The mechanism of the catalytic functionalization of haloalkanes by carbene insertion: An experimental and theoretical study

Article abstract of DOI:10.1021/om9006888

Carbon-halogen (C-X) bonds (X = Cl, Br) can be easily functionalized with ethyl diazoacetate (N₂CHCO₂Et) in the presence of silver-based catalysts containing the Tp^xAg core (Tp^x = hydrotrispyrazolylborate ligand

Full text of DOI:10.1021/om9006888

5968 Organometallics 2009, 28, 5968–5981
DOI: 10.1021/om9006888
The Mechanism of the Catalytic Functionalization of Haloalkanes by
Carbene Insertion: An Experimental and Theoretical Study
‡
Juan Urbano, Ataualpa A. C. Braga, Feliu Maseras,* Eleuterio Alvarez,
†
,†
§
ꢀ
,‡
,‡
ꢀ
M. Mar Dıaz-Requejo,* and Pedro J. Perez*
´
‡
ꢀ
´
Laboratorio de Catalisis Homogenea, Departamento de Quımica y Ciencia de los Materiales, Unidad
ꢀ
Asociada al CSIC, Campus de El Carmen s/n, Universidad de Huelva, 21007-Huelva, Spain, †Institute
of Chemical Research of Catalonia (ICIQ), 43007 Tarragona, Catalonia, Spain, and §Instituto de
´ ꢀ
Investigaciones Quımicas, CSIC-Universidad de Sevilla, Avenida Americo Vespucio 49, 41092 Sevilla, Spain
Received August 4, 2009
Carbon-halogen (C-X) bonds (X=Cl, Br) can be easily functionalized with ethyl diazoacetate
(N2CHCO2Et) in the presence of silver-based catalysts containing the TpxAg core (Tpx =hydro-
trispyrazolylborate ligand). Polyhalomethanes are converted into products derived from the formal
insertion of the carbene CHCO2Et units into the C-X bond. In the case of monohaloalkanes
(C4-C6), cleavage of the C-X bond is observed, with formation of XCH2CO2Et and the
corresponding olefin. Experimental evidence and theoretical calculations have led to the proposal
of a novel mechanism to account for these transformations, in which the metal participates along the
pathway in all the reaction steps. Among the experimental data, the first example of a metal-induced,
asymmetric functionalization of a C-Cl bond by carbene insertion is included (ee=14 ( 2%).
Introduction
functionalization and extraction of the modified substrate is
still rare. Alternative methods are based on the use of a strong
base, in processes lacking regio- or stereochemical control.5
Very recently, Yorimitsu, Oshima. and co-workers have
reported a cobalt-based system for the regioselective dehydro-
halogenation with Grignard’s reagents.6 However, the use of
an efficient procedure that would result in the removal of the
halide from the hydrocarbon chain with high atomic economy
is still of interest and remains to be developed.
The activation of carbon-chlorine (C-Cl) bonds is a
transformation of interest since they constitute part of the
skeleton of non-environmental friendly substances such as
polyvinyl chloride (PVC) and hydrochlorofluorocarbons
(HCFCs).1 Several heterogeneous systems have been de-
scribed for the functionalization of C-Cl bonds.2 In the
homogeneous phase, most efforts have been directed to aryl
halides, a common reagent in many metal-mediated C-N
bond formation reactions.3 This is in contrast with chlor-
oalkanes, where the C-Cl bond resembles those existing in
HCFC and PVC, among others. Although the activation of
this unit with transition metal complexes has been achieved,
very often through an oxidative addition step,4 the subsequent
Half a century ago, seminal work by Urry et al.7 described
the photochemical reaction of diazomethane or methyl
diazoacetate with polyhalomethanes (Scheme 1a), leading
to products derived from the insertion of the :CHR units into
one or more C-X bonds (X = Cl, Br). Reaction of ethyl
diazoacetate with monohaloalkanes also proceeded under
irradiation conditions, although when a β-hydrogen atom
(with respect to the halogen) is present, the reaction affords a
mixture of the haloacetate and an olefin (Scheme 1b).8 In the
absence of that hydrogen, the reaction proceeds through
insertion into the C-X bond. The main drawback of these
transformations is that they occur through the intermediacy
of photolytically generated free carbenes, and therefore no
control of the selectivity can be exerted.
*Corresponding authors. E-mail: perez@dqcm.uhu.es; mmdiaz@dqcm.
uhu.es; fmaseras@iciq.es.
(1) (a) Wiersma, A.; van de Sant, E. J. A. X.; Makkee, M.; van
Bekkum, H.; Moulijn, J. A. Stud. Surf. Sci. Catal. 1996, 101, 369.
(b) Manzer, L. E.; Rao, V. N. M. Adv. Catal. 1993, 39, 329.
(2) See for example: (a) Park, K. H.; Jung, I. G.; Chung, Y. K.; Han,
J. W. Adv. Synth. Catal. 2007, 349, 411. (b) York, S. C.; Cox, D. F. J. Phys.
Chem. B 2003, 107, 5182. (c) Zhou, G.; Gellman, A. J. J. Catal. 2002, 194,
233.
(3) (a) Buchwald, S. L.; Mauger, C.; Mignani, G.; Scholz, U. Adv.
Synth. Catal. 2006, 348, 23. (b) Ley, S. V.; Thomas, A. W. Angew. Chem.,
Int. Ed. 2003, 42, 5400.
(4) (a) Vetter, A. J.; Rieth, R. D.; Brennessel, W. W.; Jones, W. D.
J. Am. Chem. Soc. 2009, ASAP doi: 10.1021/ja9002316. (b) Lobana, T. S.;
Isobe, K.; Kitayama, H.; Nishioka, T.; Kinoshita, I. Angew. Chem., Int. Ed.
2004, 43, 213. (c) Aballay, A.; Godoy, F.; Buono-Core, G. E.; Klahn, A. H.;
(5) (a) Bartsch, R. A.; Za0vada, J. Chem. Rev. 1980, 80, 453. (b) Ma, Y.;
Ramirez, A.; Singh, K. J.; Keresztes, I.; Collum, D. B. J. Am. Chem. Soc.
2006, 128, 15399.
(6) Kobayashi, T.; Ohmiya, H.; Yorimitsu, H.; Oshima, K. J. Am.
Chem. Soc. 2008, 130, 11276.
~
Oelckers, B.; Garland, M. T.; Munoz, J. C. J. Organomet. Chem. 2003, 688,
(7) (a) Urry, W. H.; Eiszner, J. R. J. Am. Chem. Soc. 1951, 73, 2977.
(b) Urry, W. H.; Eiszner, J. R. J. Am. Chem. Soc. 1952, 74, 5822. (c) Urry,
W. H.; Wilt, J. W. J. Am. Chem. Soc. 1954, 76, 2594. (d) Urry, W. H.;
Eiszner, J. R.; Wilt, J. W. J. Am. Chem. Soc. 1957, 79, 918.
(8) (a) Marchand, A. P.; Brockway, N. M. Chem. Rev. 1974, 74, 441.
(b) Marchand, A. P.; Brockway, N. M. J. Am. Chem. Soc. 1970, 92, 5801.
168. (d) Dorta, R.; Shimon, L. J. W.; Rozenberg, H.; Milstein, D. Eur. J.
Inorg. Chem. 2002, 1827. (e) Stohr, F.; Sturmayr, D.; Kickelbick, G.;
Schubert, U. Eur. J. Inorg. Chem. 2002, 2305. (f) Haarman, H. F.; Ernsting,
J. M.; Kranenburg, M.; Kooijman, H.; Veldman, N.; Spek, A. L.; van
Leeuwen, P. W. N. M.; Vrieze, K. Organometallics 1997, 16, 887.
€
r
pubs.acs.org/Organometallics
Published on Web 09/11/2009
2009 American Chemical Society

Article
Organometallics, Vol. 28, No. 20, 2009 5969
Scheme 1. Photochemical Functionalization of C-X (X=halogen)
with Diazo Compounds
Figure 1. Trispyrazolylborate ligands (Tpx) relevant to this
work.
Scheme 3. Silver-Catalyzed Functionalization of C-X(X=halogen)
with Diazo Compounds
Scheme 2. First Examples of Metal-Catalyzed Carbene
Insertion into C-Cl Bonds
In spite of the low number of contributions regarding the
reaction of haloalkanes and diazocompounds, this is an
extremely interesting transformation, which could find a
potential application for the degradation of polyvinyl chlo-
ride, PVC. Equation 1 shows a hypothetical decomposition
of such material using the methodology of carbene transfer
from ethyl diazoacetate in the presence of a metal-based
catalyst. The PVC polymer chain would undergo a formal
dehydrochlorination process in such a way that an unsatu-
rated polyolefin and ethyl chloroacetate would be the final
products. The former could be reused in different ways,
whereas the latter, ClCH2CO2Et, is a raw material employed
in chemical industry in several processes. However, the
design of the appropriate catalysts requires the knowledge
of the mechanism that governs this transformation. During
the past decade, our group has been involved in the use of
group 11 metal-based catalysts for the functionalization of
hydrocarbons, either saturated or
Late transition metal complexes are known to promote
diazo compound decomposition and subsequent transfer of
the carbene moiety to unsaturated or saturated substrates.9 In
the early 1990s, Pirrung and co-workers studied the behavior
of Rh2(OAc)4 as the catalyst for dipolar cycloaddition reac-
tions using cyclic diazo compounds. During their studies they
observed10 that the reaction of 2-diazo-1,3-cyclohexanedione
with furan could be performed in several solvents such as
THF, benzene, or fluorobenzene; however, in the case of
methylene or ethylene chloride as the solvent no reaction
occurred due to the formation of another product that was
later characterized11 as the formal HCl abstraction product
(Scheme 2a). To explain the formation of this compound, the
authors invoked the rhodium-catalyzed formation of a chlor-
onium ylide that undergoes loss of chlorine-containing moi-
eties. Further work by Muller and co-workers12 with the same
€
system led to the observation of different compounds in which
a CH2Cl group migrates from the chloronium to the vicinal
carbonyl O-atom (Scheme 2b). Migration of ArCH2 was also
observed13 by Lee and co-workers with a related system using
aryl halides ArCH2Cl as the substrate in the reaction with
2-diazo-1,3-cyclohexanedione.
unsaturated, with ethyl diazoacetate.15 We have been parti-
cularly interested in the conversion of simple hydrocarbons
with EDA, in a reaction in which the :CHCO2Et group (from
EDA) formally inserts into the C-H bond of the alkane.16-18
Among the catalysts developed for such transformation, the
[TpBr3Ag]2 (1) complex was found18 to promote high conversions
A few years ago Dias, Lovely, et al. reported14 the first
example based on this methodology of a metal-catalyzed
functionalization of a C-X bond (X=Cl, Br) using simple
ethyl diazoacetate (EDA) as the carbene source. They observed
a similar behavior to that previously reported under irradia-
tion: polyhalomethanes underwent formal insertion of the
carbene group, whereas in the case of halocyclohexane, ethyl
haloacetate and cyclohexene were obtained (Scheme 3).
ꢀ
´
(15) (a) Dıaz-Requejo, M. M.; Perez, P. J. Chem. Rev. 2008, 108,
ꢀ
3379. (b) Díaz-Requejo, M. M.; Perez, P. J. J. Organomet. Chem. 2005, 690,
ꢀ
5441. (c) Díaz-Requejo, M. M.; Perez, P. J. J. Organomet. Chem. 2001, 617,
110.
(16) (a) Dıaz-Requejo, M. M.; Belderrain, T. R.; Nicasio, M. C.;
´
ꢀ
(b) Caballero, A.; Díaz-Requejo, M. M.; Belderrain, T. R.; Nicasio, M. C.;
(9) Doyle, M. P.; McKervey, M. A.; Ye, T. Modern Catalytic
Methods for Organic Synthesis with Diazo Compounds; John Wiley &
Sons: New York 1998.
Trofimenko, S.; Perez, P. J. J. Am. Chem. Soc. 2002, 124, 896.
ꢀ
Trofimenko, S.; Perez, P. J. J. Am. Chem. Soc. 2003, 125, 1446.
(10) Pirrung, M. C.; Zhang, J.; McPhail, A. T. J. Org. Chem. 1991, 56,
6269.
(11) Pirrung, M. C.; Zhang, J.; Lackey, J.; Sternbach, D. D.; Brown,
(c) Caballero, A.; Díaz-Requejo, M. M.; Belderrain, T. R.; Nicasio, M. C.;
Trofimenko, S.; Perez, P. J. Organometallics 2003, 22, 4145.
ꢀ
(17) (a) Fructos, M. R.; Belderrain, T. R.; de Fremont, P.; Scott,
ꢀ
ꢀ
´
N. M.; Nolan, S. P.; Dıaz-Requejo, M. M.; Perez, P. J. Angew. Chem.,
F. J. Org. Chem. 1995, 60, 2112.
€
ꢀ
(12) Muller, P.; Allenbach, Y. F.; Bernardinelli, G. Helv. Chim. Acta
Int. Ed. 2005, 44, 5284. (b) Fructos, M. R.; de Fremont, P.; Scott, N. M.;
ꢀ
2003, 86, 3164.
Nolan, S. P.; Díaz-Requejo, M. M.; Perez, P. J. Organometallics 2006, 25,
(13) Lee, Y. R.; Cho, B. S.; Kwon, H. J. Tetrahedron 2003, 59, 9333.
(14) Dias, H. V. R.; Browning, R. G.; Polach, S. A.; Diyabalange,
H. V. K.; Lovely, C. J. J. Am. Chem. Soc. 2003, 125, 9270.
2237.
(18) Urbano, J.; Belderraı
´
ꢀ
n, T. R.; Nicasio, M. C.; Trofimenko, S.;
az-Requejo, M. M.; Perez, P. J. Organometallics 2005, 24, 1528.
Dı
´

5970 Organometallics, Vol. 28, No. 20, 2009
Urbano et al.
Scheme 4. Functionalization of Polyhalomethanes with EDA
Using [TpBr3Ag]2 as the Catalyst Precursora
Scheme 5. Competition Experiments with Polyhalomethanes
equimolar mixtures of dichloro- and dibromomethane were
reacted with EDA, a 7-fold excess of the product derived
from the bromo derivative was observed. This trend is in
accord with the well-known order of bond dissociation
energies (BDE) for C-Cl and C-Br bonds, the latter being
ca. 16 kcal/mol lower than the former. In another competi-
tion experiment with dichloromethane and chloroform, only
the product formed from the latter was obtained.
a Diethyl fumarateandmaleate account for the remaining initial EDA.
when employed with several plain alkanes (pentane, hexane,
2-methylbutane, 2,3-dimethylbutane, among others), using the
alkane as the reaction solvent. Given the similarity of this
complex to the Dias-Lovely catalyst14 (complex Tp(CF3)2
Ag(thf); see Figure 1 for Tpx ligands), we decided to investigate
the behavior of our [TpBr3Ag]2 catalyst in the presence of
carbon-halogen bonds. In this contribution we present the
results obtained with these silver complexes as catalysts for the
above transformations, including the first example of an
asymmetric insertion of a carbene group into the C-Cl bond.
Experimental and theoretical data have allowed the proposal
of a novel reaction mechanism that will serve as the basis for
the design of useful catalysts for these processes.
Once it was demonstrated that the TpBr3-containing silver
complex was also capable of inducing the functionalization
of C-X bonds by this methodology, we explored this feature
toward monohaloalkanes. Chloro- and bromobutane or
-hexane were employed as the substrates, again with a 1:20
ratio of [catalyst]:[EDA]. In all cases, mixtures of several
products were obtained. For example, the use of 1-chlor-
ohexane as the substrate afforded a mixture of ethyl chlor-
oacetate, 1-hexene, and several compounds derived from the
insertion of the carbene unit into the C-H bonds of
1-chlorohexane (Scheme 6). This is a notable variance with
the results reported by Dias, Lovely, et al., which did not
mention the functionalization of C-H bond with haloalk-
anes as the substrates: only the formation of the ethyl
haloacetate was mentioned,14 in spite of the reported cata-
lytic capabilities of their Tp(CF3)2-containing complex for
this transformation.19 Back to our system, the four haloalk-
anes studied displayed the same behavior using [TpBr3Ag]2 as
the catalyst (Scheme 6), the respective amounts of products
derived from C-X cleavage and the C-H functionalization
being nearly similar. It could seem that the bromoalkanes are
less reactive than the chloroalkanes, but these results just
reflect the relative reactivity of each haloalkane compared
with EDA dimerization. Intermolecular competition experi-
ments provided intractable mixtures of compounds, so we
decided to evaluate the intramolecular competition using
1-chloro-4-bromobutane as the substrate (Scheme 7). Pro-
ducts derived from the cleavage of both the C-Cl and the
C-Br bonds were observed, the ratio of products indicating
that the C-Br bond is more prone to be activated, in good
accord with its lower value of BDE. Chloro- and bromo-
cyclohexane have also been studied in this transformation
(Scheme 7). They converted mainly into a mixture of ethyl
haloacetate and cyclohexene, the bromo derivative being
more active toward the functionalization. Small amounts of
the products derived from C-H functionalization were
also detected (<10%) by GC and characterized by GC-MS
(see Supporting Information).
Results and Discussion
Catalytic Functionalization of Polyhalomethanes and
Monohaloalkanes with [TpBr3Ag]2 as the Catalyst Precursor.
A first series of experiments was carried out with [TpBr3Ag]2
as the catalyst in the reaction of ethyl diazoacetate and
several polyhalomethanes as the substrate (and solvent). A
1:20 [catalyst]:[EDA] ratio was employed in all cases
(considering TpxAg monomeric units). As shown in
Scheme 4, chloro and bromo derivatives were converted into
the corresponding products derived from the formal inser-
tion of the :CHCO2Et unit into the C-X bond. In no case did
the iodo analogues verify any reaction, all the initial EDA
being converted into a mixture of diethyl fumarate and
maleate after extended reaction times. Some decomposition
was observed, and therefore the formation of the olefins
could be due to the presence of some naked silver in the
mixture. We believe that the iodo-containing substrates
induce a certain blocking of the silver center, avoiding
further catalytic reaction. In the case of Cl- and Br-contain-
ing substrates, the conversions observed are quite similar
to those reported14 with Tp(CF3)2Ag(thf), as it could be
expected, since both catalysts have also been reported to
induce a similar activity in the C-H bond functionalization
by carbene insertion from EDA.18,19 Competition experi-
ments (Scheme 5) have also been carried out to determine the
relative reactivity of C-Cl and C-Br bonds. Thus, when
Therefore, at this stage we can conclude that the
[TpBr3Ag]2 complex catalyzes the functionalization of halo-
alkanes with a dual reactivity: polyhalomethanes CXnH4-n
(X=Cl, Br; n=2-4) can be converted into products derived
from the insertion of the carbene unit into the C-X bonds.
(19) (a) Dias, H. V. R.; Browning, R. G.; Richey, S. A.; Lovely, C. J.
Organometallics 2005, 24, 5784. (b) Dias, H. V. R.; Browning, R. G.;
Richey, S. A.; Lovely, C. J. Organometallics 2004, 23, 1200. (c) Dias,
H. V. R.; Lovely, C. J. Chem. Rev. 2008, 108, 3233.

Article
Organometallics, Vol. 28, No. 20, 2009 5971
Scheme 6. Functionalization of 1-Halohexane with EDA Using [TpBr3Ag]2 as the Catalyst Precursora
a Diethyl fumarate and maleate account for the remaining initial EDA.
Scheme 7. Intramolecular Competition Experiments with
1-Chloro-4-butane and Functionalization of Halocyclohexanes
In the case of monohalogenated substrates, with available
β-hydrogen atoms, the overall reaction supposes the cleavage of
the carbon-halogen bond and the formation of ethyl haloacetate
and an olefin. This reaction competes with the functionalization
of the C-H bond of the starting material by carbene insertion. In
all cases, the reactions are performed at room temperature, and
the yields are moderate. It is worth mentioning that we have not
optimized the conversions, at this stage, but focused in a general
screening with the aim of studying the reaction mechanism.
Figure 2. X-ray structure of the molecules of the complex
[TpBr3Ag]2, 1.
Mechanistic Studies
Ag-Ag distance falls in the range expected for the so-called
argentophilic interactions,23 weak bonding forces that are
similar in strength to hydrogen bonding. In good accord with
this, determination of the molecular weight in thf solution by
Signer’s method24 has shown that an equilibrium between
the dinuclear and monomeric species exists in solution at
room temperature, favoring the latter by a factor of 3.
Two other complexes containing the core TpxAg, with the
Tp*,Br and TpMs ligands, respectively, previously prepared in
our laboratory,18,25 corresponded with the formulations [Tp*,
BrAg]2 (2) and[TpMs2Ag3][NO3] (3). The structure of the former
is similar to that of 1 (Figure 3), with a 2.80 A separation
between both silver centers. In the case of the TpMs derivative, a
trinuclear silver structure is sustained by two TpMs ligands, the
overall positive charge being balanced with a nitrate counterion
(Figure 3). Each silver center is coordinated to two pyrazolyl
rings of different TpMs ligands. This μ3-η1:η1:η1 trinucleating
bridging mode of the TpMs structure is rare; a precedent has been
reported for the complex [TpAn2Ag3][ClO4] (An = anisyl).26
Nature of the Catalytic Species. The above results have
shown that the complex [TpBr3Ag]2 (1) can be employed as a
model to study the catalytic functionalization of haloalk-
anes. The catalyst precursor was first isolated and character-
ized with the composition [TpBr3Ag]2 1/2(acetone).18 We
3
have now fully characterized this complex by X-ray studies,
the structure of the molecule being shown in Figure 2. The
[TpBr3Ag]2 group exists in the solid state as a consequence of
the coordination of each silver ion to two pyrazolyl rings of
one TpBr3 ligand and a third pyrazolyl ring of a second TpBr3
ligand. This structure is very similar to that already reported
for [Tp*Cu]2.20 It is only worth mentioning that the distance
Ag-Ag of 2.89 A is above the sum of the covalent radii21
(1.33 A þ 1.33 A=2.66 A). The related [Tp*Ag]2 complex
displayed a 2.78 A distance,22 although with a distinct
ligand distribution around the two silver centers. This
(20) Mealli, C.; Arcus, C. S.; Wilkinson, J. L.; Marks, T. J.; Ibers,
J. A. J. Am. Chem. Soc. 1976, 98, 711.
(21) Bayler, A.; Schier, A.; Bowmaker, G. A.; Schmidbaur, H. J. Am.
Chem. Soc. 1976, 98, 711.
(24) Zoellner, R. W. J. Chem. Educ. 1990, 67, 714.
ꢀ
(25) Gomez-Emeterio, B. P.; Urbano, J.; Dıaz-Requejo, M. M.;
´
(22) Humphrey, E. R.; Reeves, Z.; Jeffery, J. C.; McCleverty, J. A.;
Ward, M. D. Polyhedron 1999, 18, 1335.
ꢀ
(23) Mohamed, A. A.; Perez, L. M.; Fackler, J. P., Jr. Inorg. Chim.
ꢀ
Perez, P. J. Organometallics 2008, 27, 4126.
(26) Humphrey, E. R.; Harden, N. C.; Rees, L. H.; Jeffery, J. C.;
Acta 2005, 358, 1657.
McCleverty, J. A.; Ward, M. D. J. Chem. Soc., Dalton Trans. 1998, 3353.

5972 Organometallics, Vol. 28, No. 20, 2009
Urbano et al.
Figure 3. X-ray structures of the molecules of the complexes [Tp*,BrAg]2, 2 (left), and [TpMs2Ag3][NO3], 3 (right).
Scheme 8. Competition Experiments Carried out with CHCl3
and CDCl3
Scheme 9. Effect of the Tpx Ligand on the Ratio of C-Cl and
C-Br Functionalization
Figure 4. Monitoring of N2 evolution in the reaction of EDA
and dihalomethanes using [TpBr3Ag]2 as the catalyst precursor.
Experimental errors are <1%. Experiments were carried out
three times.
In spite of the oligonuclear structures of 2 and 3, molecular
weight studies in solution by Signer’s method have shown the
existence of a dissociation equilibrium, similar to that reported
for complex 1. Therefore, in all cases, the oligomeric complexes
1-3 deliver in solution a certain amount of the unsaturated
fragments TpxAg that are available to react with EDA.
Kinetic and Isotopic Experiments. It is well established that
the reaction of ethyl diazoacetate with late transition metal
complexes leads to the formation of highly reactive, electro-
philic metallocarbene intermediates,27 usually this step being
rate-determining.28 In addition, previous theoretical work
from our laboratories has led to the knowledge of the
geometries and energies of the TpBr3MdC(H)CO2Me inter-
mediates (M=Cu, Ag).29 We have first carried out kinetic
measurements of the nitrogen evolution in the reactions of
EDA and dihalomethanes CH2X2 (X=Cl, Br) to assess this
proposal. First-order kinetics are observed for N2 evolution,
e.g., for EDA consumption (each molecule of EDA con-
sumed originates one molecule of N2). The kobs values for
CH2Br2 and CH2Cl2 were nearly identical (1.95 ꢀ 10-3 and
1.93 ꢀ 10-3 s-1, Figure 4). This is in good agreement with the
above statement for the metallocarbene formation as the
rate-determining step, the effect of the halomethane being
null in the reaction rate. It is worth mentioning that the
kinetic experiment is very sensitive to changes in the tem-
perature of (1 °C, due to its effect on the pressure of the
evolved nitrogen.
Competition experiments using deuterated solvents have
been carried out to determine the existence of any isotopic
effect. Thus, when equimolar amounts of chloroform and
deuterochloroform, CHCl3 and CDCl3, were employed,
a ca. 1.5:1 ratio of products derived from the insertion into
the C-Cl bond was obtained favoring the protio derivative
(27) Kirmse, W. Angew. Chem., Int. Ed. 2003, 42, 1088.
(28) See: Dıaz-Requejo, M. M.; Caballero, A.; Belderrain, T. R.;
´
ꢀ
Nicasio, M. C.; Trofimenko, S.; Perez P. J. J. Am. Chem. Soc. 2002, 124,
978, and references therein.
(29) Braga, A. A. C.; Maseras, F.; Urbano, J.; Caballero, A.;
ꢀ
Dıaz-Requejo, M. M.; Perez, P. J. Organometallics 2006, 25, 5292.
´

Article
Organometallics, Vol. 28, No. 20, 2009 5973
Figure 5. 1H NMR (400 MHz) spectra of a racemic, commercial sample of Cl2CHC(H)(Cl)CO2Et (top) and the mixture of enantiomers
obtained: AgOTf and a chiral bis(oxazoline) as the catalyst (bottom). Europium tris[3-(heptafluoropropylhydroxymethylene)-(þ)-
camphorate] was employed to resolve the mixture of enantiomers. Impurities of DEF and DEM are marked with *.
(Scheme 8a). A similar behavior was also observed with
methylene chloride, for which a 1.25 isotopic effect was
obtained (Scheme 8b). These results are rather surprising
since they support the existence of a C-H cleavage step in the
reaction mechanism, a nontrivial step when thinking in the
reaction pathway of this transformation.
depending on the catalyst employed could be interpreted as
a consequence of the effect of
Effect of the Ligand: Observation of Asymmetric Induction.
A series of reactions using complexes 1-3 to induce the
reaction of EDA and equimolar mixtures of dichloro- and
dibromobutane have been performed. In all cases, we de-
tected a mixture of products derived from the insertion of
the :CHCO2Et unit into the C-Cl or the C-Br bonds, the
latter being the major compound formed (Scheme 9). Inter-
estingly, the ratio of products depended on the catalyst
precursor employed.
the Tpx ligand in the overall insertion process. In order to
demonstrate that the silver center plays a crucial role in this
transformation, we have performed a series of experiments
with silver triflate and a chiral bisoxazoline as the catalyst
precursor, and dicloromethane as the substrate, since the
formal insertion of the carbene group into the C-Cl bond
The observance of a different distribution of products
when using mixtures of dihalomethanes as the substrate

5974 Organometallics, Vol. 28, No. 20, 2009
Urbano et al.
Scheme 10. Previously Proposed Mechanism for Haloalkane
Functionalization with Ethyl Diazoacetate (top) That Is Formally
Similar to a Stevens Rearrangement (bottom)
Scheme 11. Possible Pathway for the Silver-Catalyzed
Functionalization of C-X Bondsa
a The metal center must participate along the pathway, with no free
ylides being involved.
(Scheme 10).33 This mechanism supposes that the ylide, a metal-
free intermediate, would control the formation of the products,
with no possible influence of the metal catalyst on the selectivity of
the reaction. Such a proposal has been later applied by the same
authors to the reaction of related allyl and propargyl halides
and diazoacetates.34 However, the aforementioned informa-
tion collected with our model system is somewhat contradictory
to such a proposal. Thus, the observance of a different dis-
tribution of products when using mixtures of dihalomethanes
as the substrate depending on the catalysts employed or the
unprecedented observation of enantiomeric excesses when
employing the silver-bisoxazoline catalyst precursor cannot
be explained on the basis of a Stevens rearrangement of a free
ylide. On the other hand, the transformation must occur along
a pathway in which the metal center exerts the appropriate
influence to induce those values of regio- (with several Tpx
ligands) or enantioselection (with the bisox ligand). It is also
worth mentioning that the observance of a certain deutero
effect must be related to the existence of some C-H cleavage
step during the catalytic transformation, after the silver-car-
bene intermediate is formed. This step also defies the free-ylide-
based mechanism. Therefore, the overall transformation, i. e.,
the formal insertion of a carbene group into a C-X bond, must
therefore be accomplished throughout a more complex me-
chanism that the mere formation of an ylide10-14 or a concerted
insertion of the carbene group into such a bond, similar to the
related alkane C-H bond functionalization.29
On the basis of the available data, we believe that the
transformation described in this contribution occurs through
the intermediacy of a silver-coordinated ylide (CY) formed
from the silver metallocarbene (MC) and the haloalkane,
either polyhalomethanes or monohaloalkanes (Scheme 11).
The rate-determining step is the formation of the silver-
carbene from the [LAg] core and EDA (L = Tpx, bisox).
The coordinated ylide CY subsequently undergoes further
rearrangement, to yield the observed products. Although
the final steps of this mechanism cannot be proposed
from experimental data, it is worth mentioning that (i) the
routes in Scheme 11 would explain the asymmetric induction
by the metal catalyst during the formation of the product
derived from the formal insertion of the CHCO2Et and (ii) the
missing steps should reproduce the observation of kinetic
effects in the experiment carried out with deuterated sub-
strates. The low enantiomeric excess might not arise from
generates a stereogenic center (eq 2) following the procedure
described by Ma and co-worker.30 Although the transforma-
tion is quite slow (50% of EDA is consumed after 36 h),
a ca. 14% of enantiomeric excess has been observed by NMR
studies using europium salts as discriminating reagent
(Figure 5). At this stage, this result has little synthetic
relevance in terms of development of an efficient enantiose-
lective process. However, it is remarkable in terms of provid-
ing an unambiguous indication that the silver center, now in
a chiral environment, is responsible for exerting a certain
asymmetric induction. The involvement of free versus com-
plexed ylides is a rather subtle question that has been
thoroughly discussed by Hodgson some years ago.31,32 Sup-
posedly, if the coordinating metal carries chiral ligands, one
should observe enantioselectivity if the ylide is complexed,
but no enantioselectivity in the absence of complexation.
Free Ylide versus the Metal-Based Mechanism. The pre-
vious data can be summarized in the following: (i) the first
step in this transformation is the formation of a silver-
carbene intermediate, from the reaction of TpxAg and ethyl
diazoacetate (EDA); (ii) the interaction of this intermediate
with the haloalkanes does not affect the reaction rate; (iii) a
process involving C-H cleavage must occur in order to
explain the observance of kinetic effects with deuterated
substrates; (iv) the formation of the products must occur in
the coordination sphere of the metal complex inferred from
the observation of different ratio products when varying the
Tpx ligand and, more interestingly, from the induction of a
certain asymmetry when employing a silver-bisoxazoline
catalyst precursor.
Seminal work on this reaction10-12 invoked the formation of
metal-free ylides as intermediates (Scheme 2). In the recent
contribution using the silver catalyst,14 Dias and Lovely pro-
posed a similar mechanism in which an ylide is formed from the
interaction of the silver-carbene intermediate and the halo-
methane. The free ylide subsequently undergoes a rearrange-
ment that actually corresponds to a Stevens rearrangement
(30) Ma, S.; Wu, S. New J. Chem. 2001, 25, 1337.
(31) Hodgson, D. M.; Pierard, F. Y. T. M.; Stupple, M. A. Chem.
Soc. Rev. 2001, 30, 50.
(32) (a) Hodgson, D. M.; Stupple, M. A.; Johnstone, C. Chem.
Commun. 1999, 2185. (b) Hodgson, D. M.; Stupple, M. A.; Pierard,
F. Y. T. M.; Labande, A. H.; Johnstone, C. Chem.;Eur. J. 2001, 7, 4465.
(c) Hodgson, D. M.; Labande, A. H.; Pierard, F. Y. T. M. Synlett 2003, 59.
(d) Hodgson, D. M.; Labande, A. H.; Pierard, F. Y. T. M.; Castro, M. A. E.
J. Org. Chem. 2003, 68, 6153.
(33) Smith, M. B; March, J. Advanced Organic Chemistry, 6th ed.;
Wiley-Interscience: New York.
(34) Krishnamoorthy, P.; Browning, R. G.; Singh, S.; Sivappa, R.;
Lovely, C. J.; Dias, H. V. R. Chem. Commun. 2007, 731.

Article
Organometallics, Vol. 28, No. 20, 2009 5975
Scheme 12. Operating Mechanism for Alkane C-H
Functionalization with Ethyl Diazoacetate29
metal involvement but from other less usual processes, for
example, selective destruction by the catalyst of one of the
enantiomers of the product. In view of the novelty of the
mechanism that does not follow the previous proposals based
on free ylides, we decided to complete this work from a
theoretical point of view in order to propose a complete
pathway that accounts for the experimental data. This com-
putational study will provide, together with the available
experimental evidence, a complete mechanistic picture of the
process at hand.
Theoretical Calculations. a. Polyhalomethanes. In order
to collect information relevant to the mechanism governing
the transformation described above, computational DFT
studies at the B3LYP level have been carried out on the
mechanism of the reactions for both the polyhalomethanes
and the monohaloalkane systems. For the polyhalomethane
systems, the main purpose was to find a general mechanism
that could accommodate the experimental data presented
above. The two simplest mechanistic alternatives would be
the Dias-Lovely mechanism, actually a Stevens rearrange-
ment, outlined above (Scheme 10)14 and the direct σ bond
cleavage process analogous to that previously reported for
C-H activation (Scheme 12).29 Although only the second one
is compatible with the observed asymmetric induction, none of
them can easily explain the presence of isotope effects in the
CHCl3/CDCl3 and CH2Cl2/CD2Cl2 competition experiments.
We have studied the reaction of [TpBr3AgdC(H)CO2Me]
(R1) with the three possible polychloromethane systems
CH2Cl2 (RA), CHCl3 (RB), and CCl4 (RC). As will be shown
below, the mechanisms are quite complex in terms of the
presence of intermediates and transition states, and because
of this, we will use a systematic labeling system. We will use a
unique label for each process considered, A for the particular
case of dichloromethane, and classify the structures as
reactants (R), products (P), intermediates (I), or transition
states (TS). Intermediates will be numbered consecutively as
they appear in a given mechanism, and transition states will
be labeled according to the species they connect.
Figure 6 presents schematically the overall mechanism for
processes A, B, and C corresponding to the transition metal-
catalyzed C-X activation in CH2Cl2 (RA), CHCl3 (RB), and
CCl4 (RC), respectively. The corresponding energetics are
collected in Table 1. The overall mechanism is similar, but
the number of intermediates changes. The simplest case is
that of dichloromethane, A. The initial approach of the
reactants produces an ylide intermediate, AI1, where the
chlorine binds to the carbenic center, which remains attached
to the metal center. A reorganization then takes place; the
ylide ligand retains its general shape, but ends up attached to
the metal through its carbonylic oxygen, in AI2. For dichlor-
omethane, the C-Cl activation takes place directly from
AI2, resulting in intermediate AI5, where the formal inser-
tion has already taken place. The movement of the product
away from the silver center results in the products of this
process, [TpBr3Ag] (P1) and ClCH2CHClCO2Me (PA).
Figure 6. Computed mechanism for the C-Cl activation in
polyhalomethane systems CH2Cl2, CHCl3, and CCl4.
Table 1. Computed B3LYP Energetics (kcal/mol) for Species
Involved in the C-Cl Activation in Polyhalomethane Systems
CH2Cl2, CHCl3, and CCl4
reaction A
reaction B
reaction C
R1 þ
I1
I2
I3
I4
0.0
-5.3
-1.1
0.0
-1.1
0.0
-3.4
-12.1
-15.2
-5.7
-76.3
-63.9
-1.9
-6.5
-12.7
I5
-77.2
-64.1
4.9
-77.1
-64.7
1.9
P1 þ
TS1-2
TS2-5
TS2-3
TS3-5
TS3-4
TS4-5
10.7
6.4
3.6
-1.6
-2.3
4.9
Figure 7 presents some selected structures involved in
reaction A. The initial intermediate AI1 clearly has an ylide
nature, the chlorine interacting with two carbon atoms. The
C-Cl distances are 1.870 A for the methyl carbon and 1.916
A for the carbene carbon. These distances have to be
compared with the 1.790 A C-Cl distance computed for
free dichloromethane. This strong interaction weakens
coordination of the carbene to silver; the Ag-C distance is
2.213 A in AI1, to be compared with the 2.051 A in the
starting complex R1. AI1 is significantly different from the
adducts we had previously reported between the same silver
carbene complex and alkanes.29 AI1 does not directly under-
go C-Cl cleavage, but evolves to another intermediate, AI2,
where the ylide coordinates silver through the carbonyl
oxygen. AI2 has a relative energy of -1.1 kcal/mol, compar-
able to that of AI1, and they are connected through the
transition state ATS1-2 with a low relative energy of
4.9 kcal/mol. AI2 undergoes a C-Cl cleavage step through
transition state ATS2-5, also shown in Figure 7. This
transition state is quite strained and is the rate-determining
step of the whole process (referred to the metallocarbene

5976 Organometallics, Vol. 28, No. 20, 2009
Urbano et al.
Figure 7. Computed structures for intermediate AI1 and transition state ATS2-5. Selected distances are given in A.
Figure 8. Computed structures for intermediate BI3 and transition state BTS2-3. Selected distances are given in A.
ylide proposed above, or alternatively through an intra-
intermediate). ATS2-5 is 10.7 kcal/mol above the reactants.
In fact, it involves an inversion at the carbon center and thus
goes against the rules of orbital symmetry conservation.
These rules apply however only to concerted mechanisms,
and we see this process rather as an arrested dissociation; the
carbene fragment goes away from the metal and then it
returns. Its energy is still relatively low, likely because of
the high exothermicity of the process. The following inter-
mediate, AI5, is already 77.2 kcal/mol below the reactants.
A computational estimation of the isotope effect associated
with ATS2-5 produces a value of 1.35:1, in reasonable
agreement with the experimental measurement of 1.25:1.
No bond involving a hydrogen atom is cleaved at this step,
but the movement of the CH2Cl unit in the reaction coordi-
nate involves the hydrogen centers. It is difficult to give a
qualitative explanation for the sign of this isotope effect, but
this is not surprising given the complexity of the transition
state. The good agreement between experimental and theo-
retical values provides in any case strong support to the
computed mechanism.
molecular activation and reduction elimination. This possi-
bility was investigated for reaction A. A putative intermedi-
ate, altAI1, where there is an Ag-Cl bond (2.365 A) resulting
from the addition of the C-Cl bond to the AgdC double
bond, is certainly quite stable, 35.2 kcal/mol below the
separated reactants. However, there is no low-energy path
connecting it to the reactants. We could not locate the
transition state for the direct connection, but we did estimate
a lower limit for the barrier through constrained geometry
optimizations. The late stages of the CH2Cl2 approach to the
reactant were considered in three steps, with the Cl-Ag
distance frozen at 2.6, 2.5, and 2.4 A in structures altAaux1,
altAaux2, and altAaux3. These calculations proved that
approach of a chlorine atom to the metal is not sufficient
to detach the chlorine from the alkane. Even in altAaux3
(available in the Supporting Information), with a Ag-Cl
distance of 2.400 A, the elongation of the C-Cl bond was
very small (1.806 A). Even more significant were the energies
of altAaux1, altAaux2, and altAaux3 with respect to the
reactants. They were 10.8, 18.1, and 22.2 kcal/mol, respec-
tively. The initial point of this alternative approach,
altAaux1, has an energy already higher than that of the
A completely different alternative to the mechanism just
discussed could be envisaged where the chloride coordinates
initially at silver, and then this complex reorganizes to the

Article
Organometallics, Vol. 28, No. 20, 2009 5977
Figure 9. Computed energy profile for C-Cl activation in CH2Cl2 with (path A) and without (path D) involvement of the silver
complex throughout the whole process.
rate-limiting transition state ATS2-5 of the favored path-
way. It thus seems unlikely that a low-energy path can be
found where the haloalkane attacks first the metal center
through one of its halogen atoms.
more hindered position, therefore making scrambling un-
likely. BI3 is quite stable (12.7 kcal/mol below reactants,
Table 1) and is reached through transition state BTS2-3,
which has the highest energy (6.4 kcal/mol) in this profile.
Our calculations found out a computed isotope effect of 1.40
for this step, in reasonable agreement with the experimen-
tally measured value of 1.5. BI3 evolves toward products
through transition state BTS3-5 (energy 3.6 kcal/mol)
where the CCl2 carbene unit inserts into a C-H bond.
Another difference between reactions A and B is in the
relative energies of intermediates I1 and I2. I1 is 4.2 kcal/mol
more stable in case A, but the situation is inverted in cases B,
with I2 being 5.4 kcal/mol more stable. The reason can be
understood by analyzing the structures reported in the
Supporting Information. The change in bond distances from
I1 to I2 indicates an increased transfer of a chloride from the
Process B involves CHCl3 (RB) and is also summarized in
Figure 6 and Table 1. The key difference with process A is
that there is no direct connection between I2 and I5. Instead,
there is a new intermediate, BI3, and the corresponding
transition states BTS2-3 and BTS3-5. The geometries of
BI3 and BTS2-3 are presented in Figure 8. BI3 is an unusual
structure, which may better be described as a complex with a
C(O)(OMe)(CH2Cl) unit coordinated to the silver complex
plus a free CCl2 carbene. This description is nevertheless
simplistic, as the CCl2 unit stays much closer (by 1.5 A) to
one of the two adjacent C-H bonds, which are thus far from
equivalent. Moreover the C-H bond further away is in a

5978 Organometallics, Vol. 28, No. 20, 2009
Urbano et al.
haloalkane to the carbene. As mentioned above,
methyl-carbon and carbene-carbon distances for inter-
mediate AI1 are 1.870 and 1.916 A, respectively. The values
evolve to 2.010 and 1.686 A, respectively, in AI2. The trend is
the same for BI1 and BI2, with corresponding values of
1.940, 1.876 and 2.283, 1.668 A, respectively. A chloride
fragment thus moves from the haloalkane to the carbene,
and the resulting haloalkyl fragment C4-xClx-2 ends up in a
carbocationic form. The more substituted R-halogeno car-
benium ions are known to be more stable,35 and this explains
the increased stability of I2 in the case of reaction B.
The results corresponding to system C, consisting of
[TpBr3AgdC(H)CO2Me] plus CCl4, are also included in
Figure 6 and Table 1, as well as in the Supporting Informa-
tion, but will not be discussed here in detail. We will just
mention that a new intermediate, CI4, appears in the path
and that the overall barrier is also low, with the highest
energy being the 4.9 kcal/mol for CTS4-5.
Reaction A was also evaluated computationally for the
cases of CH2Br2 and CH2I2. The mechanism is very similar to
that of the chlorine case, the rate-determining transition
states being BrATS2-5 and IATS2-5. The structures are
given in the Supporting Information. The energies of
BrATS2-5 and IATS2-5 relative to the corresponding
reactants are 9.8 and 9.5 kcal/mol, respectively. Both values
are lower than the 10.7 kcal/mol reported above for the
chlorine case. This result fits well with our experimental
observation of an easier reaction for bromoalkanes than
chloralkanes and is likely explained by the diminution in the
C-X bond dissociation energies when X changes from
chlorine to bromine and to iodine.
We completed this part of the theoretical study on the
polyhalomethane systems by analyzing the mechanism
previously proposed by Dias and Lovely,14 where the ylide
is detached from the metal complex before cleavage of the
C-X bond, following a Stevens rearrangement (see
Scheme 10). The study was carried out only for the CH2Cl2
system, in what we label as system D (Figure 9). The starting
point is the same as for system A, the ylide adduct AI1, with
an energy 5.3 kcal/mol below the separated reactants. After-
ward, we separated the ylide from the silver complex, result-
ing in species DI2, which rearranges through transition state
DTS2-P to produce directly the product PA. The relation-
ship, from both energetic and structural points of view
between DI2 and DTS2-P, is very similar to that between
AI2 and ATS2-5. In fact, the energetics favor even slightly
the D process in this regard: DTS2-P is 10.5 kcal/mol above
DI2, while the corresponding difference for the A process
was 11.8 kcal/mol. However, this mechanism has to be
discarded because of the relative energy of the DI2 inter-
mediate. It is 18.2 kcal/mol above the separate reactants and
23.5 kcal/mol above intermediate AI1. This brings DTS2-P
to a prohibitively high energy of 28.7 kcal/mol above the
reactants, unable to compete with the 10.7 kcal/mol of
ATS2-5. The silver fragment certainly does not seem to
enhance in any way the ylide rearrangement, but it is
necessary for the ylide formation, and there is simply no
reason for dissociation before the reaction is completed.
b. Monoalkanes. The second part of the computational
study concerns the monohaloalkane systems. In this case the
purpose was to justify the experimental observation of the
Figure 10. Computed mechanism for the dehydrohalogenation
of CH3-CH2Cl. Relative energies are indicated in kcal/mol.
simultaneous formation of different products, coming from
either dehydrohalogenation or C-H activation in different
positions of the chain. The computational study was in this
case carried out on the system [TpBr3AgdC(H)CO2Me] (R1)
plus CH3CH2Cl (RE), which is the simplest possible, and
avoids the conformational problems that would appear if
longer chains were considered. Three different processes
were considered for the system: dehydrohalogenation (E)
and C-H cleavage leading to either MeCO2CH2CH2CH2Cl
(F) or MeCO2CH2CHClCH3 (G).
The computed mechanism for dehydrohalogenation is
summarized in Figure 10. The initial steps are analogous to
those reported above for the polychloromethane systems.
The approach of chloroethane to the carbene complex
produces the ylide intermediate EI1, which is attached to
the metal through the carbon, and this reorganizes to
isomeric form EI2, where the ylide attached to the metal
through the carbonyl oxygen. Here the mechanisms deviate.
There is a new conformational rearrangement of this ylide
complex affecting the ethyl chain that produces intermediate
EI3, which undergoes β elimination through transition state
ETS3-4, which yields directly the product adduct EI4. This
contains already the units of each of the three products:
[TpBr3Ag] (P1), methyl chloroacetate (PE1), and ethylene
(PE2).
The energetics, also shown in Figure 10, deserve some
comment. The transition state ETS3-4 for β elimination has
a low relative energy of 2.1 kcal/mol below reactants, which
€
(35) Grutzmacher, H.; Marchand, C. M. Coord. Chem. Rev. 1997,
163, 287.

Article
Organometallics, Vol. 28, No. 20, 2009 5979
Figure 11. Computed structures for transition states ETS1-2 and ETS3-4. Selected distances are given in A.
Figure 12. Computed structures for transition states FTS1-2 and GTS1-2. Selected distances are given in A.
is in fact lower than that for the initial ylide rearrangement
through ETS1-2 (3.2 kcal/mol), which becomes, because of
this, the rate-determining step of the overall process (relative
to the initial metallocarbene complex). The structures of
both ETS1-2 and ETS3-4 are presented in Figure 11.
ETS1-2 corresponds to the movement of the ylide in the
coordination sphere of the metal, and ETS3-4 corresponds
to the β elimination step. ETS3-4 is only 1.0 kcal/mol above
the previous intermediate EI3. β elimination is thus very
easy, which explains the absence of formal carbene insertion
in the C-X bond when β hydrogens are available.
Processes F and G, involving insertion of the carbene into
a C-H bond, are mechanistically simpler. They take place in
one single step from the initial ylide intermediate EI1. The
corresponding transition states, FTS1-2 and GTS1-2, are
shown in Figure 12. They differ only in the carbon where the
C-H insertion is taking place. It is that where chlorine is not
attached in FTS1-2 and that with chlorine in GTS1-2. The
most relevant part concerning the placement of these struc-
tures in the overall reactivity of the system is in the energetics.
The relative energies above the reactants of FTS1-2 and
GTS1-2 are 3.1 and 2.8 kcal/mol, respectively. The values
are in fact close to that of 2.8 kcal/mol previously obtained
for C-H activation in ethane.29 The similarity to the value of
3.2 kcal/mol reported above for ETS1-2 is in any case
critical for the reactivity of CH3CH2Cl.
The system [TpBr3AgdC(H)CO2Me] plus CH3CH2Cl thus
has three possible reaction paths, which we have labeled E, F,
and G, leading to different products, and their respective
barriers are 3.2, 3.1, and 2.8 kcal/mol. This points out to a
mixture of the three products, and this is exactly what is
experimentally observed in these systems, with mixture of
hydrodehalogenation and C-H insertion products. The
computed barriers are in fact so close that it is not possible
to make a computational prediction of the exact propor-
tions, which would require the use of a more realistic model
(i.e., 1-chlorobutane instead of chloroethane) and the intro-
duction of solvation and entropic effects.
Mechanistic Proposal. The overall mechanistic picture
encompassing all considered systems is summarized in
Scheme 13, which could be also in principle be applied to
the reaction of other haloalkane systems with [TpBr3Md
CHCOOEt] complexes. The reaction starts, in all cases, with
the formation of an ylide complex where the halogen atom
binds to the carbene without losing its connection to the
alkane chain. This initial ylide can evolve directly toward
C-H insertion through a transition state similar to that
of the reaction of unsubstituted alkanes. An alternative

5980 Organometallics, Vol. 28, No. 20, 2009
Urbano et al.
Scheme 13. Overall Mechanistic Proposal for the Reactivity of
Haloalkanes CnH2nþ2-mXm with [TpBr3AgdCHCOOEt]
([Ag] = TpBr3Ag)
has been shown in other cases to be based on subtle combi-
nations of steric effects36 and is out of the scope of the present
work.
Conclusions
Silver complexes containing the TpxAg core can be em-
ployed as catalysts in the reaction of ethyl diazoacetate and
haloalkanes. There are two different types of reactivity
depending of the nature of the substrate. For polyhalo-
methanes, only the product derived from the insertion of
the CHCO2Et unit into the C-X bond is observed. In the
case of monohaloalkanes, a more complex reaction takes
place, resulting in a mixture of an alkene and ethyl haloace-
tate. Theoretical calculations and experimental studies, in-
cluding the use of a chiral silver complex to induce
enantiomeric excesses in the functionalization of CH2Cl2,
have led to the proposal of a novel mechanism in which the
metal center is involved in all the steps. Development of
related catalysts for PVC dechlorination is currently under-
way in our laboratories.
Experimental Section
General Methods. All preparations and manipulations were
carried out under an oxygen-free nitrogen atmosphere using
conventional Schlenk techniques or inside a drybox. All the
haloalkanes and the ethyl diazoacetate were purchased from
Aldrich and employed without further purification. The com-
plexes TpxAg (Tpx=TpBr3, TpMs, Tp*) were prepared according
to literature methods.18,25 NMR experiments were run in a
Varian Mercury 400 MHz spectrometer. GC and GC-MS
studies were carried out using Varian 3900 and Varian Saturn
2100 instruments, respectively.
General Catalytic Experiment. The silver catalyst was dis-
solved (0.02 mmol based on TpxAg units, 20.6 mg) in 5 mL of the
haloalkane, and ethyl diazoacetate (0.75 mmol) was added in
one portion. The flask was covered with aluminum foil, and the
mixture was stirred until all the starting diazo compound was
consumed (verified by GC). The products were identified by
NMR, GC, and GC-MS, by comparison with reported data or
commercial samples (see Supporting Information). The low
boiling point of most of the products induced loss of material
during workup. Therefore, the conversions were measured by
NMR studies of the reaction mixture, using an internal stan-
dard. In the case of CBr4 as the substrate, methylene chloride
was used as the solvent of this solid haloalkane.
For competition experiments the different substrates were
mixed before the catalyst was added. For example, 10 mmol of
CHCl3 and 10 mmol of CDCl3 were mixed at room temperature
and 0.02 mmol (290.6 mg) of TpBr3Ag was added. After 15 min
of stirring, 0.1 mmol of EDA was added in one portion, and the
mixture was stirred for 8 h. After that time, the mixture was
investigated by 1H NMR, the relative amount of insertion
products obtained from each substrate being established by
integration.
Asymmetric Induction. The catalyst precursor was prepared as
described in the literature.30 (-)-2,20-Isopropylidenebis[(4S)-4-
phenyl-2-oxazoline] (0.03 mmol, 10 mg) was dissolved in 15 mL
of dichloroethane, and 0.025 mmol of silver triflate was added to
the stirring solution. After 1.5 h the solvent was removed under
vacuum and the residue was employed as the catalyst precursor
in the reaction of methylene chloride with EDA, following the
above procedure. After 36 h of stirring, GC studies revealed
ca. 50% of EDA consumed. Volatiles were removed at 0 °C, and
pathway is that in which the initial C-bound ylide rearranges
to bind to the metal through an oxygen. This O-bound ylide
can then undergo two different processes. If β hydrogens are
available, it undergoes a low-barrier dehydrohalogenation.
If this is not possible, as is the case of halomethanes, it can
undergo a C-X insertion through a multistep process,
involving in some cases C-H cleavage steps.
It may seem surprising that C-H insertion occurs after the
ylide formation. This is however consistent with our previous
computational study on C-H insertion by the same type of
silver complexes in alkanes with no halogen substituents.29
The energies of the van der Waals adducts between alkane
and metal complex were between -0.7 and -1.0 kcal/mol
depending on the substitution at the alkane. This is much less
stable than the ylide complex EI1 reported above (-8.4 kcal/
mol), which should thus be formed quantitatively through an
essentially barrierless complex.
The above proposal supports the involvement of the metal
center in the various steps and is in contrast with the
previously proposed mechanism that is based on the forma-
tion of a free ylide.10-14 The contribution of the silver center
in this transformation is of interest when designing catalysts
for the dechlorination of the aforementioned PVC, since the
appropriate tuning of the metal center will exert a certain
effect in the reaction products.
This study cannot provide a detailed mechanism for
the asymmetric induction observed in the case of the
bis(oxazoline) ligands, which has been presented only
to prove metal involvement in the reaction. The stereo-
induction mechanism by this type of ligand, which will
replace the (tris)pyrazolyl in the silver coordination sphere,
(36) Drudis-Sole, G.; Maseras, F.; Lledos, A.; Vallribera, A.;
Moreno-Manas, M. Eur. J. Org. Chem. 2008, 5614.

Article
Organometallics, Vol. 28, No. 20, 2009 5981
the residue was purified by column chromatography (SiO2, 9:1
petroleum ether/diethyl ether as eluent) to give the product as an
oil.
associated valence double-ζ basis set for the outer ones. Polar-
ization shells, of f and d type, were added for Ag and Br,
respectively.40 The 6-31G(d) basis set was used for H, B, C, N,
O, and Cl atoms.41 The structures of reactants, intermediates,
transition states, and products were fully optimized with no
symmetry restrictions. Transition states were identified by hav-
ing one negative eigenvalue in the Hessian matrix, and their
connectivity to intermediates was checked through small step
geometry optimizations from the transition state. All reported
energies are potential energies. This method was proved to give
correct results in a previous study by our groups,29 where it was
calibrated against CCSD(T) calculations, and entropic and
solvation effects were shown to have a minor effect.
The procedure to measure ee by NMR was the following. A
stock solution of europium tris[3-(heptafluoropropylhydroxy-
methylene)-(þ)-camphorate] (75 mg in 0.5 mL of C6D6) was
prepared. A separate solution of the product was prepared by
dissolving 5 μL of the isolated oil in 0.5 mL of C6D6. After
several additions of the europium solution, the enantiomers
were resolved by adding ca. 200 μL of the Eu3þ solution to that
of the product. Integration of the resonances gave a ca. 14 (2% ee.
X-ray Studies. Experimental procedures and cif files for the
determination of the structures of complexes 1, 2, and 3 are given
in the Supporting Information. CCDC 687284, 687285, and
687286, respectively, contain the supplementary crystallo-
graphic data for the compounds described in this paper. These
data can be obtained free of charge via www.ccdc.cam.ac.uk/
conts/retrieving.html (or from the Cambridge Crystallographic
Data Centre, 12 Union Road, Cambridge CB21EZ, UK; fax:
þ44 1223-336-033; or deposit@ccdc.cam.ac.uk).
Acknowledgment. Financial support from MICINN
(Grants CTQ2008-06866-CO2-02/BQU, CTQ2008-00042/
BQU, Consolider Ingenio 2010 CSD2006-0003), the
Catalan DURSI (project 2005SGR00715), the Junta de
Andalucı
foundation is acknowledged. J.U. also thanks the Junta
de Andalucıa for a student fellowship. A.A.C.B. thanks
´
a (Grant P07-FQM 2870), and the ICIQ
Kinetic Studies. A solution of the catalyst (0.02 mmol) in
methylene chloride (4 mL) was placed into a 25 mL round-
bottom flask connected to a computer-connected pressure
gauge. Ethyl diazoacetate was added in one portion (0.5 mmol),
and the evolution of nitrogen was monitored until the pressure
was maintained constant, i.e., until all EDA was consumed.
Calculations. Calculations were performed at the DFT level
by means of the B3LYP functional37 using the Gaussian suite of
programs.38 Ag and Br atoms were described using an effective
core potential (LANL2DZ)39 for the inner electrons and its
´
the Spanish MICINN for a “Juan de la Cierva” contract.
Supporting Information Available: Full reference for the
Gaussian03 program, Cartesian coordinates, and summary of
energetics for all stationary points. This material is available free
of charge via the Internet at http://pubs.acs.org.
€
(40) (a) Ehlers, A. W.; Bohme, M.; Dapprich, S.; Gobbi, A.;
Hollwarth, A.; Jonas, V.; Kohler, K. F.; Stegmann, R.; Veldkamp, A.;
€
€
€ €
Frenking, G. Chem. Phys. Lett. 1993, 208, 111. (b) Hollwarth, A.; Bohme,
(37) (a) Lee, C.; Parr, R. G.; Yang, W. Phys. Rev. B 1988, 37, 785.
(b) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (c) Stephens, P. J.; Devlin,
F. J.; Chabalowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994, 98, 11623.
(38) Frisch, M. J. Gaussian 03, Revision C.02; Gaussian, Inc.:
Wallingford, CT, 2004.
€
M.; Dapprich, S.; Ehlers, A. W.; Gobbi, A.; Jonas, V.; Kohler, K. F.;
Stegmann, R.; Veldkamp, A.; Frenking, G. Chem. Phys. Lett. 1993, 208,
237.
(41) Francl, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley, J. S.;
(39) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284.
Gordon, M. S.; Defrees, D. J.; Pople, J. A. J. Chem. Phys. 1982, 77, 3654.