3)2), 1.00 (d, J ) 7 Hz, 3H, CH(CH3)2),
-0.03 (d, J ) 7 Hz, 3H, CH(CH3)2). 13C{1H} NMR (C6D6): δ
155.5, 150.9, 145.9, 144.2, 139.3, 131.92, 130.4, 130.2, 129.6,
128.5. 127.2, 126.6, 126.2, 125.6, 124.9, 124.2, 113.5, 111.2,
66.2, 56.3, 37.9, 35.3, 33.2, 29.2, 28.2, 28.1, 27.3, 27.1, 26.9,
26.3, 25.8, 25.5, 25.1, 24.5, 23.1. HREIMS: m/z calcd for
Exp er im en ta l Section
Gen er a l Meth od s. Unless otherwise noted, all manipula-
tions were carried out under an inert atmosphere in a Vacuum
Atmospheres 553-2 drybox with an attached M6-40-Dritrain
or with use of standard Schlenk or vacuum line techniques.
Degassed solutions were frozen to -196 °C, evacuated under
high vacuum, and thawed.
Sealed NMR tubes (Wilmad 505-PP and 504-PP) were
prepared by attaching Cajon adapters directly to Kontes
vacuum stopcocks and flame-sealing. J . Young tubes refer to
resealable NMR tubes fitted with Teflon stopcocks. Glassware
was dried in an oven at 150 °C before use. Pentane, hexanes,
diethyl ether, toluene, benzene, and THF were distilled from
Na/benzophenone. Sieves (4 Å) were activated by heating
under vacuum at 200 °C for 24 h. The reagents 2,6-dimethy-
laniline and 2,6-diisopropylaniline were purchased from Ald-
rich, distilled from CaH2, and stored over 4 Å sieves in the
glovebox. Diphenylacetylene was purchased from Aldrich and
sublimed prior to use. Carbodiimides 3a -e were purchased
from Aldrich and used without further purification. Complexes
2a -c,25a,26 3f,25b 4a -d ,26 6a -d ,9,10,25 8a -e,26 and 9a ,b26 were
prepared by literature methods and crystallized prior to use.
All 1H and 13C{1H} NMR spectra were recorded on com-
mercial Bruker AMX 300 and DRX 500 spectrometers. Chemi-
cal shifts (δ) are reported in parts per million (ppm) relative
to residual protiated solvent: C6D6 (7.15 ppm), THF-d8 (3.58
ppm). All 13C{1H} NMR spectra were recorded at 100 MHz
and are reported in ppm relative to the carbon resonance of
the deuterated solvent C6D6 (128.0 ppm). Mass spectrometric
analyses were obtained at the University of California, Ber-
keley Mass Spectrometry Facility, on a VT Prospec mass
spectrometer. Elemental analyses were performed at the
University of California, Berkeley Microanalytical Facility, on
a Perkin-Elmer 2400 Series II CHNO/S analyzer. Kinetic
studies were performed with a Hewlett-Packard 8453 UV-
vis spectrometer.
Cp 2Zr (N(t-Bu )CdN(m -Tol)N(m -Tol)) (4e). A glass reac-
tion vessel equipped with a Teflon stopcock was charged with
Cp2(THF)ZrdN-t-Bu (2a ; 257 mg, 0.70 mmol) and C6H6 (10
mL). The solution was stirred, and 1,3-di-m-tolylcarbodiimide
(156 mg, 0.70 mmol) was added dropwise. The solution turned
from yellow to purple immediately. After the mixture was
stirred at 25 °C for 20 min, the solvent was removed under
reduced pressure to afford a purple solid. The solid was
crystallized from benzene layered with hexanes at 25 °C for 2
days to afford deep purple needles (259 mg, 72% yield). 1H
NMR (C6D6): δ 7.02 (t, J ) 7 Hz, 1H, aryl), 6.91 (m, 3H, aryl),
6.52 (s, 1H, aryl), 6.50 (d, J ) 7 Hz, 1H, aryl), 6.45 (d, J ) 7
Hz, 1H, aryl), 5.99 (s, 10OH, C5H5), 2.21 (s, 3H, CH3), 2.09 (s,
3H, CH3), 1.48 (s, 9H, C(CH3)3); 13C{1H} NMR (C6D6): δ 151.2,
150.2, 142.8, 137.4, 136.9, 128.3, 123.9, 121.9, 120.6, 119.9,
119.8, 117.9, 115.8, 108.8, 34.4, 30.2, 25.8, 21.8, 21.6. Anal.
Calcd for C29H33N3Zr: C, 67.66; H, 6.46; N, 8.16. Found: C,
67.99; H, 6.64; N, 7.86.
C
49H59N3Zr 779.3760, found 779.3756.
Meth od B. Diazametallacyclobutene 6e (100 mg, 0.17
mmol) and 1,3- dicyclohexylcarbodiimide (36 mg, 0.17 mmol)
were dissolved in benzene (3 mL). An orange color developed
over 30 min. The solution was stirred for 1 h at 25 °C and
evaporated to dryness at reduced pressure to afford an orange
powder. The solid was redissolved in Et2O (2 mL), and the
resulting solution was stored at -35 °C for 2 days. Complex
10a was isolated as orange crystals (99 mg, 75% yield) of 10a .
10b. Meth od A. Diphenylacetylene (58 mg, 0.326 mmol)
and diazametallacycle 9b (170 mg, 0.326 mmol) in benzene (5
mL) were added to a glass reaction vessel that was equipped
with a Teflon stopcock and stir bar. The solution was heated
to 75 °C, and a yellow color developed after heating for 20 h.
The solvent was removed at reduced pressure to leave a yellow
oil which was dissolved in Et2O (2 mL) and stored at -35 °C
for 2 days to afford analytically pure 10b as a yellow powder
(162 mg, 71%). 1H NMR (C6D6): δ 7.50 (t, J ) 6 Hz, 1H, aryl),
7.17-6.45 (13H, aryl), 6.34 (s, 5H, C5H5), 5.92 (s, 5H, C5H5),
3.69 (m, 1H, CH(CH3)2), 2.89 (m, 1H, CH(CH3)2), 2.68 (m, 1H,
CH(CH3)2), 1.95 (d, J ) 5 Hz, 3H, CH(CH3)2), 1.77 (d, J ) 5
Hz, 3H, CH(CH3)2), 1.60 (d, J ) 5 Hz, 3H, CH(CH3)2), 1.52 (d,
J ) 5 Hz, 3H, CH(CH3)2), 1.30 (d, J ) 5 Hz, 3H, CH(CH3)2),
1.00 (d, J ) 5 Hz, 3H, CH(CH3)2), 0.27 (d, J ) 5 Hz, 3H, CH-
(CH3)2), -0.03 (d, J ) 5 Hz, 3H, CH(CH3)2). 13C{1H} NMR
(C6D6): δ 154.9, 150.8, 145.8, 144.6, 144.2, 143.9, 139.3, 131.9,
130.2, 130.1, 129.5, 128.6, 128.5, 128.4, 127.8, 127.4, 127.2,
126.7, 126.2, 125.7, 124.9, 124.2, 113.6, 113.5, 111.3, 111.2,
55.2, 48.0, 28.2, 27.4, 27.0, 26.4, 25.0, 24.8, 24.7, 24.4, 23.2,
22.4, 19.5. HREIMS: m/z calcd for C43H51N3Zr 701.3140, found
701.3133.
Meth od B. Diazametallacyclobutene 6e (81 mg, 0.14 mmol)
and 1,3- diisopropylcarbodiimide (28 mg, 0.14 mmol) were
dissolved in 3 mL of benzene. The color changed from green
to yellow while the solution was stirred at 25 °C for 1 h. The
solvent was removed at reduced pressure and redissolved in
Et2O (2 mL). The solution was stored at -35 °C for 2 days. A
yellow powder (80 mg, 82% yield) of 10b was obtained.
10c. 1,3-Di-p-tolylcarbodiimide (46 mg, 0.21 mmol) and
metallacyclobutene complex 6e (70 mg, 0.21 mmol) were
dissolved in benzene (5 mL). The solution turned from green
to orange while it was stirred at 25 °C for 1 h. The reaction
mixture was evaporated to dryness, the residue was redis-
solved in Et2O (3 mL), and this solution was stored at -35 °C
for 1 day. Analytically pure 10c was obtained as an orange
Cp 2Zr (N(2,6-i-P r 2P h )C(P h )dC(P h )) (6e). The zirconocene
imido complex 2c (380 mg, 0.81 mmol) and diphenylacetylene
(145 mg, 0.81 mmol) were dissolved in 5 mL of benzene. A
forest green color developed immediately. After it was stirred
for 3 h at 25 °C, the reaction mixture was evaporated to
dryness, and the resulting green solid was triturated with Et2O
(2 × 3 mL). The last traces of solvent were removed at reduced
pressure to obtain analytically pure 6e as a dark green powder
(359 mg, 77% yield). 1H NMR (C6D6): δ 7.02 (t, J ) 15 Hz,
2H, aryl), 7.19 (m, 3H, aryl), 7.01 (m, 4H, aryl), 6.95 (d, J ) 7
Hz, 2H, aryl), 6.92 (d, J ) 7 Hz, 2H, aryl), 5.8 (s, 10H, C5H5),
3.46 (m, 2H, CH(CH3)2), 1.21 (d, J ) 6 Hz, 6H, CH(CH3)2), 0.93
(d, J ) 6 Hz, 6H, CH(CH3)2); 13C{1H}NMR (C6D6): δ 181.8,
1
powder (101 mg, 73% yield). H NMR (C6D6): δ 7.28 (d, J ) 5
Hz, 2H, aryl), 7.09 (m, 6H, aryl), 6.86 (d, J ) 5 Hz, 2H, aryl),
6.70 (m, 11H, aryl), 6.43 (s, 5H, C5H5), 5.59 (s, 5H, C5H5), 4.24
(m, 1H, CH(CH3)2), 2.65 (m; 1H, CH(CH3)2), 2.17 (s, 3H, CH3),

1806 Organometallics, Vol. 20, No. 9, 2001
Zuckerman and Bergman
2.04 (s, 3H, CH3), 1.88 (d, J ) 5 Hz, 3H, CH(CH3)2), 1.65 (d, J
) 5 Hz, 3H, CH(CH3)2), 0.95 (d, J ) 5 Hz, 3H, CH(CH3)2),
-0.05 (d, J ) 10 Hz, 3H, CH(CH3)2). 13C{1H} NMR (C6D6): δ
159.1, 146.7, 146.1, 145.6, 145.6, 144.0, 139.9, 129.0, 128.5,
128.5, 127.4, 127.1, 126.7, 126.5, 125.0, 124.8, 124.4, 121.2,
114.8, 111.6, 65.8, 27.6, 27.6, 27.5, 25.7, 24.6, 23.3, 21,0. 21.0,
20.8, 15.5. Anal. Calcd for C51H51N3Zr: C, 76.84; H, 6.45; N,
5.27. Found: C, 76.45; H, 6.68; N, 5.03.
methylsilyl)carbodiimide were formed in 97% and 100% yields,
respectively, by integration against the internal standard 1,4-
dimethoxybenzene. Complex 8c was identified by comparison
of its NMR spectrum with those in the literature.26 1H NMR
(C6D6): δ 7.18 (d, J ) 7 Hz, 2H, aryl), 6.96 (t, J ) 15 Hz, IH,
aryl), 5.80 (s, 10H, C5H5), 4.01 (m, 2H, CH(CH3)2), 2.21 (s, 6H,
aryl CH3’s), 1.32 (d, J ) 6 Hz, 6H, CH(CH3)2), 1.18 (d, J ) 6
Hz, 6H, CH(CH3)2).
11. Metallacyclobutene complex 6a (70 mg, 0.21 mmol) and
1,3-di-p- tolylcarbodiimide (45 mg, 0.20 mmol) were dissolved
in benzene (5 mL). The solution turned from green to orange
while it was stirred at 25 °C for 1 h. After it was stirred
overnight at 25 °C, the reaction mixture was evaporated to
dryness to afford an orange solid. This solid was dissolved in
hexanes (3 mL) and stored at -35 °C for 7 days. Analytically
pure 11 was obtained as an orange powder (137 mg, 97% yield).
1H NMR (C6D6): δ 7.62 (d, J ) 7 Hz, 2H, aryl), 7.51 (d, J ) 7
Hz, 1H, aryl), 7.05 (m, 10H, aryl), 6.98 (m, 5H, aryl), 6.20 (s,
5H, C5H5), 5.54 (s, 5H, C5H5), 2.14 (s, 3H, CH3), 2.14 (s, 3H,
CH3), 0.67 (s, 9H, C(CH3)3). 13C{1H} NMR (C6D6): δ 150.3,
147.1, 141.5, 140.4, 133.3, 131.9, 131.6, 130.2, 129.4, 128.9,
128.6, 128.5, 128.5, 127.5, 127.2, 126.6, 123.8, 122.2, 112.8,
110.1, 89.6, 21.9, 21.8, 11.9, 5.7. HREIMS: m/z calcd for
Kin etic Stu d ies of th e Rea ction betw een 4a a n d
Dip h en yla cetylen e (5a ). UV-vis spectrometry was used to
study the exchange reaction between diazametallacycle com-
plex 4a and diphenylacetylene (5a ) to generate azametalla-
cyclobutene 6a and 1,3-bis(trimethylsilyl)carbodiimide (3a ). A
stock solution of 4a (7.5 × 10-4 M) was prepared by weighing
44.7 mg (0.094 mmol) into a 25.00 ( 0.01 mL volumetric flask.
Dimethylzirconocene (12.6 mg, 0.05 mmol) was added, and the
mixture was diluted to the final volume with benzene to give
a concentration of 3.75 × 10-3 M for 4a . The solution was
mixed well, and 1 mL was transferred with a volumetric pipet
to a second 5.00 ( 0.01 mL volumetric flask, followed by
dilution to the final volume with benzene. The 7.5 × 10-4
M
solution of 4a was stored in the drybox at -35 °C. A typical
3.75 × 10-2 M stock solution of 5a was prepared by weighing
diphenylacetylene (66.8 mg, 0.38 mmol) into a 10 mL volu-
metric flask and diluting to the final volume with benzene. A
stock solution of 1,3-bis(trimethylsilyl)carbodiimide (3a ; 7.5 ×
10-2 M) was prepared by weighing 69.9 mg (0.375 mmol) of
3a into a 10.00 ( 0.01 mL volumetric flask and diluting to
the final volume with benzene. A serial dilution to afford a
7.5 × 10-2 M stock solution of 3a was done by transferring 1
mL of a 3.75 × 10-2 M solution of 3a with a volumetric pipet
into a 5 mL volumetric flask and diluting to the final volume
with benzene. Both stock solutions of 3a and 5a were stored
in the drybox at -35 °C.
In a typical kinetic run, stock solutions of 4a , 3a , and 5a
were allowed to thaw in the drybox and 1 mL from a 7.5 ×
10-4 M stock solution of 4a , 1 mL from a 3.75 × 10-2 M stock
solution of 5a , and 1 mL of 7.5 × 10-2 M 3a were transferred
with volumetric pipets to a 1 mm gastight cuvette. The final
concentrations were 2.50 × 10-4 M for 4a , 1.25 × 10-2 M for
5a , and 2.50 × 10-3 for 3a . The cuvette was equipped with a
Kontes Teflon stopcock and placed in the temperature control
unit of the UV-vis spectrometer. Data were collected at 95 s
time intervals for over 8 half-lives at 15.4 ( 0.1 °C. The
absorbance of reactant azametallacyclobutene 6a (λ ) 420 nm)
was plotted vs time, and the data were analyzed with an
exponential fitting program.
C
43H43N3Zr 692.2564. found 692.2582.
12. A J . Young NMR tube was charged with Cp2Zr(N(2,6-
Me2C6H3)CdN(SiMe3)N(SiMe3)) (8a ; 10.0 mg, 0.02 mmol), 1,4-
dimethoxybenzene (5.2 mg, 0.038 mmol), diphenylacetylene
(5.0 mg, 0.028 mmol), and 0.6 mL of C6D6. A green color
developed immediately. A single-pulse 1H NMR spectrum was
acquired in which the resonances of the cyclopentadienyl
ligands of 6c were integrated against the methoxy groups of
1,4-dimethoxybenzene. 1,3-Di-p-tolylcarbodiimide (6.2 mg,
0.028 mmol) was added to the reaction mixture. After 1 h,
1
another H NMR spectrum was measured as described above;
a 90% yield of six-membered zirconacycle 12 was obtained.
1
Complex 12 was only characterized by H NMR spectroscopy
and MS analysis since its analogue, zirconacycle complex 11,
1
was fully characterized. H NMR (C6D6): δ 7.55 (d, J ) 7 Hz,
2H, aryl), 7.1 (m, 7H, aryl), 6.95 (m, 12H, aryl), 6.44 (s, 5H,
C5H5), 5.47 (s, 5H, C5H5), 2.26 (s, 3H, CH3), 2.11 (s, 3H, CH3),
2.05 (s, 6H, CH3). MS (EI): m/z calcd 739, found 739.
Rea ction s of Dia za m eta lla cycle Com p lexes 4a , 8a, a n d
8b w ith Alk yn es. In a typical reaction, a J . Young NMR tube
was charged with Cp2Zr(N(2,6-Me2C6H3)CdN(t-Bu)N(t-Bu))
(8b; 10.0 mg, 0.02 mmol), 1,4-dimethoxybenzene (5.5 mg, 0.04
1
mmol), and 0.6 mL of C6D6. A single-pulse H NMR spectrum
was acquired in which the resonances of the cyclopentadienyl
ligands of 8b were integrated against the methoxy groups of
1,4-dimethoxybenzene. Diphenylacetylene (5.0 mg, 0.03 mmol)
was added to the NMR tube at 25 °C in the glovebox, and the
solution turned from purple to green within seconds after the
addition. After 20 min, another 1H NMR spectrum was
obtained as described above; a 97% yield of azametallacy-
Kin etic Stu d ies of th e Rea ction betw een 8a a n d
Dip h en yla cetylen e (5a ). UV-vis spectrometry was used to
study the exchange reaction between diazametallacycle com-
plex 8a and diphenylacetylene to generate azametallacy-
clobutene 6c and carbodiimide 3a . A typical 0.002 M stock
solution of metallacycle 8a was prepared by weighing 24.7 mg
(0.05 mmol) 8a into a 5.00 ( 0.01 mL volumetric flask.
Dimethylzirconocene (9.5 mg, 0.04 mmol) was added, and the
mixture was diluted to the final volume with benzene. The
solution was mixed well, and 1 mL was transferred with a
volumetric pipet to a second 5.00 ( 0.01 mL volumetric flask
followed by dilution to the final volume with benzene. The
0.002 M solution of 8a was stored in the drybox at -35 °C. A
typical (0.05 M) stock solution of 5a was prepared by weighing
diphenylacetylene (44.6 mg, 0.25 mmol) into a 5 mL volumetric
flask and diluted to the final volume with benzene. A serial
dilution to afford a 0.01 M stock solution of 5a was done by
transferring 1 mL of a 0.05 M solution with a volumetric pipet
into a 5 mL volumetric flask and diluting to the final volume
with benzene. A similar procedure was used for the generation
of a 0.05 M stock solution of carbodiimide 3a . 1,3-Bis-
(trimethylsilyl)carbodiimide was weighed into a 5.00 ( 0.01
mL volumetric flask (46.6 mg, 0.25 mmol) and diluted to the
1
clobutene 6c was measured. H NMR spectrometry indicated
that 1,3-di-tert-butylcarbodiimide (δ 1.16 in C6D6) was formed
in 91% yield. The 1H NMR spectrum of 1,3-di-tert-butylcar-
bodiimide was identical with that of an authentic sample
purchased from Aldrich. The 1H NMR spectrum for 6c was
compared with literature values.9 1H NMR (THF-d8): δ 7.3 (t,
2H, aryl), 6.91 (m, 6H, aryl), 6.75 (m, 4H, aryl), 6.51 (t, 1H,
aryl), 6.12 (s, 10H, C5H5), 2.16 (s, 6H, CH3).
Rea ction s of 4a a n d 8a w ith Sym m etr ica l Ca r bod iim -
id es. The general procedure used was identical with that
described above for the reaction of 8b with diphenylacetylene
(5a ). In a typical reaction, a J . Young NMR tube was charged
with Cp2Zr(N(2,6-Me2C6H3)CdN(SiMe3)N(SiMe3)) (8a ; 10.0
mg, 0.02 mmol), 1,4-dimethoxybenzene (9.6 mg, 0.07 mmol),
and 0.6 mL of C6D6. 1,3-Diisopropylcarbodiimide (3.5 mg, 4.4
µL, 0.03 mmol) was added to the NMR tube. The reaction
mixture turned purple immediately. It was determined that
Cp2Zr(N(2,6-Me2C6H3)CdN(i-Pr)N(i-Pr)) (8c) and 1,3-bis(tri-

Reactions between Zr Metallacycles and Organic Substrates
Organometallics, Vol. 20, No. 9, 2001 1807
final volume with benzene. Both stock solutions of 3a and 5a
were stored in the drybox at -35 °C.
Me3SiNdCdNSiMe3 (3a ) relative to the concentration of
diazametallacycle complex 4a . Stock solutions with weighed
amounts of 4a , 3a , and 3e in benzene were prepared and
stored at -35 °C in the drybox prior to individual kinetic runs.
Data were collected at 60 s time intervals at 15.4 ( 0.1 °C for
over 5 half-lives, and the absorbance of diazametallacycle
complex 4f was monitored (λ 600 nm). The absorbance of 4f
was plotted vs time, and the data were analyzed with an
exponential fitting program.
X-r a y Str u ctu r e Deter m in a tion of 4a a n d 10a . Both
crystals were mounted on quartz fibers using Paratone N
hydrocarbon oil. For both compounds, measurements were
made on a SMART CCD area detector with graphite-mono-
chromated Mo KR radiation. Data were integrated by the
program SAINT. The data were corrected for Lorentz and
polarization effects and analyzed for agreement and possible
absorption using XPREP.36 An empirical absorption correction
based on comparison of redundant and equivalent reflections
as applied using SADABS37 (7d, Tmin ) 0.91, Tmin ) 0.86; 10a,
Tmin ) 0.897, Tmin ) 0.818; 10b, Tmin ) 0.97, Tmin ) 0.81). The
structures were solved by direct methods and expanded using
Fourier techniques. The non-hydrogen atoms were refined
anisotropically. Plots of ∑w(|Fo| - |Fc|)2 vs |Fo|, the reflection
order in data collection, (sin θ)/λ, and various classes of indices
showed no unusual trends. All calculations were performed
using the teXsan38 crystallographic software package of Mo-
lecular Structure Corp.
In a typical kinetic run, stock solutions of 8a , 5a , and 3a
were allowed to thaw in the drybox, and 0.5 mL from a 0.002
M stock solution of 8a and 1 mL of 0.05 M 3a were transferred
with volumetric pipets to a 1 mm gastight cuvette with a
sidearm. The cuvette was equipped with a Kontes Teflon
stopcock. An additional 0.5 mL of benzene was added via
volumetric pipet. The stopcock was sealed to the UV-vis cell,
and 1.0 mL of 0.01 M solution of 5a was transferred to the
sidearm of the cell. A rubber septum was wired onto the orifice
of the sidearm to prevent the introduction of any adventitious
oxygen or water into the system. The cuvette was removed
from the drybox, and the solution was cooled to 5.0 ( 0.1 °C
in the temperature control unit of the UV-vis spectrometer.
Once the solution in the cuvette was cooled, the stopcock was
adjusted to allow the flow of diphenylacetylene (5a ) solution
into the cell. The stopcock was tightly closed, and the cuvette
was shaken to ensure a homogeneous solution. Data acquisi-
tion began after the solution was allowed to equilibrate to the
bath temperature (ca. 60 s), and spectra were recorded at
regular intervals (10 s) for over 4 half-lives of the reaction.
The absorbance of metallacyclobutene 6c was monitored (λ )
550 nm) and plotted vs time. The data were analyzed with an
exponential fitting program.
Kin etic Stu d ies of th e Rea ction betw een 8b a n d
Dip h en yla cetylen e (5a ). A procedure similar to that used
for the kinetic analysis of the reaction between 8a and 5a was
utilized in this kinetic study. Stock solutions with weighed
amounts of 8b, 3b, and 5a in benzene were prepared and
stored at -35 °C in the drybox prior to individual kinetic runs.
Data were collected at 10 s time intervals for over 4 half-lives,
and the absorbance of product azametallacyclobutene 6c was
plotted vs time.
Kin etic Stu d ies of th e Rea ction betw een Aza m eta lla -
cyclobu ten e Com p lex 6e a n d 1,3-Diisop r op ylca r bod iim -
id e (3c). A procedure similar to that carried out for the kinetic
analysis of the reaction between diazametallacycle complex 4a
and diphenylacetylene (5a ) was utilized in this kinetic study.
Stock solutions with weighed amounts of 6a and 3c in benzene
were prepared and stored at -35 °C in the drybox prior to
individual kinetic runs. Data were collected at 20 s time
intervals for over 11 half-lives at 15.4 ( 0.1 °C. The absorbance
of reactant azametallacyclobutene 6e (λ 420 nm) was plotted
vs time, and the data were analyzed with an exponential fitting
program.
Kin etic Stu d ies of th e Rea ction betw een 4a a n d 1,3-
Di-p-tolylca r bod iim id e (3e). A procedure similar to that
carried out for the kinetic analysis of the reaction between
diazametallacycle complex 4a and diphenylacetylene (5a ) was
utilized in this kinetic study. This system was modified by
using excess concentrations of 4a and Me3SiNdCdNsiMe3 (3a )
relative to the concentration of (p-Tol)NdCdN(p-Tol) (3e).
Stock solutions with weighed amounts of 4a , 3a , and 3e in
benzene were prepared and stored at -35 °C in the drybox
prior to individual kinetic runs. Data were collected at 60 s
time intervals at 15.4 ( 0.1 °C for over 5 half-lives, and the
absorbance of diazametallacycle complex 4e was monitored (λ
420 nm) and was plotted vs time.
Kin etic Stu d ies of th e Rea ction betw een 4a a n d 1,3-
Di-m -tolylca r bod iim id e (3f). A procedure similar to that
used for the kinetic analysis of the reaction between diaza-
metallacycle complex 4a and (p-Tol)NdCdN(p-Tol) (3e) was
utilized in this kinetic study. This system was modified by
using excess concentrations of (m-Tol)NdCdN(m-Tol) (3f) and
For compound 4a , all non-hydrogen atoms were refined
anisotropically except for Si3 and C2. Hydrogen atoms were
included but not refined. Disorder was found in the structure
t
of the molecule between one SiMe3 group and the Bu group.
A partially occupied silicon atom, Si3, was placed in a fixed
idealized location along the N1-C2 bond axis 1.67 Å from N1.
The population of the majority silicon, Si1, was allowed to vary,
and the populations of Si3 and C2 were tied to that of Si1; the
occupancy of Si3 was constrained to be 1 - occ(Si1), and the
occupancy of C2 was constrained to equal the occupancy of
Si1. This model of the disorder resulted in percent occupancies
of Si in the two sites, Si1 and Si3, of 84% and 16%, respectively.
Crystallographic data for compounds 4a and 10a are listed
in Table 2, and selected bond lengths and bond angles are
given in Table 1 for 4a and Table 5 for 10a ; full details are
given in the Supporting Information.
Ack n ow led gm en t. We thank Drs. Frederick Hol-
lander and Patricia Goodson for solving the crystal
structures of 4a and 10a , the National Institutes of
Health (Grant No. GM-25459) for financial support of
this work, and Prof. Kristopher McNeill (University of
Minnesota) and Dr. Mark Aubart (Ato-Fina) for fruitful
discussions. The Center for New Directions in Organic
Synthesis is supported by Bristol-Myers Squibb as
Sponsoring Member.
Su p p or tin g In for m a tion Ava ila ble: Tables of concentra-
tions used in the kinetic runs, one kinetic plot, and crystal-
lographic data for 4a and 10a . This material is available free
of charge via the Internet at http://pubs.acs.org.
OM010091O
(36) XPREP, version 5.03 (part of the SHELXTL Crystal Structure
Determination Package); Siemens Industrial Automation, Inc., 1995.
(37) Sheldrick, G. SADABS: Siemens Area Detector Absorption
correction program; Siemens Industrial Automation, Inc., 1996.
(38) teXsan: Crystal Structure Analysis Package; Molecular Struc-
ture Corp., 1985 and 1992.