Mechanism of the reaction of alkynes with a "constrained geometry" zirconaaziridine. PMe3 dissociates more rapidly from the constrained geometry complex than from its Cp2 analogue

Article abstract of DOI:10.1021/om801009p

The "constrained geometry" (cg) zirconaaziridines Me ₄C₅SiMe₂N(tBu)Zr-(n²)-[N(Ph)CH(Ph)] (PMe2R) (R = Me, Ph) have been synthesized, and the R = Ph derivative has been structurally characterized by X-ray crystal

Full text of DOI:10.1021/om801009p

Organometallics 2009, 28, 493–498
493
Mechanism of the Reaction of Alkynes with a “Constrained
Geometry” Zirconaaziridine. PMe3 Dissociates More Rapidly from
the Constrained Geometry Complex than from its Cp2 Analogue
Kathleen E. Kristian, Masanori Iimura, Sarah A. Cummings, Jack R. Norton,*
Kevin E. Janak, and Keliang Pang
Department of Chemistry, Columbia UniVersity, New York, New York 10027
ReceiVed October 20, 2008
The “constrained geometry” (cg) zirconaaziridines Me4C5SiMe2N(tBu)Zr-(η2)-[N(Ph)CH(Ph)](PMe2R)
(R ) Me, Ph) have been synthesized, and the R ) Ph derivative has been structurally characterized by
X-ray crystallography. Treatment of these zirconaaziridines with unsaturated electrophiles such as
diphenylacetylene results in insertion. Kinetic data for these irreversible reactions indicate that PMe3
dissociation must occur prior to insertion and that PMe3 dissociation from the cg zirconaaziridine 4a is
faster (k1 ) 0.204(7) s-1) than from the Cp2 zirconaaziridine 5a (k1′ ) 0.0013(1) s-1). The reaction of
4a with diphenylacetylene is several orders of magnitude faster than the reaction of 5a with
diphenylacetylene.
Chart 1
Introduction
In addition to the many developments in metallocene catalysis
for olefin polymerization over the past 20 years,1 there has been
great interest in nonmetallocene-mediated olefin polymeriza-
tion.2 Nonmetallocene catalysts can offer increased thermal
stability and increased regio- and stereoselectivity. Since Bercaw
and co-workers first reported the dimeric “constrained geometry”
scandium complex [[(η5-C5Me4)Me2Si(η1-NCMe4)](PMe3)ScH]2
in 1990,3,4 there has been interest in complexes of the
“constrained geometry” (cg) ligand, Me4C5SiMe2N(tBu)2-, with
Group 45,6 and other early transition metals (Group 3,7,8 Group
5,9 Ln and An10). cgTi and cgZr complexes are highly active
olefin polymerization catalysts that are particularly good for the
copolymerization of ethylene and R-olefins.
Although the polymerization activity of constrained geometry
Ti and Zr complexes has been studied widely, the origin of the
enhanced reactivity (“constrained geometry effect”) remains
unclear. Breitling has suggested that the high activity “most
probably results from a more Lewis-acidic transition metal
centre”;5 Bercaw has characterized cgM complexes as “more
acidic and even more electron deficient [than Cp2M]”.8 The
enhanced Lewis acidity has been attributed to both electronic
factors and steric effects, such as the less crowded coordination
sphere and smaller Cpcentroid-M-N bite angle in a cgM complex
(vs those found in Cp2M complexes with M ) Ti, Zr).11
* Corresponding author. E-mail: jrn11@columbia.edu.
(1) Alt, H. G.; Koppl, A. Chem. ReV. 2000, 100, 1205–1222. Anger-
mund, K.; Fink, G.; Jensen, V. R.; Kleinschmidt, R. Chem. ReV. 2000, 100,
1457–1470. Coates, G. W. Chem. ReV. 2000, 100, 1223–1252. Fink, G.;
Steinmetz, B.; Zechlin, J.; Przybyla, C.; Tesche, B. Chem. ReV. 2000, 100,
1377–1390. Resconi, L.; Cavallo, L.; Fait, A.; Piemontesi, F. Chem. ReV.
2000, 100, 1253–1346.
(2) Boffa, L. S.; Novak, B. M. Chem. ReV. 2000, 100, 1479–1494.
Butenschon, H. Chem. ReV. 2000, 100, 1527–1564. Gibson, V. C.;
Spitzmesser, S. K. Chem. ReV. 2003, 103, 283–316. Hlatky, G. G. Chem.
ReV. 2000, 100, 1347–1376. Ittel, S. D.; Johnson, L. K.; Brookhart, M.
Chem. ReV. 2000, 100, 1169–1204. Siemeling, U. Chem. ReV. 2000, 100,
1495–1526.
(3) Shapiro, P. J.; Bunel, E.; Schaefer, W. P.; Bercaw, J. E. Organo-
metallics 1990, 9, 867–869.
(4) Shapiro, P. J.; Schaefer, W. P.; Labinger, J. A.; Bercaw, J. E.; Cotter,
W. D. J. Am. Chem. Soc. 1994, 116, 4623–4640.
(5) Braunschweig, H.; Breitling, F. M. Coord. Chem. ReV. 2006, 250,
2691–2720.
(6) Brintzinger, H. H.; Fischer, D.; Mulhaupt, R.; Rieger, B.; Waymouth,
R. M. Angew. Chem., Int. Ed. Engl. 1995, 34, 1143–1170. Cano, J.; Kunz,
K. J. Organomet. Chem. 2007, 692, 4411–4423. Lian, B.; Thomas, C. M.;
Navarro, C.; Carpentier, J. F. Organometallics 2007, 26, 187–195. Lian,
B.; Thomas, C. M.; Navarro, C.; Carpentier, J. F. Macromolecules 2007,
40, 2293–2294.
(7) Gromada, J.; Carpentier, J.-F.; Mortreux, A. Coord. Chem. ReV. 2004,
248, 397–410. Trifonov, A. A.; Spaniol, T. P.; Okuda, J. Organometallics
2001, 20, 4869–4874. Trifonov, A. A.; Spaniol, T. P.; Okuda, J. Dalton
Trans. 2004, 2245–2250.
However, quantitative data explaining the enhanced reactivity
are scarce. Bercaw and co-workers noted that PMe3 dissociates
readily from the original cgSc dimer,8 but did not measure the
(9) Alcalde, M. I.; Gomez-Sal, M. P.; Royo, P. Organometallics 1999,
18, 546–554. Alcalde, M. I.; Gomez-Sal, M. P.; Royo, P. Organometallics
2001, 20, 4623–4631. Postigo, L.; Sanchez-Nieves, J.; Royo, P. Inorg. Chim.
Acta 2007, 360, 1305–1309.
(10) Arredondo, V. M.; McDonald, F. E.; Marks, T. J. Organometallics
1999, 18, 1949–1960. Li, H. B.; Marks, T. J. Proc. Natl. Acad. Sci. U.S.A.
2006, 103, 15295–15302. Seyam, A. M.; Stubbert, B. D.; Jensen, T. R.;
O’Donnell, J. J.; Stern, C. L.; Marks, T. J. Inorg. Chim. Acta 2004, 357,
4029–4035. Stubbert, B. D.; Marks, T. J. J. Am. Chem. Soc. 2007, 129,
4253–4271. Stubbert, B. D.; Marks, T. J. J. Am. Chem. Soc. 2007, 129,
6149–6167. Stubbert, B. D.; Stern, C. L.; Marks, T. J. Organometallics
2003, 22, 4836–4838.
(8) Shapiro, P. J.; Bunel, E.; Schaefer, W. P.; Bercaw, J. E. Organo-
metallics 1990, 9, 867–869.
(11) Stevens, J. C. Stud. Surf. Sci. Catal. 1996, 101, 11–20.
10.1021/om801009p CCC: $40.75
2009 American Chemical Society
Publication on Web 12/19/2008

494 Organometallics, Vol. 28, No. 2, 2009
Kristian et al.
Scheme 1
rate. Jordan12 reported activation barriers for the dissociation
of pendant olefins in cgTi complexes, finding them to be larger
than the barrier to dissociation in a Cp2Zr complex or a rac-
(EBI)Zr complex (Chart 1). Petersen13 compared the solid-state
structures of cgTiCl2 and Cp*(NiPr2)TiCl2 and concluded that
the cg ligand is a poorer π donor than separate Cp and amido
fragments.
Cp2Zr(Me)(OTf) with a lithiated amine, forming a transient
methyl amide complex that spontaneously loses methane to form
a zirconaaziridine (eq 1).16,17 CH4 elimination appears to occur
via σ-bond metathesis, requiring a syn orientation of the Zr-Me
bond and the C-H bond in the benzylamido ligand.
Comparisons of tethered and untethered ligand systems can
lend insight into both steric and electronic effects, but directly
analogous systems containing identical amido substituents
should be employed. Erker14 utilized a constrained geometry
“CpC1O”Zr dimer for olefin polymerization and found that the
dimer, not the less sterically encumbered monomer, was the
active catalyst. This suggests that steric effects are not respon-
sible for the enhanced reactivity of cg complexes relative to
the corresponding metallocenes.
Kinetic data for substitution reactions of cg and Cp2 systems
should give a clearer picture of the origin of the “constrained
geometry effect”. Thus, when our long-standing interest in
zirconaaziridine chemistry15 led us to prepare a cg analogue to
Cp2 zirconaaziridines, we decided to compare the substitution
kinetics of the cg and Cp2 systems. Herein we report the
synthesis of that “constrained geometry” zirconaaziridine, its
reactions with unsaturated electrophiles, and comparative kinetic
data for the substitution reactions of the cg and Cp2 complexes.
We first attempted the synthesis of a cg zirconaaziridine by
this method, from the known complex cgZrMe2.18 Unfortu-
nately, TfOH protonated the amido portion of the cg ligand and
led to decomposition (Scheme 1). Treatment of cgZrMe2 with
1 equiv of TMSOTf, however, gave cgZr(Me)(OTf) (2) in
excellent yield.
X-ray analysis of cgZr(Me)(OTf) (2) (Figure 1) showed that
its crystals contain a triflate-bridged dimer. At room temperature
the 1H NMR of 2 showed significant broadening of two methyl
resonances, the Zr-Me and one of the four Cp methyl groups,
suggesting a monomer-dimer equilibrium in solution.
Treatment of 2 with LiN(Ph)C(H)(Ph) yielded 3, cgZr(Me)-
(N(Ph)CH2Ph). However, unlike its Cp2 analogue, 3 did not
spontaneously lose CH4 at room temperature to yield a zircon-
aaziridine. An NOE study of the cg complex 3 explained this
lack of reactivity by revealing the nonplanar conformation
shown in Figure 2. A NOE was detected between one of the
methylene protons on the benzylamido ligand and the remaining
Results and Discussion
Synthesis and Characterization of Constrained Geometry
Complexes. Metallocene-derived zirconaaziridines Cp2Zr-
(η2-N(R)CH(R′))L15 are available from the reaction of
(12) Carpentier, J. F.; Maryin, V. P.; Luci, J.; Jordan, R. F. J. Am. Chem.
Soc. 2001, 123, 898–909.
(13) Carpenetti, D. W.; Kloppenburg, L.; Kupec, J. T.; Petersen, J. L.
Organometallics 1996, 15, 1572–1581. Kloppenburg, L.; Petersen, J. L.
Organometallics 1997, 16, 3548–3556.
(14) Kunz, K.; Erker, G.; Kehr, G.; Frohlich, R.; Jacobsen, H.; Berke,
H.; Blacque, O. J. Am. Chem. Soc. 2002, 124, 3316–3326.
(15) Cummings, S. A.; Tunge, J. A.; Norton, J. R. Synthesis and
Reactivity of Zirconaaziridines. In New Aspects of Zirconium Containing
Organic Compounds; Marek, I., Ed.; Springer: New York, 2005; pp
1-39.
(16) Buchwald, S. L.; Watson, B. T.; Wannamaker, M. W.; Dewan, J. C.
J. Am. Chem. Soc. 1989, 111, 4486–4494.
(17) Gately, D. A.; Norton, J. R.; Goodson, P. A. J. Am. Chem. Soc.
1995, 117, 986–996.
(18) Canich, J. M. EP 420436-A1, 1991. Stevens, J. C.; Timmers, F. J.;
Wilson, D. R.; Schmidt, G. F.; Nickias, P. N.; Rosen, R. K.; Knight, G. W.;
Lai, S. Y. EP 416815-A2, 1991.
Figure 1. Molecular structure of 2 (20% probability level). Select
hydrogen atoms are omitted for clarity.

“Constrained Geometry” Zirconaaziridine
Organometallics, Vol. 28, No. 2, 2009 495
Figure 2. Conformation of 3 supported by NOE experiments. A
NOE was observed between the indicated protons.
Figure 5. Molecular structure of 6 (20% probability level). Select
hydrogen atoms are omitted for clarity. Selected bond lengths (Å):
Zr-N2 2.097(2), Zr-C(50) 2.263(3), Zr-N1 2.080(2), C(50)-C(40)
1.338(4).
Scheme 2
Figure 3. Molecular structure of 4b (20% probability level). Select
hydrogen atoms are omitted for clarity. Selected bond lengths (Å)
and angles (deg): Zr-N2 2.056(3), Zr-C31 2.321(5), Zr-N1
2.128(3), Zr-P 2.815(1), C-N2-C31 123.3(4), C-N2-Zr 150.6(3),
C31-N2-Zr 81.3(2), N2-Zr-C31 37.5(1), N2-C31-Zr 61.1(2).
the methylene protons in 3 disappeared, with concurrent growth
of a 1H singlet at δ 3.10, which is assigned to the methine proton
in 4. In the 31P{1H} NMR a peak at δ -27 appeared that can
be assigned to coordinated PMe3. The zirconaaziridine could
not be trapped by weaker donors such as THF or pyridine. The
1H NMR of 4 indicated the formation of the stereoisomer shown
in Scheme 1, along with its enantiomer. A NOE was observed
between the methine H and one of the Cp methyl groups but
not between this H and the tBu group, indicating that the
methine H is oriented toward the Cp.
Single crystals of 4b were obtained from slow diffusion of
1
hexanes into a benzene solution. Their H NMR was identical
to that of 4b formed in solution. X-ray analysis showed the
structure in Figure 3.
Figure 4. Molecular structure of 5b (20% probability level). Select
hydrogen atoms are omitted for clarity. Selected bond lengths (Å)
and angles (deg): Zr-N 2.100(1), Zr-C30 2.377(1), Zr-P 2.7227(3),
C-N-C30 124.0(1), C-N-Zr 144.10(8), C30-N-Zr 82.65(7),
N-Zr-C30 36.16(4), N-C30-Zr 61.19(6).
methyl group on Zr, and between the other methylene proton
and the tBu group in the cg ligand (Figure 2).
When heated at 60 °C for 2-3 h a benzene solution of 3 and
excess PMe2R (R ) Me, Ph) changed from yellow to bright
orange. The 1H and 31P{1H} NMR spectra showed the formation
of the zirconaazirdines cgZr(η2-N(Ph)CH(Ph))PMe2R (4, Scheme
Figure 6. Plot of kobs vs [PMe3]-1 for the reaction of 4a (0.02 M)
with 10 equiv of PhCCPh at 50 °C in C6D6.
1
1). In the H NMR the AB quartet (δ 4.25) corresponding to

496 Organometallics, Vol. 28, No. 2, 2009
Kristian et al.
upfield shift for the ring methine. By analogy with previously
characterized carbodiimide insertion products,20 we believe these
products to be the azazirconacyclopentanes resulting from
insertion into the Zr-C and Zr-N bonds (the Zr-C insertion
product is shown as 7 in eq 3b).
Figure 7. Plot of kobs vs [PhCCPh] for the reaction of 4a (0.02 M)
in the presence of 10 equiv of PMe3 at 50 °C in C6D6.
In order to quantify the rate of insertion into a PMe3-ligated
Cp2 zirconaaziridine, we treated 5a with 10 equiv of dipheny-
lacetylene at 50 °C. It was indeed slow, with t1/2 ≈ 30 h in
C6D6 and >96 h21 in THF-d8 (eq 4), more than 60 times slower
than the reaction of 4a with diphenylacetylene.
For comparison with 4b, we synthesized Cp2Zr(η2-N(Ph)-
CH(Ph))PMe2R (5a, R ) Me; 5b, R ) Ph) by a modification17,19
of Buchwald’s procedure.16 The X-ray structure of 5b is shown
in Figure 4. The bond angles within the three-membered ring
in the cg zirconaaziridine 4b are comparable to those in 5b as
well as to those in other Cp2 zirconaaziridines.15 The Zr-P bond
length in the Cp2 complex 5b is 2.7227(3) Å, considerably
shorter than the Zr-P bond in the cg complex 4b (2.816(1) Å).
However, the Zr-N and Zr-C bonds within the zirconaaziridine
ring are slightly longer in 5b (Zr-N ) 2.100(1) Å, Zr-C )
2.377(1) Å) than in 4b (Zr-N ) 2.056(3) Å, Zr-C ) 2.321(5)
Å). Thus the η2 imine is slightly closer to the Zr in the cg
complex 4b, while the PMe3 is considerably closer to the Zr in
the Cp2 complex 5b.
Reactivity of the cg Zirconaaziridine 4a and Its Cp2
Analogue 5a. Heating 3 at 60 °C in benzene with dipheny-
lacetylene gave 6, an azazirconacyclopentene resulting from
insertion into the Zr-C bond (eq 2). X-ray analysis of a crystal
of 6 confirmed this structure (Figure 5). Compound 6 was also
obtained (along with PMe3) when the PMe3 adduct 4a was
treated with diphenylacetylene. Heating 6 at 60 °C with excess
PMe3 gave no reaction even after 24 h, indicating that the
reaction of 4a with PhCt CPh is irreversible.
Mechanisms of the Insertion Reactions. The great disparity
between the reaction rates of the Cp2 complexes 5 and the cg
complexes 4 led us to consider the mechanisms by which they
occur. Previous work had shown that the insertion of carbodi-
imides into THF-ligated Cp2 zirconaaziridines required dis-
sociation of the THF and suggested that the same mechanism
was operative for the insertion of cyclic carbonates and
isocyanates.17,20 We therefore expected that insertion into the
Zr-C bond of the Cp2 zirconaaziridine 5a would require
dissociation of the PMe3, though we expected PMe3 dissociation
to be slower than THF dissociation. Indeed, mixing 5a and
diphenylacetylene in the presence of 10 equiv of PMe3 resulted
in no reaction even after 36 h at 50 °C, confirming our
hypothesis that PMe3 dissociation must occur prior to insertion.
The cg zirconaaziridine 4a is electronically unsaturated (even
with a coordinated PMe3), and an associative mechanism would
seem possible for its insertion reactions. However, the addition
of excess PMe3 to reaction 3a decreased the rate (t1/2 increased
to approximately 3 h), implying a dissociative mechanism for
that insertion.
Such a mechanism (Scheme 2) will obey eq 5 or 6 if either
of two different simplifying assumptions is valid.20 If the
equilibrium between 4a and D + PMe3 is rapidly maintained,
i.e., if k1 and k-1[PMe3] are . k2[PhCCPh], the overall rate
law will be eq 5. If instead we apply the steady-state ap-
proximation to reactive intermediate D, i.e., if k-1[PMe3] +
k2[PhCCPh] is . k1, the rate law is given by eq 6.
Cp2 zirconaaziridines form very stable complexes (e.g., 5)
with PMe3. Their insertion reactions are slower than those of
THF adducts because the dissociation of the PMe3 is slower
than that of THF.15,16 Given that “constrained geometry”
complexes were purported to be more electron deficient and
Lewis acidic than Cp2 complexes, we suspected that 4a might
dissociate PMe3 even more slowly than 5a and that 4a would
therefore insert electrophiles very slowly. However, the reaction
of 10 equiv of diphenylacetylene with 4a proceeded rapidly at
50 °C and even at room temperature to form the insertion
product 6 and free PMe3 (eq 3a).
K1k2[PhCCPh][4a]
[PMe3] + K1
d[6]
dt
)
(5)
(6)
k1k2[PhCCPh][4a]
d[6]
)
dt
k-1[PMe3] + k2[PhCCPh]
The reaction of 4a with 10 equiv of the carbodiimide
(Me3Si)NCN(SiMe3) at room temperature proceeded even more
rapidly. Two products were formed, each of which showed an
Determining the order in PMe3 was complicated by the fact
that, as reaction 3a proceeds, 1 equiv of PMe3 is produced and
(20) Tunge, J. A.; Czerwinski, C. J.; Gately, D. A.; Norton, J. R.
Organometallics 2001, 20, 254–260.
(21) At 50 °C, decomposition of 5a began after about 96 h.
(19) Tunge, J. A.; Gately, D. A.; Norton, J. R. J. Am. Chem. Soc. 1999,
121, 4520–4521.

“Constrained Geometry” Zirconaaziridine
Organometallics, Vol. 28, No. 2, 2009 497
Figure 8. (a) Plot of peak intensity at 1.7 ppm (methyl group in cg ligand of 4a) vs time (s) for the reaction of 4a (0.02 M) with 0.202 M
PhCCPh in the presence of 10 equiv (0.0404 M) of added PMe3 at 50 °C. The data are fit with a first-order equation. (b) Plot of the
experimental data for the reaction of 4a (0.02 M) with 0.202 M PhCCPh in the presence of 2 equiv (0.0404 M) of added PMe3 at 50 °C,
and the best fit curve from a Kintecus simulation.
Figure 9. Left-hand side of eq 8 vs time for (a) the reaction of 4a (0.02 M) with 0.213 M PhCCPh in the presence of 2 equiv of added
PMe3 at 50 °C and (b) the reaction of 4a (0.02 M) with 0.213 M PhCCPh in the presence of 5 equiv of added PMe3 at 50 °C.
[PMe3] therefore varies with time. The reaction of 4a with
excess PhCCPh without added PMe3 yielded kinetic traces that
deviated from first-order behavior, while with large amounts
of added PMe3 the reaction was too slow for kinetic measure-
ments to be practical. With modest amounts (2-10 equiv) of
added PMe3 we were able to follow the progress of the reaction
and determined approximate first-order rate constants kobs for
the conversion of 4a to 6. A plot of kobs as a function of 1/[PMe3]
(Figure 6, [4a] ) 0.02 M) proved to be linear over the
concentration range 0.040-0.222 M.
These measurements were made in the presence of excess (10
equiv) PMe3 to render deviations from first-order behavior
negligible.
The linearity of the plots in Figures 6 and 7 suggests the rate
of conversion of 4a f 6 is proportional to [PhCCPh]/[PMe3].
A rate law with this form (the “equilibrium steady-state”
approximation,22 eq 7) can be obtained from either eq 5 or eq
1
6. Because we cannot observe D by H NMR in a solution of
4a, K1 must be j10-5 M, and during a kinetics experiment
[PMe3] . K1 and eq 5 reduces to eq 7. Because Figure 7 is
linear, with no evidence for saturation behavior, k2[PhCCPh]
must be , k-1[PMe3], and eq 6 reduces to eq 7.
We also found kobs to be linear in [PhCCPh] over the
concentration range 0.213-0.823 M (Figure 7, [4a] ) 0.02 M).
K1k2[PhCCPh][4a]
d[6]
dt
)
(7)
[PMe3]
(22) Pyun, C. W. J. Chem. Educ. 1971, 48, 194–196.
(23) Ianni, J. C. Kintecus, Windows Version 3.95, 2008; www.kintecus.
com.
(24) We are unaware of any previous report of the integration of such
a rate law, with a concentration in the denominator increasing during the
reaction.
The validity of the rate law in eq 7 for eq 3a was tested with
the kinetic simulation program Kintecus,23 which calculated the
variation of [PMe3] during the reaction. We set k1 and k-1 to
give a reasonable estimate of K1 and kept k1 and k-1 fixed as
we refined k2; refinements converged to a constant K1k2 over a
range of these k1 and k-1 estimates (details are given in the
Supporting Information). Iteration on data from the reaction of
4a with excess PhCCPh and 2 equiv of added PMe3 gave K1k2
as 6.4(1) × 10-5 s-1 (Figure 8b), which is indistinguishable
(25) The K1 equilibrium in Scheme 2 lies far to the left, i.e., k1 . k-1
.
We cannot observe D directly, and only the product K1k2 is available from
kobs, or from fitting time-dependent data to eq 7 or eq 8. In theory it should
be possible to obtain K1 and k2 individually from a data set where [PMe3]
varies a great deal (i.e., the reaction of 4a with excess PhCCPh when no
PMe3 is added), but we have been unable to refine these parameters
separately with Kintecus.

498 Organometallics, Vol. 28, No. 2, 2009
Kristian et al.
from the value of K1k2 (6.47(6) × 10-5 s-1) obtained with a
pseudo-first-order treatment of data from the analogous reaction
with 10 equiv of added PMe3 (Figure 8a). Thus eq 7 adequately
describes the kinetics of reaction 3a.
We also ran Kintecus simulations (see Supporting Informa-
tion) for reaction 3a with 5 equiv and 10 equiv of added PMe3.
The K1k2 values thus obtained (6.8(1) × 10-5 and 6.5(1) × 10-5
s-1 for 5 and 10 equiv, respectively) are in good agreement
with both the value from the simulation of the 2 equiv case
and the value from pseudo-first-order treatment of the 10 equiv
case. The average of the three values of K1k2 from Kintecus
A 1H EXSY spectrum of the cg complex 4a with added PMe3
showed exchange cross-peaks for free and coordinated PMe3,
indicating that k1 could be measured by a selective population
inversion (SPI) experiment.26 Due to the complexity of the 1H
NMR spectrum, we carried out the SPI experiment using
31P{1H} NMR.27 Figures S-1 through S-4 in the Supporting
Information display peak intensity vs mixing time data from
these experiments. The rate constant for phosphine dissociation
simulations is 6.6(2) × 10-5 s-1
.
If we assume that [PMe3]t ) [PMe3]0 + [6], we can integrate
eq 7 to eq 8 (the derivation is given in the Supporting
Information).24 The function on the left-hand side of eq 8 is
indeed linear with time for our data (Figure 9), offering
additional evidence that reaction 3a obeys the rate law in eq 7.
However, there is only fair agreement (a discrepancy of 27%)
between the values of K1k2 obtained from the slope of such plots
and the values obtained from either (1) iteration with Kintecus
or (2) a first-order treatment of the data from a run with 10
equiv of PMe3. The discrepancies probably arise from the fact
that [PMe3]t is not exactly equal to [PMe3]0 + [6] (details are
given in the Supporting Information).
1
from 4a (k1) proved to be 0.204(7) s-1 at 298 K. A H EXSY
spectrum of the Cp2 complex 5a with added PMe3 showed that
the rate constant k1′ is too small to be quantified by SPI at room
temperature. Therefore, we measured k1′ at several temperatures
over the range 269.9-287.7 K by monitoring the exchange of
PMe3 and PMe3-d9 described in eq 10. Eyring analysis (Figure
S-5) determined k1′ to be 0.0013(1) s-1 at 298 K.
PMe3 dissociation from the cg zirconaaziridine 4a is therefore
faster than from the Cp2 zirconaaziridine 5a, a result consistent
with the much longer Zr-P distance in crystalline 4a. While
the relative dissociation rates may not be the same for other
ligands (e.g., olefins), our results call into question the idea3,5,11
that constrained geometry complexes are better Lewis acids than
their Cp2 counterparts.
[4a]0
-[6] + ([PMe3]0 + [4a]0)log
) K1k2[PhCCPh]0t
[4a]0 - [6]
(8)
PMe3 Dissociation Rate Constants. Although we know the
product K1k2 for the cg system, we cannot usefully compare
that system with the Cp2 system without determining an
individual equilibrium or rate constant for both.25 We have
therefore measured k1, the rate constant for PMe3 dissociation
from 4a, and the analogous rate constant k1′ for 5a (eqs 9 and
10). These rate constants enable us to quantitatively compare
directly analogous cg and Cp2 complexes and may therefore
lend insight into the origin of the differences in these systems.
The phosphine exchange equilibria described in eq 9 were
established before an NMR spectrum could be obtained, but
no broadening was observed in the room-temperature 1H NMR
spectra of 4a, indicating k1 to be on the order of 10-2 to 10-1
s-1. The analogous equilibria for 5a (eq 10) showed the same
behavior in both C6D6 and THF-d8.
Acknowledgment. This work was supported by the
National Science Foundation (CHE-0749537). K.E.K. grate-
fully acknowledges a fellowship from the NSF Graduate
Teaching Fellows in K-12 Education (GK-12) Program. The
authors are grateful to G. Parkin for assistance with the X-ray
structure determinations and to the National Science Founda-
tion (CHE-0619638) for the acquisition of an X-ray diffrac-
tometer; to A. D. Bain and R. A. Petros for assistance with
CIFIT; to J. Espenson, A. Bakac, and A. Haim for consulta-
tion on the kinetics; to Boulder Scientific for gifts of
Cp2ZrCl2 and Me4C5(H)SiMe2NH(tBu); and to A. Vosk-
oboynikov for a gift of Me4C5SiMe2NH(tBu)ZrCl2.
Supporting Information Available: Synthetic procedures and
details of the SPI, EXSY, simulations, and other kinetic experi-
ments. This information is available free of charge via the Internet
at http://pubs.acs.org.
(26) Bain, A. D. CIFIT; McMaster University, 2003. Bain, A. D.; Duns,
G. J.; Rathgeb, F.; Vanderkloet, J. J. Phys. Chem. 1995, 99, 17338–17343.
(27) Fischer, A.; Wendt, O. F. J. Chem. Soc., Dalton Trans. 2001, 1266–
1268.
OM801009P