Catalytic double-bond metathesis without the transition metal

Article abstract of DOI:10.1021/ja020494v

Iminophosphoranes of the type X₃P=NR (X = CI, pyrrolyl; R = alkyl, aryl) catalytically metathesize C=N bonds of carbodiimides via an addition/elimination mechanism that, despite the lack of d orbital participation in P-N bonding, conserves the key features of metal-catalyzed olefin metathesis. Diazaphosphetidine intermediates, produced by the formal [2 + 2] addition of carbodiimides to the P=N bond, have been isolated and characterized. All phosphorus-containing species in the complex catalytic reaction mixtures have been identified and their origins explained. The kinetics of addition of diisopropylcarbodiimide to Cl₃P=NPrⁱ and subsequent elimination were studied, and rate constants were determined: k_add 1.7 × 10^-3 (±0.1 × 10^-3) M s^-1 and k_elim = 4.0 × 10^-4 (±0.3 × 10^-4) s^-1. The rate of these reactions corresponds well with the observed catalytic TOF of 1.44 TO/P/h.

Full text of DOI:10.1021/ja020494v

Published on Web 08/17/2002
Catalytic Double-Bond Metathesis without the Transition
Metal
Stephen A. Bell, Tara Y. Meyer,* and Steven J. Geib
Contribution from the Department of Chemistry, UniVersity of Pittsburgh,
Pittsburgh, PennsylVania 15260
Received April 5, 2002
Abstract: Iminophosphoranes of the type X3PdNR (X ) Cl, pyrrolyl; R ) alkyl, aryl) catalytically metathesize
CdN bonds of carbodiimides via an addition/elimination mechanism that, despite the lack of d orbital
participation in P-N bonding, conserves the key features of metal-catalyzed olefin metathesis. Diaza-
phosphetidine intermediates, produced by the formal [2 + 2] addition of carbodiimides to the PdN bond,
have been isolated and characterized. All phosphorus-containing species in the complex catalytic reaction
mixtures have been identified and their origins explained. The kinetics of addition of diisopropylcarbodiimide
to Cl3PdNPri and subsequent elimination were studied, and rate constants were determined: kadd ) 1.7 ×
10-3 ((0.1 × 10-3) M s-1 and kelim ) 4.0 × 10-4 ((0.3 × 10-4) s-1. The rate of these reactions corresponds
well with the observed catalytic TOF of 1.44 TO/P/h.
Introduction
Chauvin-type mechanism that would allow us to exploit the
reaction for materials synthesis via ring-opening metathesis
Can a main group element do the job of a transition metal?
There are in certain organic transformations, such as hydro-
formylation, sp2-carbon coupling, and catalytic double bond
metathesis, which conventional wisdom would say are practical
only in the presence of elements that bear energetically
accessible d-orbitals. Although in some cases one might find
main group elements that accomplish the same overall reaction,
they generally do so by acting as simple acids or bases, rather
than following the same overall pathway as the transition metal
catalyst they replace. It is tempting to assume that the main
group versions will lack the specificity, activity, and potential
for stereocontrol of their transition metal counterparts. There
is, however, increasing evidence that certain main group systems
are capable of activating substrates by pathways that, heretofore,
were thought only accessible to transition metals. In particular,
it is interesting to consider the recent reports of Ziegler-Natta-
type polymerizations initiated by aluminum alkyls1 and stereo-
selective ring-opening polymerizations catalyzed by single-site
magnesium and aluminum alkoxide catalysts.2
polymerization (ROMP) and small molecule synthesis via ring-
closing metathesis (RCM). In the course of studying the
mechanisms of our own systems, however, the similarities
between metathesis and the classic Wittig reaction became
increasingly difficult to discount, despite our bias in favor of
transition metals (Figure 1). In particular, we were struck by
the fact that both reactions are proposed to proceed through a
four-membered ring intermediate, the formal addition product
of the unsaturated substrate and the catalyst.10,11 These similari-
ties prompted us to explore the potential of iminophosphoranes
as CdN metathesis catalysts.
The major difference between a main group imide and a
transition metal imide that would be expected to impact
mechanism is the relative importance of d orbitals in bonding.
(3) (a) Grubbs, R. H.; Tumas, W. Science 1989, 243, 907-915. (b) Kim, S.-
H.; Bowden, N.; Grubbs, R. H. J. Am. Chem. Soc. 1994, 116, 10801-
10802. (c) Fu, G. C.; Grubbs, R. H. J. Am. Chem. Soc. 1992, 114, 5426-
5427. (d) La, D. S.; Sattely, E. S.; Ford, J. G.; Schrock, R. R.; Hoveyda,
A. H. J. Am. Chem. Soc. 2001, 123, 7767-7778. (e) Nomura, G.; Takahashi,
S.; Imanishi, Y. Macromolecules 2001, 34, 4712-4723. (f) Buchmeiser,
M. R. Chem. ReV. 2000, 100, 1565-1604. (g) Phillips, A. J.; Abell, A. D.
Aldrichimica Acta 1999, 32, 75-89. (h) Schrock, R. R. In Ring-Opening
Polymerization; Brunelle, D. J., Ed.; Hanser: Munich, 1993; pp 129-156.
(4) (a) Schrock, R. R. Acc. Chem. Res. 1990, 23, 158-165. (b) Trnka, T. M.;
Grubbs, R. H. Acc. Chem. Res. 2001, 34, 18-29.
We have been interested in the catalytic metathesis of CdN
bonds. Not surprisingly, given the precedent of olefin metath-
esis,3,4 early efforts, both in our group5 and others,6-8 focused
on transition metal-based catalysts. Although CdN metathesis
is possible in the presence of simple acids,9 we hypothesized
that a transition metal would be required to promote a controlled
(5) (a) Cantrell, G. K.; Meyer, T. Y. Organometallics 1997, 16, 5381-5383.
(b) Cantrell, G. K.; Meyer, T. Y. J. Am. Chem. Soc. 1998, 120, 8035-
8042. (c) Cantrell, G. K.; Meyer, T. Y. J. Chem. Soc., Chem. Commun.
1997, 1551-1552.
(6) (a) Birdwhistell, K. R.; Lanza, J.; Pasos, J. J. Organomet. Chem. 1999,
584, 200-205. (b) Meisel, I.; Hertel, G.; Weiss K. J. Mol. Catal. 1986,
36, 159-166. (c) Zuckerman, R. L.; Krska, S. W.; Bergman R. G. J. Am.
Chem. Soc. 2000, 122, 751-761. (d) Zuckerman, R. L.; G., B. R.
Organometallics 2000, 19, 4795-4809.
* To whom correspondence should be addressed. E-mail: tmeyer@
pitt.edu.
(1) Jordan, R. F.; Coles, M. P. J. Am. Chem. Soc. 1997, 119, 8125-8126.
(2) (a) Chisholm, M. H.; Eilerts, N. W.; Huffman, J. C.; Iyer, S. S.; Pacold,
M.; Phomphrai, K. J. Am. Chem. Soc. 2000, 122, 11845-11854.; (b)
Radano, C.; Baker, G. L.; Smith, M. R., III J. Am. Chem. Soc 2000, 122,
1552-1553.; (c) Chamberlain, B. M.; Cheng, M.; Moore, D. R.; Ovitt, T.
M.; Lobkovsky, E. B.; Coates, G. W. J. Am. Chem. Soc. 2001, 123, 3229-
3238.; (d) Ovitt, T. M.; Coates, G. W. J. Am. Chem. Soc. 1999, 121, 4072-
4073.
(7) Zuckerman, R. L.; Bergman R. G. Organometallics 2001, 20, 1792-1807.
(8) Bruno, J. W.; Li, X. J. Organometallics 2000, 19, 4672-4674.
(9) Toth, G.; Pinter, I.; Messmer, A. Tetrahedron Lett. 1974, 9, 735-738.
(10) Johnson, A. W. In Ylides and Imines of Phosphorus, 1st ed.; John Wiley
& Sons: New York, 1993; pp 221-305.
(11) Ivin, K. I.; Mol, J. C. Olefin Metathesis; Academic Press: London, 1997.
9
10698
J. AM. CHEM. SOC. 2002, 124, 10698-10705
10.1021/ja020494v CCC: $22.00 © 2002 American Chemical Society

Catalytic Double-Bond Metathesis
A R T I C L E S
Figure 2. Iminophosphorane dimer/monomer equilibrium.
Scheme 1
Figure 1. Similarities between the Wittig reaction and olefin/imine
metatheses.
catalysts. These main-group imide complexes23 mediate the
dNR exchange between carbodiimides to give statistical
mixtures of symmetric and unsymmetric carbodiimides (Scheme
1).24 Herein, we explore the generality of the reaction, identify
key intermediates, and provide mechanistic data consistent with
an addition/elimination pathway that resembles, in most of its
features, the Chauvin pathway proposed for olefin metathesis.
Both experimental observations12 and theoretical studies12-14
on transition metal imide complexes invoke a high degree of
participation of d-orbitals in the multiple bond and the avail-
ability of empty d-orbitals for coordination of substrate. In
contrast, modern calculations on main group hypervalent
compounds suggest that although the presence of d-orbitals is
necessary to successfully model any of the heavier main group
elements, they do not play a significant role in bonding.10,15,16
It has been proposed instead that negative hyperconjugation
(πE f σ*PX, E ) C, N) may be responsible for any π character
in the P-C or P-N bonds.16 These results have been upheld
by recent calculations dealing specifically with the Wittig-type
reactivity of phosphorus ylides17 and iminophosphoranes.18,19
Note that in this paper we will continue to represent the
iminophosphoranes as doubly bonded for the sake of visual
simplicity.
Another more directly observable difference between the
known main group phosphorus ylide reactions and the transition
metal systems is the stoichiometry. Traditional Wittig-type
reactivity, in contrast with olefin metathesis, is stoichiometric
and relies on the formation of thermodynamically stable
phosphorus-oxygen bonds to drive the formation of CdC,
CdN, or CdP bonds.10,20 There are, however, some reports of
phosphorus acting catalytically for example the disproportion-
ation of isocyanates to give carbodiimides,21 and anecdotal
reports of stoichiometric metatheses of iminophosphoranes with
carbodiimides.22
Results
Iminophosphoranes. Iminophosphoranes with chloride sub-
stituents on the phosphorus (Cl3PdNR) have, thus far, proven
to be the most successful carbodiimide metathesis catalysts. In
contrast with ubiquitous P-phenyl-substituted iminophosphorane
(Ph3PdNPh), the trichloroiminophosphoranes naturally dimerize
to form four-membered self-adducts (Figure 2). This tendency
toward self-dimerization, which was reported previously by
Becke-Goering,25 Fluck,26 and Zhmurova27,28 and co-workers
is facilitated by the relatively small size and electron-withdraw-
ing nature of the chloride substituents.
To explore the generality of the dimerization reaction and to
probe the correlation between dimer stability and catalyst
efficiency, a series of iminophosphoranes with varying P- and
N-substitution was prepared using adaptations of literature
methods. In Table 1, we report the phosphorus NMR spectro-
scopic shifts for both monomer and dimer (if observable) forms
and the percentage of material that is monomeric at 95 °C. Our
NMR correlate well with those reported previously for com-
pounds 1, 4, and 6.29
The isolation of a pure sample of the isopropyliminophos-
phorane 2 from the standard reaction of lithium isopropylamide
with PCl5 proved difficult. Distillation did not separate the
product from unreacted PCl5. We were, however, able to obtain
31P NMR spectral data and from that determine that 2 is
monomeric in solution. The formulation of 2 was confirmed
by derivatization with DIC to yield an isolable and well-
characterized diazaphosphetidine derivative (diazaphosphetidine
synthesis and characterization is discussed in detail later in this
report).
We have discovered that, despite these differences, imino-
phosphoranes of the type Cl3PdNR can act as CdN metathesis
(12) Nugent, W. A.; Mayer, J. M. Metal-Ligand Multiple Bonds; John Wiley &
Sons: New York, 1988.
(13) (a) Cundari, T. R. Chem. ReV. 2000, 100, 807-818. (b) Wang, C.-C.; Tang,
T.-H.; Wang, Y. J. Phys. Chem. A 2000, 104, 9566-9572. (c) Monteyne,
K.; Ziegler, T. Organometallics 1998, 17, 5901-5907.
(14) (a) Cundari, T. R.; Gordon, M. S. Organometallics 1992, 11, 55. (b) Folga,
E.; Ziegler, T. Organometallics 1993, 12, 325-337.
(15) Magnussson, E. J. Am. Chem. Soc. 1993, 115, 1051-1061.
(16) Reed, A. E.; von Rague Schleyer, P. J. Am. Chem. Soc. 1990, 112, 1434-
1445.
(17) (a) Doblado, J. A.; Martinez-Garcia, H.; Molina, J. M.; Sundberg, M. R. J.
Am. Chem. Soc. 2000, 122, 1144-1149. (b) Yamataka, H.; Nagase, S. J.
Am. Chem. Soc. 1998, 120, 7530-7536. (c) Restrepo-Cossio, A. A.;
Gonzalez, C. A.; Mari, F. J. Phys. Chem. 1998, 102, 6993-7000. (d) Naito,
T.; Nagase, S.; Yamataka, H. J. Am. Chem. Soc. 1994, 116, 10080-10088.
(18) (a) Lu, W. C.; Liu, C. B.; Sun, C. C. J. Phys. Chem. A 1999, 103, 1078-
1083. (b) Lu, W. C.; Sun, C. C.; Zang, Q. J.; Liu, C. B. Chem. Phys. Lett.
1999, 311, 491-498.
(23) Main group imides are also known for the heavier members of the
pnictogens and chalcogens. (a) Chivers, T. Can. J. Chem. 2001, 79, 1841-
1850. (b) Beswick, M. A.; Wright, D. S. Coord. Chem. ReV. 1998, 176,
373-406.
(24) A preliminary account of this work has been published previously. Bell,
S. A.; Geib, S. J.; Meyer, T. Y. J. Chem. Soc., Chem. Commun. 2000, 16,
1375-1376.
(19) Koketsu, J.; Ninomiya, Y.; Suzuki, Y.; Koga, N. Inorg. Chem. 1997, 36,
694-702.
(25) Becke-Goehring, M.; Leichner, L.; Scharf, B. Z. Anorg. Allg. Chem. 1966,
343, 154-164.
(20) Shah, S.; Protasiewicz, J. D. Coord. Chem. ReV. 2000, 210, 181-201.
(21) Campbell, T. W.; Monagle, J. J.; Foldi, V. J. Am. Chem. Soc. 1962, 84,
3673-3677.
(22) (a) Appel, R.; Guth, E. Z. Naturforsch., B: Chem. Sci. 1960, 15, 57-69.
Huisgen, R.; Wulff, J. Chem. Ber. 1969, 102, 1833-1840. (b) Bodeker, J.;
Kockritz, P.; Courault, K. Z. Chem. 1979, 19, 59. (c) Hall, R. C.; Smith,
D. J. H. J. Chem. Soc., Perkin Trans. 2 1977, 1379-1382.
(26) Fluck, E., Wachtler, D. Liebigs. Ann. Chem. 1980, 1651-1658.
(27) Zhmurova, I. N.; Kisilenko, A. A.; Kirsanov, A. V. Zh. Obsch. Khim. 1962,
32, 2580-2586.
(28) Zhmurova, I. N.; Drach, B. S. Zh. Obsch. Khim. 1964, 34, 1441-1446.
(29) (a) Gutmann, V.; Utvary, K.; Bermann, M. Monatsh. Chem. 1966, 97,
1745-1762. (b) Thonnessen, H.; Siedentop, T.; Jones, P. G.; Schmutzler,
R. Z. Anorg. Allg. Chem. 2001, 627, 731-741.
9
J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002 10699

A R T I C L E S
Bell et al.
Table 1. Physical and Spectroscopic Information for
Iminophosphoranes
predominant % monomer
a
cmpd
iminophosphorane
Cl3PdN-Prnb
31P NMR (δ)
species at rt
95 °C
1
-67.8 (monomer)e dimer
-78.4 (dimer)
-70.6 (monomer) monomer
-81.3 (monomer) monomer
-48.6 (monomer) dimer
-77.0 (dimer)
-53.5 (monomer) dimer
-78.1 (dimer)
-45.6 (monomer) dimer
-76.4 (dimer)
2
2
3
4
Cl3PdN-Pric
Cl3PdN-Butc
Cl3PdN-Phd
>95f
>95f
78
5
6
Cl3PdN-p-tolyld
Cl3PdN-2-fluorophenyld
Ph3PdNPhd
50
>95f
7
8
-1.2 (monomer) monomer
>95f
>95f
Figure 3. Sample GC catalysis data for carbodiimide metathesis catalyzed
by Cl3PdNPrn.
(pyrrolyl)3PdN-2-fluorophenyl -32.4 (monomer) monomer
a In toluene at room temperature, relative to 85% H3PO4. b Prepared via
e
method A. c Prepared via method B. d Prepared via method C. Measured
f
at 100 °C. Complete disassociation into monomer by 31P NMR spectros-
copy.
Scheme 2
The nature of the nitrogen substituent is an important
determinant in the monomer/dimer equilibrium of a particular
Cl3PdNR iminophosphorane. Hence, the basic, nonbulky N-
alkyl iminophosphorane, 1, is a strong dimer that does not
readily dissociate significantly below ∼90 °C. The more bulky
N-iso- and N-tert-alkyl iminophosphoranes, 2 and 3, in contrast,
are monomeric at room temperature. Aryliminophosphoranes
4-6 form relatively weak self-adducts, dissociating significantly
to monomeric form above 55 °C. The steric bulk and electronic
nature of the phosphorus substituent also affects the monomer/
dimer equilibrium. Both of the iminophosphoranes with phos-
phorus substituents other than chlorine, 7 and 8, are monomeric,
which is consistent with the increased steric bulk and the
decreased electrophilicity at phosphorus.
Figure 4. Comparison of the initial production of CIC for the catalysts
used (data for catalyst 7 and “no catalyst” coincide on the baseline with no
activity).
Scheme 3
react with carbodiimides to give diazaphosphetidine cycload-
ducts.24 For example, stoichiometric addition of DIC to 2-fluo-
rophenyliminophosphorane, followed by heating at 50 °C, yields
Catalytic Carbodiimide Cross-Metathesis. All of the imino-
phosphoranes in Table 1 proved to be carbodiimide metathesis
catalysts, except for 7, which has phenyl substituents on the
phosphorus. A standard cross-metathesis of carbodiimides
involved heating a mixture of diisopropylcarbodiimide (DIC)
and dicyclohexylcarbodiimide (DCC) at 95 °C in toluene in the
1
a single product (Scheme 3). Analysis of H and 31P NMR
spectral data supports formation of diazaphosphetidine adduct
9. In particular, the 1H NMR spectrum shows two inequivalent
isopropyl resonances and a 31 Hz 3JPH coupling to only one of
the isopropyl methynes that is indicative of CdN addition across
the PdN bond, with N-to-P regiochemistry. This formulation
was further supported by X-ray crystallography (Figure 5). The
most noteworthy feature of the structure is that both nitrogens
are bound to the phosphorussthere is no indication of ring-
opened, betaine character. Details of the data collection and the
structure can be found in the original communication.24 Analo-
gous aza-Wittig oxazaphosphetidine intermediates have been
reported.30
1
presence of iminophosphorane (Scheme 2). H NMR spectra
indicated the presence of the new mixed cyclohexyl isopropyl-
carbodiimide (CIC), and the observation of additional 31P NMR
spectral resonances showed the presence of significant inter-
mediates during the reaction. These intermediates are discussed
in detail later. The reaction appears to be reasonably generals
metathesis is observed for all combinations of DIC, DCC, and
ditolylcarbodiimide (DTC).
The catalytic activities of iminophosphoranes 1-8 vary
significantly. To determine what factors were responsible for
these differences, the appearance of the mixed carbodiimide and
the disappearance of the substrates were monitored as a function
of time for each catalyst under a standard set of conditions. A
sample plot for catalyst 1 is shown in Figure 3. The differences
in activity as a function of N- and P-substituents can be
illustrated by plotting the initial rate data for appearance of CIC
for catalysts 1-8 on the same time axis (Figure 4).
The ease of formation of these diazaphosphetidines under
stoichiometric conditions suggests that they are good candidates
for the intermediates observed in the catalytic cross-metathesis
mediated by iminophosphoranes. If, in particular, we examine
the metathesis of DIC and DTC, we could expect to see up to
six possible diazaphosphetidines (all those shown in Table 2).
(30) (a) Sheldrick, W. S.; Schomburg, D.; Schmidpeter, A.; von Criegern, T.
Chem Ber. 1980, 113, 55-69. (b) Storzer, W.; Ro¨schenthaler, G.-V.;
Schmultzeler, R.; Sheldrick, W. S. Chem Ber. 1981, 114, 3609-3624. (c)
Francke, R.; Dakternieks, D.; Gable, R. W.; Hoskins, B. F.; Ro¨schenthaler,
G.-V. Chem Ber. 1985, 118, 922-930.
Diazaphosphetidine Intermediates. As we reported in our
preliminary communication, monomeric iminophosphoranes
9
10700 J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002

Catalytic Double-Bond Metathesis
A R T I C L E S
Table 3. Sample Catalysis NMR Data
31P NMR
shift (δ)
%of
signal
31P NMR
shift (δ)
%of
signal
diazaphosphetidine
diazaphosphetidinea
15
13
12
11
53.5
55.2
56.9
59.9
9.9 16
61.9 17
14.5
53.4
56.4
58.2
59.4
1.4
3.5
1.6
1.1
9
5.9 18
total % tolyl
92.2 total % fluorophenyl
7.8
a 2-fluorophenyl substituted diazaphosphetidines (Figure 6).
Figure 6. Proposed fluorophenyl-containing diazaphosphetidine intermedi-
ates (Ar ) 2-flurophenyl).
Figure 5. Molecular structure of 9. Hydrogen atoms have been omitted
for clarity, 50% probability ellipsoids.
15. It was not possible to assign a 31P NMR shift for 14 using
this methodology. We theorize that the resonance for 14 overlaps
with that of another phosphorus-containing compound in these
complex mixtures.
Table 2. Spectroscopic Details for Diazaphosphetidines
The isopropyl- and tolyl-containing diazaphosphetidines are,
in fact, present in catalytic reaction mixtures. Specifically, an
experiment involving 5 equiv each of DTC and DIC to 1 equiv
of 2-fluorophenyliminophosphorane 6 exhibited resonances for
10-13, and 15. The spectrum also contained four additional
peaks that comprised less than 8% of all the species present.
One of these minor peaks corresponds to the 2-fluorophenyl-
substituted diazaphosphetidine 9. The other three resonances
likely correspond to diazaphosphetidines that contain the original
2-fluorophenyl substituent (Table 3 and Figure 6). Tentative
assignments of these minor products have been made on the
basis of trends derived from the detailed study of the tolyl/
isopropyl diazaphosphetidine series.
Significantly, the diazaphosphetidines are also catalysts for
CdN metathesis. For example, diazaphosphetidine 10, when
added at a 5 mol % loading (the same conditions employed for
the iminophosphorane catalysts), metathesized DIC and DCC
at rates similar to those observed with pure iminophosphoranes.
These reaction mixtures also contained resonances for the other
diazaphosphetidines observed in the iminophosphorane catalytic
mixtures. A turnover frequency (TOF) of 0.33 turnovers per
phosphorus per hour (10800 s for a single turnover) was
determined for 10 at 65 °C.
Kinetics. If the diazaphosphetidines are intermediates in the
catalytic metathesis of carbodiimides, the rate of addition and
elimination of carbodiimides to iminophosphoranes is of interest.
However, initial attempts to observe the adduct formation by
addition of DIC to 2-fluorophenyliminophosphorane 6 proved
unsatisfactory. 31P NMR spectroscopic data collected during the
course of the experiment at 55 °C showed no monomer during
the course of conversion from dimer to diazaphosphetidine
adduct. Clearly, the rate-determining step was dissociation of
the iminophosphorane dimer (Scheme 4). Using monomeric
iminophosphoranes to directly observe the addition was not
possible. Catalyst 2 could not be isolated in pure form and 3
would not form stable diazaphosphetidines.
a 31P NMR shift not determined due to coincidence with other resonances.
To determine if these proposed intermediates are present in
catalytic reaction mixtures, we have measured 31P NMR
spectroscopic shifts for five of these six species. The diaza-
phosphetidines 10, 12, 13, 15 were prepared and isolated by
cycloaddition of either DIC or DTC to the appropriate imino-
phosphorane. The synthesis of the remaining two diazaphos-
phetidines was problematic, however, because two regioisomers
are possible from addition of a mixed carbodiimide to an
iminophosphorane. We, therefore, attempted to assign the 31P
NMR shifts for 11 and 14 by observing the ratios of all
diazaphosphetidine species as we titrated 10 and 15 with DTC
and DIC at sufficiently high temperatures that scrambling would
occur. For example, a sample of diazaphosphetidine 10 was
heated with 0.5 equiv of DTC. DIC was then added to shift the
equilibrium back toward propyl-containing species. Since
hysteresis was observedsthe equilibrium ratios of certain
adducts changedsit must be true that the dissociation of mixed
diazaphosphetidines or the regiochemistry of addition of mixed
carbodiimides to free iminophosphoranes or both favor certain
species over others. The relative ratios are not simply governed
by statistics. Subsequent incremental additions of DTC shifted
the equilibrium toward diazaphosphetidines with multiple tolyl
substituents. Interpolation of the data for a sequence of these
experiments allowed us to assign the resonance for 11. The
diazaphosphetidine spectroscopic assignments were further
confirmed by heating samples of isolated diazaphosphetidines
12 or 13 above 65 °C. Resonances were observed for the
expected diazaphosphetidine scrambling products, 10-13 and
The addition kinetic data were obtained instead by studying
the equilibrium between diazaphosphetidine and dissociated
iminophosphorane/carbodiimide using relaxation kinetic tech-
niques (Scheme 5). We chose to study the equilibrium mixture
9
J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002 10701

A R T I C L E S
Bell et al.
Scheme 4
Figure 8. Tris(pyrrolyl) iminophosphorane 8.
be calculated. Calculation of the time for a single turnover
assumes that metathesis is considered a simple sequence of
addition (k1), and elimination (k-1) steps, with the elimination
step being significantly slower and, thus, rate-determining.
Alternative Mechanistic Pathways. At higher temperatures
there is a significantly increased risk of metathesis-induced
byproducts of catalyst decomposition. Since a likely decomposi-
tion product in our system is HCl, a known catalyst for
carbodiimide metathesis, we carried out several experiments to
determine if it is a factor in these reactions. First, we produced
a non-chlorine-containing iminophosphorane catalyst 8 (Figure
8). Pyrrolyl was chosen as a chlorine replacement based on
Petersen’s characterization of trispyrrolylphosphorus ligands.32
Pyrrolyl is π-acidic, which allows replacement of chlorine
without loss of electrophilicity at the phosphorus centers
believed to be an important factor in the carbodiimide metathe-
ses. In fact the PdN bond length was found to be similar to
that of the analogous trichloroiminophosphoranes, indicating a
similarity in electronegativity of the ligands.33 Carbodiimide
metathesis with the trispyrrolyl catalyst was successful, but was
slower than with the trichloro catalysts. The increased time to
equilibrium can be attributed to the bulk of the pyrrolyl groups.
Although we cannot rule out the decomposition to produce
pyrrole at trace levels, we see no evidence for it spectroscopi-
cally.
Scheme 5
Table 4. Keq for Cycloaddition/Cycloreversion of 10
K
eq
temp (K)
K
temp (K)
eq
180
93.4
66.0
43.1
34.4
323
328
333
335
338
17.7
11.3
4.7
3.4
2.1
343
348
358
363
368
In addition, lowering the temperature should reduce the risk
of catalyst decomposition and HCl formation for the trichloro-
iminophosphoranes. Although initially we ran the majority of
our reactions at 120 °C, we found that successful carbodiimide
metathesis occurred at 65 °C, and 31P NMR spectra indicated
the presence of the same intermediates as in the previous
catalysis experiments. As would be expected, the time taken to
reach equilibrium (20 h) was significantly longer at the lower
temperature.
Finally, we were concerned about decomposition to produce
amines, which could also have a role in metathesis. Amine-
catalyzed processes for imine metathesis are implicated in
transition metal systems.34,35 The effect of amine on the
iminophosphorane catalytic reactions was studied by running
standard catalyses with trace amine. Although the amine
contaminated catalyses reached equilibrium faster than those
without trace amine added, we found no metathesis occurred
with control reactions containing amine and no catalyst.
Figure 7. Temperature dependence of Keq for cycloaddition-cycloreversion
of 10.
of addition/elimination products formed by dissolution of pure
diazaphosphetidine 10 because the high symmetry eliminates
the possibility of regioisomers or multiple products. Prior to
undertaking the kinetic studies, equilibrium constants over a
range of temperatures were obtained for the dissociation/
association products of 10 (Table 4). A semilog plot of Keq
versus 1/T gave the thermodynamic parameters, ∆H ) -97.6
((8) kJ/mol and ∆S ) -250 (( 20) J/(mol K) (Figure 7). The
equilibrium data were used to determine the ideal conditions
for the perturbation experiments.
To determine the kinetics of the dissociation, the sample was
equilibrated to 338 K and then rapidly cooled to 335 K, resulting
in a shift toward adduct. During the cooling process the
concentrations of all the relevant species were monitored as a
function of time. From a plot of ln(∆[2]/∆[2]0) versus time the
relaxation constant τ was determined to be 385 s. Using standard
assumptions,31 an expression for τ in terms of k1 and k-1 can
be derived. From the experimentally determined values for τ
and Keq at 335 K, k1 and k-1 can be calculated as 1.7 × 10-3
Discussion
Postulated Mechanism. The simplest proposed mechanism
for this catalytic system is based on Chauvin’s olefin metathesis
((0.1 × 10-3) M s-1 and 4.0 × 10-4 ((0.3 × 10-4) s-1
,
(32) Moloy, K. G., Petersen, J. L. J. Am. Chem. Soc. 1995, 117, 7696-7710.
respectively. A catalytic turnover frequency (TOF) of 1.44
turnovers per phosphorus per hour (2500 s per turnover) can
(33) Bell, S. A.; Geib, S. J.; Meyer, T. Y. Acta Crystallogr., Sect. C 2001, 57,
1341-1342.
(34) Burland, M. C.; Pontz, T. W.; Meyer, T. Y. Organometallics 2002, 21. In
press.
(31) Bernasconi, C. F. Relaxation Kinetics, 1st ed.; Academic Press: New York,
(35) McInnes, J. M.; Mountford, P. J. Chem. Soc., Chem. Commun. 1998, 16,
1669-1670.
1997.
9
10702 J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002

Catalytic Double-Bond Metathesis
A R T I C L E S
Figure 9. Iminophosphorane catalytic pathway.
mechanism36 and the Wittig reaction pathway (Figure 9). Others
have postulated a similar process for analogous transition metal
systems.6,7 In the proposed catalytic cycle, the iminophosphorane
dimer dissociates; subsequent cycloaddition of a carbodiimide
produces a diazaphosphetidine adduct. The new adduct can then
undergo cycloreversion to produce the first catalytically relevant
iminophosphorane. Repetition of the addition-elimination
process completes the catalytic cycle, generating another
equivalent of mixed carbodiimide. Only productive metatheses
are illustrated in the catalytic cycle. It must be assumed that
significant degenerate processes such as the reaction of isopro-
pyliminophosphorane 2 with DIC are concurrent.
There appear to be two factors that influence the activity of the
precatalysts. First, it is clear that activity is affected by the
tendency of iminophosphorane dimers to dissociate. Comparison
of the tolyl- and phenyl- iminophosphoranes, 5 and 6, in Figure
4 indicates that the fluorophenyliminophosphorane catalyst has
a slightly higher initial reactivity than the tolyl system. This
would be expected as the fluorophenyl system has a less stable
dimer. Second, one must also consider the nucleophilicity of
the iminophosphorane. We would expect the alkyl iminophos-
phoranes to be more nucleophilic than the aryl systems. This
nucleophilic trend is clearly seen in Figure 4. The most reactive
systems are the propyl (both isopropyl and n-propyl), followed
by the aryl systems. Interestingly, the nucleophilicity difference
between alkyl and aryl substituents is apparently more important
than the monomer/dimer equilibrium effects, since the alkyl
species are the fastest despite having the most stable dimers.
The tert-butyl derivative is the slowest simply because of steric
bulk inhibiting the formation of the diazaphosphetidines. We
also found that the trispyrrolyl-substituted catalysts were slower
than their chloro analogues. We believe that this is attributable
to the steric bulk of the pyrrolyl groups. It is important to note
that, while these pyrrolyl systems are slower, they do follow a
reactivity pattern similar to that of the chloro systems. Carbo-
diimide metathesis did not occur when catalyst was not added
nor when Ph3PdNPh 7 was present. The lack of reactivity 7
probably due to the steric bulk and poor electron-withdrawing
ability of the phenyl groups on the phosphorus.
Our approach to studying this mechanism considered first
the postulated catalytic intermediates. We were able to inde-
pendently isolate and analyze the majority of the proposed
adduct intermediates, and compare the resulting spectroscopic
evidence with data collected during the catalysis experiments.
The fact that all phosphorus resonances observed during the
catalytic metathesis of DIC and DTC could be attributed to either
iminophosphoranes or diazaphosphetidines that would be pre-
dicted to be present in such a reaction supports the addition/
elimination pathway. Moreover, heating isolated samples of
diazaphosphetidines that contained both alkyl and aryl groups
(e.g., 12) caused a scrambling of these substituents and produced
an equilibrium mixture of the same diazaphosphetidines found
during DIC and DTC cross-metathesis.
Changes in the relative number of the propyl and tolyl groups
present should have a statistical effect on which diazaphosphe-
tidines are present, and their approximate relative ratios.
However, we have found that no matter which catalyst was used,
adducts 12 and 13 were particularly prevalent, even when there
was an excess of propyl or tolyl groups, respectively. The
significant presence of 12 and 13 indicates that diazaphosphe-
tidines with one endocyclic propyl and one endocyclic tolyl are
likely more thermodynamically stable than other regioisomers.
Further support for our postulated mechanism comes from
the observed activity of diazaphosphetidines. As intermediates,
they should be expected to act as catalysts with approximately
the same reactivity as the “free” iminophosphoranes. This
assumption was substantiated by experimental observations. We
found that 10 is, in fact, one of the most active catalysts and is
comparative in behavior to Cl3PdNPrn 1.
Alternative Mechanistic Pathways. Although there is
significant evidence to support the diazaphosphetidine mecha-
nism our experience with transition metal systems34,37 led us to
investigate minor pathways that could initiate carbodiimide
metathesis. These include (i) decomposition of the catalyst to
A comparison of metathesis rate data gives us a qualitative
understanding of each precatalyst’s ability to convert into what
we believe is the active catalyst, a dissociated iminophosphorane
carrying a NR group from one of the carbodiimide substrates.
(36) Herrison, J. P.; Chauvin, Y. Makromol. Chem. 1970, 141, 161-176.
(37) Jordan, A. Y.; Meyer, T. Y. J. Organomet. Chem. 1999, 591, 104-113.
9
J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002 10703

A R T I C L E S
Bell et al.
produce HCl, which could subsequently propagate metathesis,
and (ii) metathesis catalyzed by trace amine impurity.
and an olefin is nonzero because d-orbitals are involved in MdC
bonding.14 Although the lack of betaine intermediates suggests
that the PdN addition reactions are also highly concerted, it
seems unlikely that the two systems can have identical pathways
given the lack of π-bonding and d-orbital participation in the
iminophosphoranes. Theoretical studies on this type of addition
have shown, however, that the lowest-energy pathway for
addition of unsaturated substrates to these dipolar compounds
can involve a related, symmetry-allowed concerted pro-
cess.18,19,39 Although no calculations have been performed
specifically on iminophosphorane-mediated CdN metathesis,
a similar concerted pathway should be accessible.40
Since HCl would also be a catalyst for the metathesis of
carbodiimides, it is important that we determine that it is not a
factor in these reactions. Although the decomposition of
P-chlorinated iminophosphoranes to produce HCl is a reasonable
postulate, our data are not consistent with its presence. First of
all, our catalysts are extremely stable. Heating for weeks at >120
°C does not cause any detectable change. Also, the catalysis
proceeds normally, but more slowly, when much lower tem-
peratures are employed (65 °C). Moreover, trispyrrolyl imino-
phosphorane 8, which cannot decompose to give HCl, shows
reasonable catalytic activity. Although it could be argued that
pyrrole, generated by an analogous decomposition, could serve
as a catalyst as well, no pyrrole was observed in the reaction
mixtures.
Conclusions
A metal is not required for a catalytic CdN metathesis
involving a Chauvin-type addition/ elimination pathway. Imino-
phosphoranes with electron-withdrawing substituents have been
shown to scramble dNR substituents of carbodiimides via
phosphetidine intermediates. Although turnover frequencies at
65 °C are relatively slow (1.44 TO/P/h), they compare favorably
to that reported, 1.77 TO/Zr/h at 105 °C, for the only catalytic
metal-based CdN metathesis system that has also been shown
definitively to proceed by sequential addition/elimination,
Cp*Cp(THF)Zr(dNBut).7 We now plan to extend this work to
imine metathesis with the aim of producing more diverse
applications for these novel iminophosphorane metathesis
catalysts.
Our final and most convincing evidence against competing
pathways, including HCl and amine-catalyzed processes, comes
from the comparison of TOFs for overall catalyses (0.33 TO/
P/h) and those determined from the kinetics of what we postulate
is the rate-determining step, the elimination of carbodiimide
from a diazaphosphetidines (1.44 TO/P/h). Although these two
TOFs are not identical (they differ by a factor of 4.3), they are
comparable given that the TOF calculated for the overall
metatheses must necessarily be a lower limit since it does not
account for the many redundant metatheses that must occur.
On the basis of simple statistical arguments, we would predict
a minimum factor of 3 underestimate of the TOF for the overall
reaction. Since the addition/elimination reaction sequence is
kinetically competent to explain the TOFs we observe, it appears
unlikely that other competing pathways are contributing sig-
nificantly to the rates.
Experimental Section
Solvents, Reagents, and General Data. Solvents were freshly
distilled under nitrogen prior to use. Toluene and benzene were distilled
from benzophenone/sodium. Chloroform, methylene chloride, and
tetrachloroethane were dried over CaH2 and fractionally distilled. All
deuterated solvents used were purified in the same manner as the
nondeuterated equivalent after being received from Cambridge Isotope
Laboratories. Phosphorus pentachloride was used as purchased from
Strem Chemicals. Solid carbodiimides were purified by sublimation,
and liquid carbodiimides were purified by fractional distillation after
stirring over sieves. All amines used were purified by fractional
distillation after drying over CaH2, and stored in darkness, under
nitrogen.
Equipment. All manipulations of air- or water-sensitive compounds
were performed using standard high-vacuum or Schlenk techniques.
Solid organometallic compounds were transferred in a nitrogen-filled
drybox and were stored at room temperature, unless otherwise stated.
1H and 31P NMR spectra were recorded with a Bruker AF300
spectrometer at 300 MHz and 121 MHZ, respectively. Chemical shifts
were referenced to residual 1H NMR signals of the deuterated solvents
or external 85% phosphoric acid standards. The GC analyses were
performed on a Hewlett-Packard 5890 chromatograph filled with a
30.0-m methyl siloxane capillary column (HP19091Z-413E) and a flame
ionization detector.
We have shown that the mechanism of iminophosphorane
metathesis does, in fact, conform to the Chauvin mechanism in
that it consists of repeated addition/elimination steps. There
remain two issues that have to be addressed before the degree
of analogy can be fully established between the main group
and transition metal pathways: precoordination and concert-
edness. Although precoordination of olefins is postulated and
has been observed in transition metal-mediated olefin metathesis
systems,38 generally these adducts are not sufficiently stable to
be observed.4 The lack of energetically accessible d-orbitals
would seem to preclude such an interaction for phosphorus. In
recent theoretical studies on the aza-Wittig reaction, however,
Koketsu et al. have found weak complexes that form by
complementary interaction of the polar P-N and substrate CdO.
These adducts exhibit some charge transfer from the nitrogen
to the substrate but not sufficient to classify these intermediates
as betaine-type structures.19 An analogous approach would be
likely for a CdN substrate although sterics (N-substituent)
should destabilize the intermediate and may explain why such
adducts are not observed.
Method A (Based on a Preparation by Fluck et al.).26 N-4-Tolyl-
p-trichloroiminophosphorane (5). In a three-necked flask equipped
with a condenser, N2-inlet and addition funnel, PCl5 (3.90 g, 18.8 mmol)
was added over a period of 10 min to toluidine (2.01 g, 18.8 mmol) in
The second issue, concertedness, could also be expected to
be different in the metal and nonmetal systems. In the Chauvin-
pathway for olefin metathesis, a high degree of concertedness
is observed because, in contrast with olefin/olefin interactions,
the π/π* overlap required for a [2π + 2π] between an alkylidene
(39) Theory suggests that concerted [2 + 2] reactions are not forbidden for
substrates that have signficant polarity. (a) Ho¨ller, R.; Lischka, H. J. Am.
Chem. Soc. 1980, 102, 4632-4635. (b) Haller, J.; Beno, B. R.; Houk, K.
N. J. Am. Chem. Soc. 1998, 120, 6468-6472. (c) Yamaguchi, K.; Fueno,
T.; Fukutome, H. Chem. Phys. Lett. 1973, 461-465. (d) Yamaguchi, K.;
Fueno, T.; Fukutome, H. Chem. Phys. Lett. 1973, 22, 466-471.
(40) It has been suggested that heterocumulene additions to some unsaturated
substrates may involve more than a [2 + 2] interaction. See, for example
Deubel, D. V. J. Phys. Chem. A 2002, 106, 431-437.
(38) (a) Kress, J.; Osborn, J. A. Angew. Chem., Int. Ed. Engl. 1992, 31, 1585-
1587. (b) Tallarico, J. A.; Bonitatebus, P. J., Jr.; Snapper, M. L. J. Am.
Chem. Soc. 1997, 199, 77157-7158.
9
10704 J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002

Catalytic Double-Bond Metathesis
A R T I C L E S
1,2-tetrachloroethane (40 mL). The mixture was stirred for 25 min at
room temperature and refluxed for a further 14 h. A white solid
precipitated upon cooling. Recrystallization from toluene afforded white
crystals (2.45 g, 56.2%). 1H NMR (300 MHz, d8-toluene) δ 7.10 (d/d,
4, aryl), 2.03 (s, 3, tolyl); 31P{1H} NMR (121 MHz, d8-toluene) δ -78.1.
Method B (Based on a Preparation by Zhmurova et al.).28 N-n-
Propyl-p-trichloroiminophosphorane (1). A three-necked flask equipped
with a condenser and N2-inlet was charged with PCl5 (2.98 g, 14.3
mmol), propylamine hydrochloride (1.36 g, 14.3 mmol), and chloro-
benzene (15 mL). The resulting mixture was heated at approximately
90-100 °C for 4 h. White crystals formed upon cooling. Recrystalli-
solution that was reduced to 2 mL. Cooling to -35 °C yielded white
crystals (0.257 g, 21.8%). 1H NMR (300 MHz, d8-toluene) δ 6.95 (m,
4, aryl), 4.4 (d/sept, 1, P-N-CH(CH3)2), 3.45 (sept, 1, P-N-C-N-
CH(CH3)2), 2.05 (s, tolyl), 1.53 (d, 7, P-N-C-(CH3)2), 0.95 (d, 7,
P-N-C-N-CH(CH3)2); 31P NMR (121 MHz, d8-toluene) δ -56.9;
Analysis: 45.08 C, 5.96 H, 11.37 N; Calculated: 45.51 C, 5.99 H,
11.22 N.
Diazaphosphetidine Scrambling. 3,4-Diisopropyl-1-tolyl-1,3-di-
azaphosphetidine 12 (18.0 mg, mmol) and d8-toluene (0.65 mL) were
added to a screw-valve NMR tube. The reaction mixture was heated
to 95 °C and monitored for 48 h.31P NMR{1H} (121 MHz, d8-toluene)
δ -54.6 (s), -55.2 (s), -56.9 (s), -59.3 (s), -70.5 (s), -78.1 (s)
Incremental Carbodiimide Addition. 3,4-Diisopropyl-1-tolyl-1,3-
diazaphosphetidine 12 (17.5 mg, 0.055 mmol), ditolylcarbodiimide (6.1
mg, 0.027 mmol), and d8-toluene (0.65 mL) were added to a screw-
valve NMR tube. The sample was heated to 95 °C for 24 h. At one-
day intervals diisopropylcarbodiimide (9 µL, 0.027 mmol), ditolylcar-
bodiimide (12.2 mg, 0.055 mmol), ditolylcarbodiimide (12.2 mg, 0.055
mmol), ditolylcarbodiimide (12.2 mg, 0.055 mmol) were added. After
each addition, the reaction mixture was heated to 95 °C for several
hours before the next 31P{1H} NMR spectrum was acquired.
Comparison of Catalyst Rates. A stock solution of diisopropyl-
carbodiimide (8.90 mmol) and dicyclohexylcarbodiimide (8.90 mmol)
in toluene (84 mL) was prepared. iminophosphorane 1 (24.8 mg, 0.128
mmol) was added to a 12-mL aliquot of this solution. The resulting
solution was then divided further into 8 × 1.5 mL samples. One sample
was used to determine the initial concentrations, and the other seven
samples were placed in an oil bath at 95 °C. The samples were removed
at set time intervals, immediately cooled in an acetone bath, and
analyzed by GC (Figure 3). Using the same initial DIC/DCC stock
solution this process was simultaneously performed with a control (no
catalyst), diazaphosphetidine 10, and catalysts 3, 4, 5, and 7.
1
zation from hexanes yielded white crystals (1.97 g, 70.3). H NMR
(300 MHz, d8-toluene) δ 3.10 (m, 2, P-CH2), 1.83 (m, 2, CH2), 0.68
(t, 3, CH3); 31P{1H} NMR (121 MHz, d8-toluene) δ -78.4 (s).
Method C (Based on a Preparation by Schwesinger et al.)41
N-Isopropyl-p-trichloroiminophosphorane (2). In a three-necked flask
equipped with a condenser, N2-inlet and addition funnel, PCl5 (4.0 g,
19.0 mmol) was added over a period of 10 min to isopropylamine
hydrochloride (1.83 g, 19.0 mmol) in 1,2-tetrachloroethane (40 mL).
The mixture was stirred for 25 min at room temperature and warmed
to 80 °C for a further 6 h. Vacuum distillation yielded a solution
containing isopropyliminophosphorane and PCl3. 1H NMR (300 MHz,
d8-toluene) δ 4.35 (d/sept, 1, P-N-CH), 1.11 (dd, 7, CH-CH3);31P-
{1H} NMR (121 MHz, d8-toluene) δ +220 (s, PCl3), -70.6 (s,
Cl3PdNPri).
N-2-Fluorophenyl-p-trispyrrolyliminophosphorane (8). Potassium
pyrrolide (0.518 g, 4.93 mmol) was slowly added to a vigorously stirred
solution of Cl3PdNAr (0.405 g, 1.643 mmol) in hexanes (45 mL) at
-78 °C. The mixture was stirred for 30 min at -78 °C, warmed to
room temperature and stirred for an additional 14 h. Filtration yielded
a colorless liquid that was reduced to 15 mL and cooled to -35 °C.
The precipitate was removed by filtration. The filtrate was again reduced
1
in volume and cooled to yield colorless crystals (0.231 g, 41.5%). H
Keq Studies. 3,4-Diisopropyl-1-tolyl-1,3-diazaphosphetidine 12 (43.5
NMR (300 MHz, d8-toluene) δ 6.79 (m, 4, aryl), 6.67 (m, 6, pyrrolyl),
6.08 (m, 6, pyrrolyl); 31P{1H} NMR (121 MHz, d8-toluene) δ -32.4;
Analysis: 65.46 C, 5.12 H, 16.26 N; Calculated: 63.74 C, 5.01 H,
16.52 N.
mg, 0.136 mmol) and d8-toluene (0.70 mL) were added to a screw-
valve NMR tube. The reaction mixture was heated to 45 °C and allowed
to equilibrate for 30 min. 31P{1H} NMR (121 MHz, d8-toluene, 318
K) δ -59.0 (s, diazaphosphetidine), -71.5 (s, Cl3PdNPri). This process
was repeated at 45, 50, 55, 60, 63, 65, 70, 75, 85, 95 °C.
Relaxation Kinetic Studies. 3,4-Diisopropyl-1-tolyl-1,3-diazaphos-
phetidine 12 (43.5 mg, 0.136 mmol) and d8-toluene (0.70 mL) were
added to a screw-valve NMR tube. The reaction mixture was heated
to 65 °C and allowed to equilibrate for 1 h. This sample was then placed
in a NMR spectroscopic probe equilibrated at 63 °C. 31P NMR spectra
were obtained over a 30-min period. Integration was then used in order
to calculate the concentration of each of the species present. 31P{1H}
NMR (121 MHz, d8-toluene, 336 K) δ -59.0 (s), -71.5 (s).
Standard NMR Carbodiimide Metathesis. 2-Fluorophenylimino-
phosphorane 6 (18.4 mg, 0.074 mmol) and d8-toluene (0.65 mL) were
1
added to a screw-valve NMR tube. Initial 31P and H NMR spectra
were acquired. Then diisopropylcarbodiimide (58.4 µL, 0.37 mmol)
and ditolylcarbodiimide (82.8 mg, 0.37 mmol) were added. The mixture
1
was heated overnight at 120 °C and further 31P and H NMR spectra
1
were obtained. H NMR (300 MHz, d8-toluene) DTC δ 6.95 (m, 4,
aryl), 2.06 (s, 3, CH3C6H4NdCdNtolyl); DIC - 3.38 (m, 1,
CH3CH(CH3)NdCdNPri), 1.04 (d, 6, CH3CH(CH3)NdCdNPri); TIC
3.38 (m, 1, CH3CH(CH3)NdCdNtolyl), 2.10 (s, CH3C6H4NdCdN-
Pri), 1.02 (d, 6, CH3CH(CH3)NdCdNtolyl); 31P NMR{1H} (121 MHz,
d8-toluene) δ -46.0 (s), -53.4 (s), -53.6 (s), -54.4 (s), -55.3 (s),
-56.4 (s), -57.0 (s), -58.2 (s), -59.5 (s).
3,4-Diisopropyl-1-tolyl-1,3-diazaphosphetidine (12). In a thick-
walled reaction vessel diisopropylcarbodiimide (0.55 mL, 3.53 mmol)
was added to a solution of tolyliminophosphorane 5 (0.775 g, 3.21
mmol) in toluene (25 mL) at room temperature. The mixture was heated
at 65 °C in a closed system for 24 h. Filtration yielded a pale-yellow
Acknowledgment. This research was supported by NSF
award CHE-0091400. T.Y.M. is a fellow of the Alfred P. Sloan
Foundation. S.A.B. thanks the University of Pittsburgh for an
Andrew Mellon Pre-Doctoral Fellowship.
Supporting Information Available: 31P NMR spectral data
for a typical carbodiimide cross-metathesis and a plot of
relaxation kinetic data (PDF). This is material is available free
of charge via the Internet at http://pubs.acs.org.
(41) Schwesinger, R.; Willaredt, J.; Schlemper, H.; Keller, M.; Schmitt, D.; Fritz,
H. Chem. Ber. 1994, 127, 2435-2454.
JA020494V
9
J. AM. CHEM. SOC. VOL. 124, NO. 36, 2002 10705