Rh-mediated polymerization of carbenes: Mechanism and stereoregulation

Article abstract of DOI:10.1021/ja073897+

Ligand variation, kinetic investigations, and computational studies have been used to elucidate the mechanism of rhodium-catalyzed diazoalkane polymerization. Variations in the "N,O" donor part of the catalyst precursors (diene)Rh^I(N,O) result in different activities but virtually identical molecular weights, indicating that this part of the precursor is lost on forming the active species. In contrast, variation of the diene has a major effect on the nature of the polymer produced, indicating that the diene remains bound during polymerization. Kinetic studies indicate that only a small fraction of the Rh (1-5%) is involved in polymerization catalysis; the linear relation between polymer yield and M_w suggests that the chains terminate slowly and chain transfer is not observed (near living character). Oligomers and fumarate/maleate byproducts are most likely formed from other "active" species. Calculations support a chain propagation mechanism involving diazoalkane coordination at the carbon atom, N₂ elimination to form a carbene complex, and carbene migratory insertion into the growing alkyl chain. N₂ elimination is calculated to be the rate-limiting step. On the basis of a comparison of NMR data with those of known oligomer fragments, the stereochemistry of the new polymer is tentatively assigned as syndiotactic. The observed syndiospecificity is attributed to chain-end control on the rate of N₂ elimination from diastereomeric diazoalkane complexes and/or on the migratory insertion step itself.

Full text of DOI:10.1021/ja073897+

Published on Web 08/23/2007
Rh-Mediated Polymerization of Carbenes: Mechanism and
Stereoregulation
Erica Jellema,† Peter H. M. Budzelaar,‡ Joost N. H. Reek,† and Bas de Bruin*,†
Contribution from the Department of Homogeneous and Supramolecular Catalysis, Van‘t Hoff
Institute for Molecular Sciences (HIMS), UniVersity of Amsterdam, Nieuwe Achtergracht 166,
1018 WV, Amsterdam, and Department of Chemistry, UniVersity of Manitoba,
Winnipeg, MB, R3T 2N2, Canada
Received May 30, 2007; E-mail: bdebruin@science.uva.nl
Abstract: Ligand variation, kinetic investigations, and computational studies have been used to elucidate
the mechanism of rhodium-catalyzed diazoalkane polymerization. Variations in the “N,O” donor part of the
catalyst precursors (diene)RhI(N,O) result in different activities but virtually identical molecular weights,
indicating that this part of the precursor is lost on forming the active species. In contrast, variation of the
diene has a major effect on the nature of the polymer produced, indicating that the diene remains bound
during polymerization. Kinetic studies indicate that only a small fraction of the Rh (1-5%) is involved in
polymerization catalysis; the linear relation between polymer yield and Mw suggests that the chains terminate
slowly and chain transfer is not observed (near living character). Oligomers and fumarate/maleate byproducts
are most likely formed from other “active” species. Calculations support a chain propagation mechanism
involving diazoalkane coordination at the carbon atom, N2 elimination to form a carbene complex, and
carbene migratory insertion into the growing alkyl chain. N2 elimination is calculated to be the rate-limiting
step. On the basis of a comparison of NMR data with those of known oligomer fragments, the stereochemistry
of the new polymer is tentatively assigned as syndiotactic. The observed syndiospecificity is attributed to
chain-end control on the rate of N2 elimination from diastereomeric diazoalkane complexes and/or on the
migratory insertion step itself.
Introduction
One alternative route to the formation of highly functionalized
polymers would be through polymerization of (substituted)
Metal-catalyzed (co)polymerization of functionalized olefins
can be considered as the “holy grail” of polymerization
catalysis.1,2 It would not only afford access to new polymers
with controlled monomer sequences, but would also allow
precise tuning of stereochemistry through ligand variation.
However, (co)polymerizing such monomers is not easy, because
functional groups compete with the double bond for coordination
to the transition metal (both before and after insertion) and also
facilitate various types of side reactions that can lead to catalyst
deactivation. Late transition metals tend to be less sensitive to
both issues than early transition metals, and significant progress
has been made in particular with Ni and Pd catalysts.2
Nevertheless, this is by no means a solved problem.
diazoalkanes. Whereas unsubstituted diazoalkanes are in-
herently unstable, derivatives bearing electron-withdrawing
substituents (such as diazoacetates, N2CHCO2R) are reasonably
stable, safe5 (even in large scale/industrial synthesis), easy to
prepare, cheap, and therefore extensively used as carbene
precursors in organic synthesis. In the presence of transition
metals they generally convert to reactive carbene complexes,
which can undergo dimerization (to form olefins), carbene
transfer to olefins (cyclopropanation), and insertion into O-H,
N-H and C-H bonds.6 These reactions are most often mediated
by mononuclear platinum, copper, and ruthenium catalysts, or
dinuclear Rh(II)-acetates. They allow efficient and often
Group transfer polymerization (GTP) is a convenient way to
prepare acrylate and methacrylate polymers with narrow mo-
lecular weight distributions.3 However, as is the case with radical
(ATRP) and anionic polymerization, GTP methods give, at best,
poor control over the polymer stereoproperties.4
(4) (a) Ute, K.; Ohnuma, H.; Shimizu, I.; Kitayama, T. Polymer. J. 2006, 38,
999-1003. (b) Ute, K.; Tarao, T.; Hongo, S.; Ohnuma, H.; Hatada, K.;
Kitayama, T. Polymer. J. 1999, 31, 177-183. (c) Banerjee, K. G.; Hogen-
Esch, T. E. Polymer Prepr. (Am. Chem. Soc., DiV. Polym. Chem.)1987,
28, 320.
(5) (a) Clarck, J. D.; Shah, A. S.; Peterson, J. C. Thermochim. Acta 2002,
392-392, 177-186. (b) Clarck, J. D.; Shah, A. S.; Peterson, J. C.; Patelis,
L.; Kersten, R. J. A.; Heemskerk, A. H. Thermochim. Acta 2002, 386, 73-
79. (c) Clarck, J. D.; Shah, A. S.; Peterson, J. C.; Patelis, L.; Kersten, R.
J. A.; Heemskerk, A. H.; Grogan, M.; Camden, S. Thermochim. Acta 2002,
386, 65-72. (d) Clarck, J. D.; Heise, J. D.; Shah, A. S.; Peterson, J. C.;
Chou, S. K.; Levine, J.; Karakas, A. M.; Ma, Y.; Ng, K.-Y.; Patelis, L.;
Springer, J. R.; Stano, D. R.; Wettach, R. H.; Dutra, G. A. Org. Process
Res. DeV. 2004, 8, 176-185.
† University of Amsterdam.
‡ University of Manitoba.
(1) Tullo, A. H. Chem. Eng. News 2001, 79(43), 35-36.
(2) (a) Williams, B. S.; Leatherman, M. D.; White, P. S.; Brookhart, M. J.
Am. Chem. Soc. 2005, 127, 5132-5146. (b) Ittel, S. D.; Johnson, L. K.;
Brookhart, M. Chem. ReV. 2000, 100, 1169. (c) Gibson, V. C.; Spitzmesser,
S. K. Chem. ReV. 2003, 103, 283 and references therein.
(3) (a) Webster, O. W. Science 1991, 251, 887. (b) Webster, O. W. J. Polymer
Sci. A 2000, 38, 2855-2860. (c) Webster, O. W. AdV. Polymer Sci. 2004,
167, 1-34.
(6) (a) Merlic, C. A.; Zechman, A. L. Synthesis 2003, 1137. (b) Ye, T.;
McKervey, A. Chem. ReV. 1994, 94, 1092. (c) Maas, G. Chem. Soc. ReV.
2004, 33, 183.
9
10.1021/ja073897+ CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 11631-11641
11631

A R T I C L E S
Jellema et al.
Scheme 1. Formation of High-Molecular Weight Stereoregular
Polymers by Rh-Mediated Polymerization of “Carbenes” Formed
from EDA
Scheme 2. Synthesis of Catalyst Precursors 1-7
stereoselective conversion of the coordinated carbene to more
or less sophisticated products.6
Carbene insertion into a transition metal-alkyl bond, which
could lead to interesting functionalized polymers, is a far less
common reaction.7 There are a few reports describing BH3-,
Cu-, and Pd-mediated oligomerization of R-carbonyl stabilized
“carbenes” formed from diazocarbonyl compounds, producing
atactic low-molecular mass oligomers, with an average degree
of polymerization ranging from a few up to ca. 100.8-10 These
materials are notable because they carry a polar functionality
(amenable to further transformations) at each main-chain carbon
atom. The only other route to such compounds has been through
free-radical polymerization of dialkyl maleates or fumarates,11
with no control over the resultant polymer stereochemistry.
Results and Discussion
Catalyst Precursors. Previously, we reported that [(N,O-
ligand)RhI(cod)] complexes are catalyst precursors for the
polymerization of carbenes from alkyl diazoacetates.12 In the
search for the nature of the active species we focused on
variation of both the N,O- and the diene-ligand of the [(N,O-
ligand)RhI(diene)] species. The previously reported catalyst
precursors, 3-5,12 and the new complexes 2, 6, and 7 used for
the polymerization of carbenes formed from ethyl diazoacetate
(EDA) are depicted in Scheme 2. Complexes 2, 6, and 7 were
prepared according to the reactions shown (Scheme 2).
We recently reported that a family of Rh(diolefin) complexes
bearing nontraditional “hard” N,O-type ligands is able to
polymerize carbenes formed from diazoacetates to yield ster-
eoregular ester-functionalized polymethylenes with a high
molecular mass.12 In particular, the use of the amino-acid
L-proline as a ligand for rhodium proved to give quite active
and selective catalysts for the polymerization of carbonyl-
substituted methylene carbenes (see Scheme 1). Polymers with
molecular weights up to 190000 (Mw) were obtained as solid
materials with interesting material properties. To date, there are
no other catalysts that produce high-molecular mass or stere-
oregular polymethylenes with functional sidegroups. In the
present paper, we assign the stereochemistry of the new polymer
as syndiotactic on the basis of comparison of NMR data with
those of reported oligomer fragments.13 We also use ligand
variation and DFT calculations to shed some light on the nature
of the catalytically active species. We first focused on under-
standing the chain propagation mechanism to provide a rationale
for the formation of stereoregular polymers. Investigations
aiming at understanding initiation and termination pathways are
ongoing and will be presented in another paper.
Similar to the previously reported complexes 3-5, the
complexes 2 and 6 reveal expected and uncomplicated NMR
spectra, showing the presence of only a single isomer in solution.
The new [(L-prolinate)RhI(endo-dicyclopentadiene)] complex 7
1
shows more complicated H and 13C NMR spectra due to the
presence of four isomers. Reaction of enantiomerically pure
L-proline (S-proline) with racemic [{(dcp)Rh(µ-Cl)}2] in a 2:1
ratio is expected to result in formation of two diastereoisomeric
[(L-prolinate)RhI(dcp)] species 7a and 7b in a 1:1 ratio. The
NMR spectra clearly reveal the presence of four isomeric [(L-
prolinate)RhI(dcp)] species, two by two (7a with 7a′ and 7b
with 7b′) species in dynamic equilibrium with each other
according to EXSY spectroscopy (see Figure 1). The species
are present in different ratios (7a:7a′:7b:7b′ ) 1:0.9:1.2:0.6),
but as expected the sum 7a + 7a′ equals the sum 7b + 7b′.
The spectroscopic difference between these species is most
(7) (a) Greenman, K. L.; Carter, D. S.; Van Vranken, D. L. Tetrahedron 2001,
5219-5225. (b) Greenman, K. L.; Van Vranken, D. L. Tetrahedron 2005,
6438.
1
pronounced for the olefinic signals, both in H NMR and in
13C NMR. The olefinic signals f and g (see Figure 1) of the
isomers 7a and 7b are substantially upfield shifted relative to
those of 7a′ and 7b′, whereas the a/b olefinic signals shift
downfield (see Table 1). This clearly points to the presence of
cis/trans isomers, with 7a and 7b having the more symmetrical
f/g olefin trans to the anionic O-donor and 7a′ and 7b′ having
the least symmetrical olefin a/b trans to this donor. Stronger
metal-to-olefin π-back-donation to the olefin trans to the (π-
donating) prolinate O-donor explains the observed upfield shifts
of the different olefin signals.
(8) Bai, J.; Burke, L. D.; Shea, K. J. J. Am. Chem. Soc. 2007, 129, 4981-
4991 and references therein.
(9) Liu, L.; Song, Y.; Li, H. Polym. Int. 2002, 51, 1047.
(10) (a) Ihara, E.; Haida, N.; Iio, M.; Inoue, K. Macromolecules 2003, 36, 36.
(b) Ihara, E.; Fujioka, M.; Haida, N.; Itoh, T.; Inoue, K. Macromolecules
2005, 38, 2101. (c) Ihara, E.; Kida, M.; Fujioka, M.; Haida, N.; Itoh, T.;
Inoue, K. J. Polym. Sci., Part A: Polym. Chem. 2007, 45, 1536-1545.
(11) Toyoda, N.; Yoshida, M.; Otsu, T. Polym. J. 1983, 15, 255.
(12) Hetterscheid, D. G. H.; Hendriksen, C.; Dzik, W. I.; Smits, J. M. M.; van
Eck, E. R. H.; Rowan, A. E.; Busico, V.; Vacatello, M.; Van Axel Castelli,
V.; Segre, A.; Jellema, E.; Bloemberg, T. G.; de Bruin, B. J. Am. Chem.
Soc. 2006, 128, 9746-9752.
(13) Yoshioka, M.; Matsumoto, A.; Otsu, T.; Ando, I. Polymer 1991, 32/15,
2741-2746.
9
11632 J. AM. CHEM. SOC. VOL. 129, NO. 37, 2007

Rh-Mediated Polymerization of Carbenes
A R T I C L E S
Figure 1. Relative energies of the DFT optimized (Turbomole, b3-lyp, TZVP) geometries of diastereoisomers 7a and 7b, each consisting of four possible
geometrical isomers as a result of cis-trans isomerism and pyramidal inversion at nitrogen (in kcal/mol, not corrected for ZPE or thermal effects).
Table 1. Chemical Shift Values of the Olefinic NMR Signals of 7a,
These molecular weights and weight distributions, therefore,
deviate slightly from the previously reported data, which were
calibrated with MALLS.12
7a′, 7b, and 7ba,b
1H
δ
(ppm)
7b
13C
7a
δ
(ppm)
7b
7a
7a′
7b
′
7a
′
7b′
If we focus on the formation of well-defined stereoregular
PEA, it is clear that the use of the Rh(cod) catalyst precursors
1-5 with different O,O- and N,O-ligands (chiral and non-chiral)
produces virtually identical polymers with comparable molecular
weights and weight distributions, albeit in different yields
(entries 1-5). These results suggest that the O,O- or N,O-ligands
might be involved in the initiation of the polymerization but
are unlikely to act as stabilizing ligands to Rh during the
propagation steps. In contrast, varying the diene ligand of the
[(L-prolinate)RhI(diene)] catalyst precursors 5-7 leads to poly-
mers with very different molecular weights and weight distribu-
tions (entries 5-7). Most remarkably, the RhI complex with
the dicyclopentadiene (dcp) ligand (7) produces substantially
higher-molecular weight polymers (Mw ) 540 kD) with a much
narrower molecular weight distribution (Mw/Mn ) 2.0) com-
pared to the previously reported cod analogue 5 (Mw ) 150
kD, Mw/Mn ) 3.6) under otherwise identical reaction conditions
(Table 2). These data clearly point to the involvement of Rh-
(diene) species in the chain propagation.14 Since high-molecular
weight polymers are obtained while using only 50 equiv of
EDA, only a few percent of the catalyst precursor can be active
in the polymerization reaction (5: ∼3%, 7: ∼1%). These low
initiation efficiencies do not exclude the possibility of Rh
adventitious impurities playing a role in the catalytic reaction.
However, different batches of the Rh(diene) catalysts, prepared
from different Rh sources, give rise to identical results in the
catalytic carbene polymerization. Using repeatedly recrystallized
batches of catalyst precursor 5 and freshly distilled EDA and
a
b
f
4.81 4.04 4.92 4.52 87.59
4.92 4.04 4.98 4.04 85.22
3.81 4.67 4.04 4.70 64.84
78.48 87.08
73.59 84.18
75.94 65.78
78.48
73.22
75.14
g
5.50 6.39 5.18 6.20 89.59 102.28 90.52 101.82
a Measured in CDCl3 at 298 K (1H: 500 MHz and 13C: 125 MHz).
b For labeling, see Figure 1.
In good agreement with the experimentally derived ratios,
DFT calculations reveal that the species 7a′ and 7b′ lie about
+0.33 and +0.19 kcal/mol higher in energy than 7a and 7b,
respectively (Figure 1).
Coordination of the prolinate nitrogen makes this donor
enantiotopic, and in principle both R and S configurations are
possible for this atom through pyramidal inversion. In a previous
report we published the X-ray structure of the [(L-prolinate)-
RhI(cod)] analogue, revealing an S configuration of the N-donor
(and no other isomers were observed with NMR spectroscopy).12
We expected 7a and 7b to behave in a similar way, and indeed
the relative energies (DFT calculations) of the species having a
nitrogen R configuration (7aN-Inverted, 7a′N-inverted, 7bN-Inverted
,
and 7b′N-inverted) are about 3 kcal/mol higher compared to the
respective species having a nitrogen S configuration (Figure 1).
In line with this, these N-inverted species were not detected by
NMR spectroscopy.
Carbene Polymerization Reactions. Polymerization reac-
tions were carried out by addition of 50 equiv of EDA to a
solution of the catalyst precursor in chloroform at room
temperature. After 14 h the solvent was evaporated, and the
polymer (poly(ethyl 2-ylidene-acetate), PEA) was isolated and
(14) Ancillary (chiral) diene ligands have also been observed to support the
active species in Rh-mediated polymerization of phenylacetylene and
asymmetric 1,4-additions to R,â-unsaturated ketones: (a) Saeed, I.;
Shiotsuki, M.; Masuda, T. Macromolecules 2006, 39, 8977-8981. (b)
Otomaru, Y.; Kina, A.; Shitani, R.; Hayashi, T. Tetrahedron: Asymmetry
2005, 16, 1673-1679. (c) Kina, A.; Ueyama, K.; Hayashi, T. Org. Lett.
2005, 7, 5889-5892.
1
washed with methanol. For all catalysts except 6, H NMR
spectroscopy indicated complete conversion of EDA after 14 h.
The polymers and oligomers were analyzed by size exclusion
chromatography (SEC) calibrated with polystyrene samples.
9
J. AM. CHEM. SOC. VOL. 129, NO. 37, 2007 11633

A R T I C L E S
Jellema et al.
Table 2. Polymerization of Carbenes Formed from EDA with Catalysts 1-7a
a Conditions: 0.04 mmol catalyst; 2 mmol EDA, 5 mL chloroform (solvent), room temperature, reaction time: 14 h. b Isolated by precipitation and
washing with methanol. c SEC analysis calibrated against polystyrene samples. d Combined methanol soluble fractions after evaporation of the solvent.
e Estimated yield; part of the combined diethyl maleate and fumarate fraction is lost during workup of the polymer and oligomer fractions, using a vacuum
f
pump. ∼35% EDA recovered. g Bimodal distribution.
Figure 2. Time-product profiles for carbene polymerization reactions. (Top Left) Development of the product distribution in time using catalyst precursor
5. (Top Right): Correlation between the molecular weights of the polymer (Mw in kD) and obtained PEA yields (%) for catalyst precursors 5 and 7.
(Bottom) Development of the polymer molecular weights and yields in time using catalyst precursors 5 and 7. The numbers represent the obtained PDI
values (Mw/Mn) for each SEC sample.
solvent has no beneficial effect on the catalytic performance at
all, and deliberately adding water and/or introducing oxygen
has only a minor negative influence on the obtained polymer
yields. Furthermore, it is relevant to note that for Rh(diene)-
mediated polymerization of phenylacetylene inefficient initiation
(<5%) is quite commonly observed, with ligand modifications
having a substantial beneficial effect on the initiation efficiencies
(reaching up to 100% as recently reported by Masuda et al.).15
Note that the behavior of 7 points to generation of a single-
site active catalyst, while the spectroscopic data of the precata-
(15) Saeed, I.; Shiotsuki, M.; Masuda, T. Macromolecules 2006, 39, 8567-
8573.
9
11634 J. AM. CHEM. SOC. VOL. 129, NO. 37, 2007

Rh-Mediated Polymerization of Carbenes
A R T I C L E S
lyst clearly revealed the presence of four different isomeric
species in solution. These data further support the idea of
prolinate dissociation upon generation of the catalytically active
species.
For the reactions with complexes 1, 2, 5, 6, and 7 we also
analyzed the methanol soluble fractions left after workup and
washing the PEA with methanol. From this it became clear that
besides the previous reported (methanol insoluble) stereoregular
polymer and small amounts of diethyl maleate and fumarate,
the reactions also produce varying amounts of oligomers with
a number degree of polymerization of only about 11-20
“carbene” units. The extremely broad NMR spectra of these
oligomers (see Figure S1, Supporting Information) point to a
poorly defined structure, which is not at all related to the well-
defined structure of the stereoregular polymer. It is clear that
all of the applied catalyst precursors produce the same poorly
defined oligomers (identical NMR spectra and identical Mw and
Mw/Mn within experimental errors, see Table 2). It is thus un-
likely that the oligomerization process involves the same cata-
lytically active species responsible for the polymerization event.
We monitored the polymerization reaction catalyzed by cod
Figure 3. (Top) 13C NMR spectrum (CDCl3, 75 MHz, 298 K) of atactic
PEA prepared by radical polymerization of diethylfumarate,12 with assign-
ment of the mmmmmm heptad microstructure signals according to ref 13.
(Middle) (CDCl3, 125 MHz, 298 K) and (Bottom) (CDCl3, 125 MHz, 323
K): 13C NMR spectra of stereoregular PEA prepared with rhodium catalyst
7 (Table 2, entry 7).
1
complex 5 in time by H NMR spectroscopy. As shown in
Figure 2 (top left), the applied conditions lead to rather fast
EDA consumption in the beginning of the reaction, with
production of most of the byproducts. Within 1 h most (>75%)
of the EDA has been consumed, and the reaction mixture
consists of ∼38% oligomers, ∼15% diethyl maleate and
fumarate dimers (combined), and only ∼21% of polymer. After
1 h the total amount of diethyl maleate/fumarate and oligomer
byproducts hardly increases further (within experimental errors),
and consumption of the remaining EDA almost exclusively leads
to stereoregular PEA (Figure 2, top left). The final product
distribution determined with NMR spectroscopy deviates only
slightly from the distribution of isolated products (Table 2),
which is reasonable considering substantial experimental and
monomer depletion, which potentially limits the molecular
weight. However, using much more EDA (100 equiv) leads to
much lower overall yields (with nonconverted EDA being
recovered after the reaction) and does not lead to an increase
in the molecular weight of the obtained polymer. This clearly
suggests that the active species deactivates in time, even in the
presence of enough monomer. Further investigations in which
different concentrations of EDA are slowly added in time may
lead to further process optimization, but future efforts aiming
at finding more efficient ways to initiate the polymerization are
likely more rewarding.
The capricious behavior of the polymer PDI values in time
for samples obtained with catalyst precursor 5 is not so easily
understood in terms of a single-site catalyst, and perhaps 5
generates more than one species active in the carbene polym-
erization process. Somewhat higher-molecular weight PEA (205
kD) is attainable with catalyst precursor 5 upon addition of an
additional amount of EDA (25% of the initial amount) after 1
h. This leads to an overall PEA yield of only 45% (60% is
expected with full conversion of, at that point, nonconverted
plus additional EDA to PEA), again suggesting that, even in
the presence of enough monomer, termination processes stop
the polymerization reaction before complete EDA conversion.17
Assignment of Stereochemistry. The polymers obtained with
[(L-prolinate)RhI(cod)] are highly stereoregular, as shown by
detailed NMR spectroscopy studies in our previous paper.12 PEA
samples obtained with other Rh(diene) species give rise to
virtually identical NMR spectra and must thus have the same
microstructures (either all isotactic or all syndiotactic). A
detailed analysis by Otsu et al. of 13C NMR data of atactic poly-
(dialkyl fumarate)s with varying isoactic enrichments (depending
on the radical polymerization method) has allowed assignments
of the carbonyl and methine 13C NMR signals to the heptad
microstructures mmmmmm, mmmmmr, mmmmrm, rmmmrm,
mmmrmm, mmmrmr, mrmrmm, and mrmrmr.13 Unfortunately,
1
integration errors associated with H NMR integrations.16
SEC analysis of stereoregular PEA samples obtained after
varying reaction times yields further information. With catalyst
precursor 5 (Figure 2, bottom left), the polydispersity values
(PDI) of the obtained polymers is quite large initially (ca. 3.0),
increases further in the first hour of the reaction to about 4.0,
and then slowly decreases to a limiting value of ca. 3.0.
Formation of PEA using catalyst precursor 7 (Figure 2, bottom
right) is slower. The molecular weight distribution is quite
narrow initially (PDI ≈1.3) but broadens over time (to a PDI
of ca. 2.0).
The nearly linear correlation between the molecular weight
and yield for polymer samples obtained with both catalyst
precursors 5 and 7 (Figure 2, top right) is an indication for a
chain-growth mechanism proceeding without detectable chain
transfer (near-living character). The gradually broader molecular
weight distribution of polymer samples obtained with catalyst
precursor 7 in time points to slow termination of the chain-
propagating species. This is happening under conditions of
(16) Especially the oligomer contents in Figure 2 (left) are not highly reliable.
The oligomer backbone CH(COOEt) protons are very broad, hardly
detectable, and overlap with other signals, and are thus not suitable for
direct 1H NMR integration. The relative oligomer contents are therefore
estimated by subtracting the expected integral contributions of EDA, PEA,
and diethyl maleate/fumarate (based on separate integration of their
CH(COOEt) signals) present in solution from the total integral of
overlapping -OCH2CH3 signals of all species present.
(17) Direct addition of 100 equiv (instead of 50 equiv) EDA to a solution of 5
in CHCl3 led to an even lower overall PEA yield of 33% (Mw ) 210 kD,
Mw/Mn ) 4.3).
9
J. AM. CHEM. SOC. VOL. 129, NO. 37, 2007 11635

A R T I C L E S
Jellema et al.
Figure 4. Relative free energies (∆G0 in kcal/mol) associated with rate-determining carbene formation steps from A and MDA (b3lyp, TZVP), corrected
for ZPE and thermal effects (298 K).
reliable chemical shifts for rrrrrr heptads of syndiotactic
microstructures are not available for comparison because the
probability of racemic propagation Pr is near zero for radical
polymerization of dialkyl fumarates. However, observed devia-
tions from expected heptad sequences according to such
probability calculations upon polymerization of dimethyl fu-
marate at higher temperatures have been interpreted in terms
of increased amounts of rr triads, giving rise to 13C NMR shifts
at ∼171 ppm (carbonyl) and ∼45.3 ppm (methine). The 13C
NMR chemical shifts of the highly stereoregular polymers
obtained with Rh and EDA (170.8, 45.4 ppm) are very different
from those of the mmmmmm heptads of isotactically enriched
poly(dialkyl fumarate)s (170.2, 46.0 ppm), while they cor-
respond to those interpreted as rr triads. On this basis we assign
the polymer being syndiotactic (Figure 3, see also ref 12).
All calculations were performed with cod as the stabilizing
diene and with methyl diazoacetate (MDA) as a smaller model
substrate for EDA. Neutral (cod)RhI{CH(COOMe)}3CH3 spe-
cies A (obtained after three consecutive carbene insertions into
a Rh-Me fragment) is probably the smallest possible model
that realistically represents the steric and electronic properties
of the active catalyst. All calculations were initiated from species
A, containing a syndiotactic growing chain fragment.18 As
expected, geometry optimization of A led to coordination of a
carbonyl oxygen atom of the growing chain to the rhodium
center. The ester methoxy oxygen atoms can also coordinate to
rhodium, but this leads to substantially higher-energy structures,
which we therefore did not consider further. The most stable
geometry of A contains a favorable five-membered ring formed
by coordination of the â-carbonyl fragment (see Figure 4).
Coordination of the R-carbonyl is also possible, leading to a
somewhat higher-energy structure A′ (+0.7 kcal/mol) which
could be considered as a “hetero-allyl” species (see Supporting
Information).
Computational Studies
The experimental results clearly suggest the involvement of
a Rh(diene) active species in chain propagation. This makes
the involvement of any RhIII complex (kinetically inert, low
affinity for alkenes) rather unlikely, leaving (diene)RhI(alkyl)
as the most plausible active species. Such a species could, for
example, be generated in situ by nucleophilic attack of the
prolinate oxygen atom on a Rh-bound carbene fragment. On
this basis we performed computational studies in search for a
plausible mechanistic explanation for the preferred formation
of syndiotactic polymers with (diene)Rh catalysts.
Generation of a Rh-carbene fragment requires binding of
the prochiral diazo-bearing carbon atom of MDA to Rh after
carbonyl dissociation from A or A′, leading to two possible
diastereoisomeric MDA adducts that are close in energy (Bsyndio
+12.6 kcal/mol; Biso +12.3 kcal/mol). MDA coordination
(18) This fragment has an S configuration of the Rh-bound methine carbon.
The methine carbons of the rr triad have S,S,R configurations, becoming
R,S,R after the next carbene insertion.
9
11636 J. AM. CHEM. SOC. VOL. 129, NO. 37, 2007

Rh-Mediated Polymerization of Carbenes
A R T I C L E S
Figure 5. Relative enthalpies ∆H0 and free energies (∆G0 given in between brackets) of isotactic and syndiotactic carbene insertion steps (b3lyp, TZVP)
in kcal/mol, corrected for ZPE and thermal effects (298 K).
through a nitrogen atom is also possible (+7.5 kcal/mol) but is
not productive19 and was therefore not considered further.
We failed to find any transition state involving a direct attack
of the Rh-bound growing chain on the coordinated MDA
fragments in structures B. Such a mechanism would lead to
structures D in a single SN2-like substitution step, similar to
the mechanism recently proposed for BH3- and Pd-catalyzed
oligomerization of EDA.8,10 It thus seems more likely that our
Rh-mediated polymerization reaction proceeds via discrete
carbenoid intermediates C, and this pathway was further
explored.
In good agreement with other computational studies,20
exothermic formation of the carbenoids C seems to proceed
via dinitrogen loss from carbon-bound MDA adducts B, which
is over-all rate limiting along the calculated reaction pathway.
The transition-state free energies TS1 are within a reasonable
range for reactions proceeding within hours at RT (Figure 4).
The transition state for formation of Csyndio (TS1syndio, +21.8
kcal/mol) lies about 3 kcal/mol lower in free energy than that
for formation of Ciso (TS1iso, +24.9 kcal/mol).
Remarkably, the RhdC bond of structure C adopts a parallel
orientation with respect to the cod CdC bond trans to it. Thus,
different d orbitals of the rhodium metal center are used for
optimal π backbonding to the carbene fragment and to this
CdC bond. As a result, the carbene fragment is pre-organized
for migratory insertion into the Rh-C bond of the growing
chain, with Csyndio having its re face and Ciso having its si face
oriented for attack of the growing chain. This corresponds to
the two possible enantiotopic insertion pathways (racemic vs
meso) leading to syndiotactic and isotactic polymers. For steric
reasons, the {CH(COOMe)}3CH3 chains of B and C adopt
approximate trans-gauche conformations (torsion angles: CMe
C-C-C ≈ 160° and C-C-C-Rh ≈ 60°). Any other confor-
mation would cause substantial strain, either within the chain
itself or between the chain and the Rh(cod) or Rh(carbene)
fragments.
-
Once the carbene fragments are formed, migratory carbene
insertion into the Rh-C bonds of C to form the insertion
products D is a strongly exothermic, low-barrier process (Figure
5). Although formation of Csyndio is substantially easier than
formation of Ciso, carbene rotation provides a relatively low-
energy equilibration pathway between these species (TS2 )
+4.4 kcal/mol). The product-forming steps (C f D) are also
calculated to have low activation energies (∆Gq ) 6.3 and 7.0
kcal/mol), so the final selectivity will not be simply determined
by a single limiting step.
It should be noted here that entropy corrections appear to
artificially disfavor TS3syndio relative to TS3iso. The calculated
enthalpy difference of 2.2 kcal/mol is reduced to only 0.7 kcal/
mol on going to free energies. This is due to contributions from
a few very low-frequency modes, for which the harmonic
approximation used in the corrections loses its validity. It may
be that the use of relative enthalpy values is more appropriate
here. Indeed, for smaller models in which we replaced the
-{CH(COOMe)}3CH3 growing chain by -{CH(COOMe)}CH3
(which is less floppy, thus causing fewer problems associated
with low-frequency modes), the free energy for the syndiotactic
migratory insertion (∆Gq ) 3.6 kcal/mol) is 1.5 kcal/mol lower
than the corresponding isotactic insertion (∆Gq ) 5.1 kcal/mol).
In any case, the final selectivity will be intermediate between
that expected from B f C (k1/k1′), as selectivity-determining
on the one hand, and rapid Csyndio/Ciso equilibration leading to
(19) Howell, J. A. S. Dalton Trans. 2007, 1104.
(20) Cohen, R.; Rybtchinski, B.; Gandelman, M.; Rozenberg, H.; Martin, J. M.
L.; Milstein, D. J. Am. Chem. Soc. 2003, 125, 6532-6546.
9
J. AM. CHEM. SOC. VOL. 129, NO. 37, 2007 11637

A R T I C L E S
Jellema et al.
Scheme 3. Selectivity of the Syndiotactic over Isotactic Carbene
Insertion (ksyndio/kiso) Depending on Rate Constants k1, k1′, k2, k2′,
k3, and k3′
Table 3. Predicted Stereoselectivity ksyndio/kiso for a Steady-State
Model and Limiting k1/k1′ and k3/k3′ Values Based on Calculated
Barriers
B
f
C limiting
C
f
D limiting
(k1/k1
)
steady-state model
(k3/k3
)
′
′
∆H‡
∆G‡
187.8
187.8
50.1 (115.6a)
3.4 (16.5a)
48.6 (113.2a)
3.4 (13.0a)
a Values between brackets based on barriers for C f D (TS3syndio and
TS3iso) obtained with smaller, shorter growing chain models (see Supporting
Information) keeping TS1 and TS2 barriers of the larger model.
Figure 7. Selection of bond lengths (Å) and angles (deg) associated with
carbene formation and migratory insertion for both the calculated isotactic
and syndiotactic pathways.
shown in Figure 7. The geometric differences between the
syndio and iso paths are clearly minor and support our
interpretation of weak steric monomer-chain end interactions
as the selectivity-determining factor when applying cod as the
ancillary ligand.21
Figure 6. Simplified representation of Rh-mediated “carbene” polymeri-
zation under chain-end control, with minimal steric interactions between
the substituents of the growing chain and the carbene fragments favoring
the Csyndio f D migratory insertion.
C f D (k3/k3′), as selectivity-determining on the other hand
(see Scheme 3).
Conclusions
[(N,O-ligand)Rh(diene)] complexes mediate formation of
high-molecular weight and highly stereoregular functionalized
polymethylenes (PEA) from ethyl diazoacetate (EDA). On the
basis of a comparison of the 13C NMR data with those obtained
for isotactic enriched polyfumarates, the polymers obtained with
Rh and EDA are found to be syndiotactic. Variation of the
anionic N,O-ligand is only of influence on the polymer yields
and not on the polymer properties, whereas variation of the diene
ligand has a marked influence on the polymer properties. This
suggests that the N,O-ligand is only involved in the initiation
while the diene ligand is the metal-supporting ligand of the
catalytically active species during the propagation steps.
Kinetic studies indicate that only a small fraction of the Rh
(1-5%) is involved in polymerization catalysis; the linear
relation between polymer yield and Mw suggests that the chains
terminate slowly and chain transfer is not observed. Oligomers
and fumarate/maleate byproducts are most likely formed from
other “active” species.
Table 3 lists the actual predicted selectivity (from a steady-
state model, see Supporting Information) and these limiting
values, based on both enthalpy and free-energy differences.
Regardless of the energies used, syndiotactic propagation is seen
to be preferred. However, all of the energy differences here
are rather small, close to the error limits of the computational
methods employed, and the differences should be treated with
caution. In particular, the balance between rotation and insertion,
which has a significant effect on predicted stereoselectivity, can
also be expected to have a relatively large error margin.
The interplay between the different rates implies that the
ancillary ligand could well influence the stereoselectivity.
Ancillary ligand effects are difficult to predict though. For cod,
competition between the two respective carbene-forming steps
(B f C) and the two respective carbene insertion steps (C f
D) is clearly determined by carbene-chain end interactions (for
a simplified representation, see Figure 6), but this could easily
change with other (diene) ligands. Electronic and steric proper-
ties of the ancillary ligand (e.g., the bite-angles) could have an
influence on the absolute and relative carbene formation and
insertion barriers. The ancillary ligand could also affect the
carbene rotation barrier, which depends on the π-acceptor
strength of the diolefin as well as its steric properties. Variation
of the rotation barrier would shift the selectivity between the B
f C and C f D limiting values in Table 3.
DFT calculations suggest a chain-growth pathway involving
consecutive migratory insertion steps of carbene fragments,
formed from the carbon-bound diazo substrate by rate-limiting
dinitrogen loss, into the Rh-C bond of the growing chain. The
(21) The computational results are in agreement with stereoregulation under
chain-end control, but do not strictly exclude the possibility of other
contributing factors such as, for example, higher-order (helical) growing
chain structures influencing the stereo-electronic environment around the
A selection of changes in bond lengths and angles associated
with carbene formation (N2 loss from MDA adduct B via TS1
to carbenoid C) and migratory insertion (from C to TS3) is
active site. Additional studies are required for
a fully quantitative
understanding, taking into account the influence of all possible stereo-errors
on the carbene formation and carbene insertion steps, using COOEt groups
instead of COOMe model fragments in the calculations.
9
11638 J. AM. CHEM. SOC. VOL. 129, NO. 37, 2007

Rh-Mediated Polymerization of Carbenes
A R T I C L E S
Table 4. Tabulated 1H NMR Chemical Shift Values (δ in ppm) of Compounds 6 and 7
COO), 2.32 (m, 4H, cod-CH2), 1.71 (dd, 3JHH ) 7.5 Hz, 4H, cod-CH2).
13C NMR (75 MHz, DMSO-d6, 292 K): δ 182.3 (COO), 77 (br, cod-
CH), 45.8 (gly-CH2), 30.1 (cod-CH2). Elemental analysis for C10H16-
NO2Rh: calcd C 42.12, H 5.66, N 4.91; found C 41.98, H 5.62, N 4.83.
[(L-Prolinate)RhI(2,5-norbornadiene)] (6). A solution of proline
(125 mg, 1 mmol) and sodium hydroxide (43 mg, 1 mmol) in methanol
(5 mL) was added to a brown suspension of [{RhI(2,5-norbornadiene)-
(µ-Cl)}]2 (250 mg, 0.5 mmol) in methanol (5 mL). The obtained clear-
brown solution was stirred for 1 h at room temperature. The solvent
was removed in vacuo, and the product was dissolved in dichlo-
romethane (15 mL). The mixture was filtrated, and subsequently the
solvent was evaporated. The product was obtained as a yellow powder
(264 mg, 0.8 mmol, 80%).
1H NMR (500 MHz, CDCl3, 298 K): δ 4.02, 3.98 (br s, 4H, nbd-
CH olefinic), 3.90 (s, 2H, nbd-CH), 3.70 (m, 1H, CH-COO), 2.86-
2.73 (m, 2H, CH2-NH), 2.19-2.12 (m, 1H, CH2-CH), 2.05-1.91 (m,
2H, CH2-CH2-CH-COO), 1.69-1.61 (m, 1H, CH2-CH2-CH), 1.25 (s,
2H, nbd-CH2). 13C NMR (125 MHz, CDCl3, 298 K): δ 184.99 (COO),
62.51 (COO-CH), 61.69 (nbd-CH2), 51.02 (nbd-CH non-olefinic), 49.13
(NH-CH2), 30.13 (COO-CH-CH2), 25.72 (NH-CH2-CH2) (olefinic
carbon signal is probably too broad to be observed). Elemental analysis
for C12H16NO2Rh‚0.2CH2Cl2: calcd C 44.54, H 5.04, N 4.24; found C
44.79, H 5.13, N 4.37. NMR reveals inclusion of ∼0.2 equiv of CH2-
Cl2 in the crystal lattice.
[(L-Prolinate)RhI(endo-dicyclopentadiene)] (7). A solution of
proline (115 mg, 1 mmol) and sodium hydroxide (40 mg, 1 mmol) in
methanol (5 mL) was added to a suspension of [{RhI(endo-dicyclo-
pentadiene)(µ-Cl)}]2 (270 mg, 0.5 mmol) in methanol (5 mL). The
obtained clear orange-brown solution was stirred for 1 h at room
temperature. The solvent was removed in vacuo, and the product was
dissolved in dichloromethane (20 mL). The mixture was filtrated, and
subsequently the solvent was evaporated. The product was obtained as
an orange-yellow powder (244 mg, 0.6 mmol, 65%).
1H NMR (500 MHz, CDCl3, 298 K) (mixture of four isomers: 7a,
7a′, 7b, and 7b′): δ 6.39 (s, 1H, CgH of 7a′), 6.18 (s, 1H, CgH of 7a′),
5.50 (s, 1H, CgH of 7a), 5.18 (s, 1H, CgH of 7b), 4.98 (s, 1H, CbH of
7b), 4.92 (overlapping s, 2H, CbH of 7a and CaH of 7b), 4.81 (s, 1H,
CaH of 7a), 4.70 (s, 1H, CfH of 7b′), 4.67 (s, 1H, CfH of 7a′), 4.52 (s,
1H, CaH of 7b′), 4.43 (m, 1H, NH), 4.22-4.16 (m, 3H, NH), 4.04
(overlapping s, 4H, CaH, and CbH of 7a′, CbH of 7b′, and CfH of 7b),
3.81 (s, 1H, CfH of 7a), 3.76, 3.69 (2× m, 4H, COO-CH), 3.33 (s, 1H,
ChH of 7a), 3.29 (overlapping s, 3H, ChH of 7a′, 7b, and 7b′), 3.05
(overlapping s, 4H, 4× CiH), 3.24, 3.12, 2.99, 2.86, 2.67, 2.62 (6× m,
computational results explain the formation of syndiotactic
polymers via chain-end control,21 modulated by carbene rotation.
These new insights provide new tools for catalyst optimization
to further improve this novel polymerization reaction.
Experimental Section
General Procedures. All manipulations were performed under an
argon atmosphere using standard Schlenk techniques. Methanol and
dichloromethane distilled from calcium hydride and toluene distilled
from sodium were used for metal complex synthesis. [{RhI(1,5-
cyclooctadiene)(µ-OAc)}]2 (1),22 [{RhI(2,5-norbornadiene)(µ-Cl)}]2,23
and [{RhI(endo-dicyclopentadiene)(µ-Cl)}]224 were prepared according
to literature procedures. The syntheses and catalytic activity of
[(picolinate)RhI(1,5-cyclooctadiene] (3), [(quinolinate)RhI(1,5-cyclooc-
tadiene)] (4), and [(L-prolinate)RhI(1,5-cyclooctadiene)] (5), have been
reported previously.12 All other chemicals were purchased from
commercial suppliers and used without further purification. NMR
spectroscopy experiments were carried out on a Varian Inova 500
1
spectrometer (500 and 125 MHz for H and 13C, respectively) or a
1
Varian Mercury 300 spectrometer (300 and 75 MHz for H and 13C,
respectively). Assignment of the signals was aided by COSY, NOESY,
13C HSQC, and APT experiments. Solvent shift reference for 1H NMR
spectroscopy: CDCl3: δH ) 7.26, DMSO-d6: δH ) 2.50. For 13C NMR
spectroscopy: CDCl3: δC ) 77.0, DMSO-d6: δC ) 39.43. Abbrevia-
tions used are: s ) singlet, d ) doublet, dd ) doublet of doublets, t
) triplet, q ) quartet, m ) multiplet, br ) broad, cod ) 1,5-
cyclooctadiene, gly ) glycinate, nbd ) 2,5-norbornadiene, dcp )
dicyclopentadiene, pro ) prolinate. Elemental analyses (CHN) were
performed by the Kolbe analytical laboratory in Mu¨lheim an der Ruhr
(Germany).
[(Glycinate)RhI(1,5-cyclooctadiene)] (2). A solution of [{RhI(1,5-
cyclooctadiene)(µ-OAc)}]2 (1) (200 mg, 0.4 mmol) in toluene (10 mL)
was added to a suspension of glycine (56 mg, 0.7 mmol) in methanol
(6 mL). The yellow solution was stirred at room temperature for 20 h.
Subsequently, volatiles were evaporated, resulting in a yellow solid.
The product was recrystallized from hot methanol, affording yellow
crystals (76 mg, 0.3 mmol, 38%).
3
1H NMR (300 MHz, DMSO-d6, 292 K): δ 3.97 (t, JHH ) 6.5 Hz,
3
2H, NH2), 3.90 (br s, 4H, cod-CH), 3.03 (t, JHH ) 6.5 Hz, 2H, CH2-
(22) Sheldrick, W. S.; Gu¨nther, B. J. Organomet. Chem. 1989, 375, 233-243.
(23) (a) Chatt, J.; Venanzi, L. M. J. Chem. Soc. 1957, 4735. (b) Abel, E. W.;
Bennett, M. A.; Wilkinson, G. J. Chem. Soc. 1959, 3178.
(24) Saeed, I.; Shiotsuki, M.; Masuda, T. Macromolecules 2006, 39, 8977.
9
J. AM. CHEM. SOC. VOL. 129, NO. 37, 2007 11639

A R T I C L E S
Jellema et al.
Table 5. Tabulated 13C NMR Chemical Shift Values (δ in ppm) of Compounds 6 and 7
8H, NH-CH2), 2.42, 2.36, 2.31 (overlapping s, CdH and CeH), 2.28,
2.25, 2.21 (3× s, CcH2), 2.18-1.88 (overlapping m, CjH2 and COO-
CH-CH2), 1.85-1.77 (m, CcH2 and COO-CH-CH2-CH2), 1.75-1.71
(m, CjH2), 1.67-1.53 (COO-CH-CH2-CH2). 13C NMR (125 MHz,
CDCl3, 298 K): δ 184.65, 184.33, 183.86 (3× s, COO), 102.28 (d,
1JRh-C ) 12 Hz, CgH of 7a′), 101.82 (br s, CgH of 7b′), 90.52 (d, 1JRh-C
) 13 Hz, CgH of 7b), 89.59 (d, 1JRh-C ) 13 Hz, CgH of 7a), 87.59 (d,
columns in series and a Shimadzu RID-10A refractive index detector,
using dichloromethane as mobile phase at 1 mL/min and T ) 40 °C.
Polystyrene standards in the range of 760-1880000 g/mol (Aldrich)
were used for calibration. Previously, SEC-MALLS experiments
showed that the values obtained by calibration against polystyrene
standards slightly underestimate the molecular weights (Mw) and
somewhat overestimate the polydispersities (Mw/Mn) of poly(ethyl
2-ylidene-acetate).12
1
1JRh-C ) 11 Hz, CaH of 7a), 87.08 (d, JRh-C ) 15 Hz, CaH of 7b),
85.22 (d, 1JRh-C ) 11 Hz, CbH of 7a), 84.18 (d, 1JRh-C ) 11 Hz, CbH
DFT Calculations. The geometry optimizations were carried out
with the Turbomole program25a,b coupled to the PQS Baker optimizer.26
Geometries were fully optimized as minima or transition states at the
b3-lyp level27 using the polarized triple-ê TZVP basis25c,f (small-core
pseudopotential25c,e on Rh or Ir). All stationary points were characterized
by vibrational analysis (numerical frequencies); ZPE and thermal
corrections (entropy and enthalpy, 298 K, 1 bar) from these analysis
are included. Energies are reported both as enthalpies and free energies
(Table S1). Optimized geometries, visualized with the PLATON28
program (rendered with POVRAY), are shown in Figures S2-S21 in
the Supporting Information.
of 7b), 78.48 (d, 1JRh-C ) 12 Hz, CaH of 7a′ and 7b′), 75.94 (d, 1JRh-C
1
) 14 Hz, CfH of 7a′), 75.14 (br s, CfH of 7b′), 73.59 (d, JRh-C ) 13
1
Hz, CbH of 7a′), 73.22 (br s, CbH of 7b′), 65.78 (d, JRh-C ) 16 Hz,
CfH of 7b), 64.84 (d, 1JRh-C ) 16 Hz, CfH of 7a), 63.77, 63.47, 63.41
(3× s, COO-CH), 59.81, 59.52, 59.35 (3× s, CjH2), 56.91 (br s), 56.48,
56.12, 55.61, 55.18, 54.33, 54.15 (br s) (7× s, ChH and CiH), 49.64,
49.42, 48.93 (3× s, NH-CH2), 44.82, 44.54, 44.48, 44.42, 44.35, 44.24
(6× s, CdH and CeH), 34.84, 34.73, 34.60, 34.39 (br) (4× s, CcH2),
30.23, 29.89, 29.62 (3× s, COO-CH-CH2), 25.35, 25.02, 24.98, 24.82
(br) (4× s, COO-CH-CH2-CH2). Elemental analysis for C15H20NO2-
Rh‚0.25CH2Cl2: calcd C 48.80, H 5.52, N 3.71; found C 48.81, H
5.57, N 3.74. NMR reveals inclusion of ∼0.25 equiv of CH2Cl2 in the
crystal lattice (Tables 4 and 5).
Acknowledgment. This work was supported by the Nether-
lands Organization for Scientific Research (NWO-CW), the
University of Amsterdam, and the Canada Foundation for
Innovation (CFI Project 10831). We thank Prof. P. W. N. M.
van Leeuwen, T. G. Bloemberg, and J. Flapper for helpful
discussions and J. A. J. Geenevasen for his assistance with the
NMR measurements.
Polymerization of Carbenes from EDA to Poly(ethyl 2-ylidene-
acetate). Ethyl diazoacetate (EDA) (2 mmol) was added to a yellow
solution of catalyst precursor (0.02 mmol of 1 and 0.04 mmol of 2-7)
in chloroform (5 mL) (2 dissolved after addition of EDA). Upon
addition, gas evolution was visible, and the color of the reaction mixture
became slightly darker. The mixture was stirred for 14 h at room
temperature. Subsequently the solvent was removed in vacuo, and
methanol was added to the oily residue. The precipitate was centrifuged
and washed with methanol until the washings were colorless. The
resulting white powder was dried in vacuo and identified as poly(ethyl
(25) (a) Ahlrichs, R; et al. Turbomole Version 5, January 2002. Theoretical
Chemistry Group, University of Karlsruhe. (b) Treutler, O.; Ahlrichs, R.
J. Chem. Phys. 1995, 102, 346-354. (c) Turbomole basis-set library;
Turbomole, Version 5, see ref 25a. (d) Scha¨fer, A.; Horn, H.; Ahlrichs, R.
J. Chem. Phys. 1992, 97, 2571-2577. (e) Andrae, D.; Haeussermann, U.;
Dolg, M.; Stoll, H.; Preuss, H. Theor. Chim. Acta 1990, 77, 123-141. (f)
Scha¨fer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100, 5829-
5835. (g) Ahlrichs, R.; May, K. Chem. Phys. 2000, 2, 943.
(26) (a) PQS version 2.4; Parallel Quantum Solutions: Fayetteville, Arkansas,
U.S.A., 2001 (the Baker optimizer is available separately from PQS upon
request). (b) Baker, J. J. Comput. Chem. 1986, 7, 385-395.
(27) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789. (b)
Becke, A. D. J. Chem. Phys. 1993, 98, 1372-1377. (c) Becke, A. D. J.
Chem. Phys. 1993, 98, 5648-5652. (d) Calculations were performed using
the Turbomole functional “b3-lyp”, which is not identical to the Gaussian
“B3LYP” functional.
(28) Spek, A. L. PLATON. A Multipurpose Crystallographic Tool; Utrecht
University: Utrecht, The Netherlands, 2003.
1
2-ylidene-acetate) (PEA) using H NMR spectroscopy (spectroscopic
data of the polymer have been reported previously).12
Experiments in which the polymerization was followed in time were
performed using identical conditions (1H NMR: 300 MHz, CDCl3, 292
K).
Polymer and Oligomer Characterization. Molecular mass distribu-
tions were measured using size exclusion chromatography (SEC) on a
Shimadzu LC-20AD system with Waters Styragel HR1, HR2, and HR4
9
11640 J. AM. CHEM. SOC. VOL. 129, NO. 37, 2007

Rh-Mediated Polymerization of Carbenes
A R T I C L E S
Supporting Information Available: 1H NMR spectrum of
the poorly defined oligomers, energy profile for the carbene
rotation, details concerning the calculation of the syndio/iso
insertion preference, energies and optimized geometries of the
minima and transitions states and complete ref 25a; zip file
containing PDB and XYZ files. This material is available free
of charge via the Internet at http://pubs.acs.org.
JA073897+
9
J. AM. CHEM. SOC. VOL. 129, NO. 37, 2007 11641