Ligand-dependent mechanistic dichotomy in iron-catalyzed allylic substitutions: σ-allyl versus π-allyl mechanism

Article abstract of DOI:10.1002/anie.200703874

(Chemical Equation Presented) Iron at the crossroads: A remarkable mechanistic dichotomy in iron-catalyzed allylic substition increases not only the scope of regioselective allylic alkylation but could also herald the development of asymmetric allylic substitution (see Scheme, MTBE = methyl tert-butyl ether). Depending on the ligand used, either a σ-or a π-allyl mechanism is observed.

Full text of DOI:10.1002/anie.200703874

Communications
DOI: 10.1002/anie.200703874
Allylic Substitution
Ligand-Dependent Mechanistic Dichotomy in Iron-Catalyzed Allylic
Substitutions: s-Allyl versus p-Allyl Mechanism**
Bernd Plietker,* AndrØ Dieskau, Katrin Möws, and Anja Jatsch
Dedicated to Prof. Wolfgang Kreiser on the occasion of his 70th birthday
À
Selective C C bond formation is vitally important in organic
chemistry. Transition-metal catalyzed allylic substitution has
evolved as a powerful tool in this field.[1] In the presence of
enantiopure transition-metal catalysts based on Pd,[2] Mo,[3]
Ir,[4] Ru[5], or Ni,[6] enantiomerically enriched products are
accessible in good yields starting from racemic material,
which is a consequence of the fluctuating character of the
intermediate p-allyl metal complexes. In the presence of
rhodium[7] or iron catalysts[8] however, a slow isomerization of
the s-allyl metal species initially formed is observed, and in
the subsequent nucleophilic substitution, retention of con-
stitution and configuration of the starting material occurs.
Hence, both procedures might be regarded as complementary
to each other.
an alternative solvent could pave the way for the allylation of
less stabilized nucleophiles, and c) the development of a
procedure for the allylic substitution by a p-allyl mechanism is
desirable and would allow the use of iron catalysts as an
alternative to the existing methods for asymmetric allylic
substitutions which use well-established metal complexes.[12]
Herein we present
a ligand-dependent mechanistic
dichotomy in iron-catalyzed allylic substitution, which not
only fulfils the requirements mentioned in (a) and (b) above,
but also allows the reaction to follow the unprecedented p-
allyl iron mechanism (c). The latter aspect represents one of
the basic requirements for the development of an asymmetric
allylic substitution.
N-Heterocyclic carbenes (NHC ligands) are amongst the
most successful ligands in transition-metal catalysis.[13] The s-
donor character of these ligands in combination with the
resulting higher nucleophilicity of the coordinated metal ions
makes these ligands an attractive choice for iron-catalyzed
allylic substitutions (Table 1).[14]
The use of an iron catalyst is attractive because of its lower
price and low toxicity.[9] Based on preliminary studies by
Roustan et al.,[10] and Xu and Zhou,[11] we recently success-
fully developed a highly regioselective iron-catalyzed allylic
alkylation[8a] and amination[8b] (Scheme 1).
The original procedure could be improved significantly.
Unlike DMF, methyl tert-butyl ether (MTBE) is more stable
toward reactive nucleophiles and therefore was the better
solvent. Furthermore, almost complete conversion was
Table 1: NHC ligands for allylic substitutions.
Scheme 1. Iron-catalyzed allylic amination and alkylation. Pip=piper-
idine.
Entry[a] Ligand[15]
Base[15]
5a/
Conv.
[%][b]
5b[b]
For synthetic application, we wished to improve our
procedure for three reasons: a) an excess of pronucleophile is
not acceptable from an economic point of view, b) the use of
1
KOtAm[d] 10:90
92
[*] Prof. Dr. B. Plietker, A. Dieskau, K. Möws, A. Jatsch
Institut für Organische Chemie
2
7
R=2,4,6-
KOtAm[d] 9:91
98
(CH3)3C6H2
Universität Stuttgart
3
4
8
9
R=2,6-(iPr)2C6H3
R=4-MeOC6H4
33:67
63:37
38
12
Pfaffenwaldring 55, 70569 Stuttgart (Germany)
Fax: (+49)711-6856-4289
E-mail: bernd.plietker@oc.uni-stuttgart.de
5
6
7
10
11
12
R=iPr
R=Ph2CH87:13
R=tBu
NaNH2 84:17
66
68
74
[**] We are grateful to the Deutsche Forschungsgemeinschaft, the
Fonds der Chemischen Industrie, the Studienstiftung des Deut-
schen Volkes (grants for K.M. and A.D.), and the Dr-Otto-Röhm-
Gedächtnisstiftung for financial support of this work.
91:9
[a] All reactions were performed on a 1-mmol scale in MTBE under an
atmosphere of N2. [b] Determined by GC. [c] Mes=2,4,6-trimethylphe-
nyl. [d] KOtAm=potassium 2-methylbutan-2-olate.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
198
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 198 –201

Angewandte
Chemie
Table 3: Influence of the nucleophile on regioselectivity.
observed using stoichiometric amounts of each reactant
(Entry 2, Table 1). Surprisingly, an inversion of the regiose-
lectivity when compared to the original reaction was
observed, leading to predominant formation of product 5b
(Entry 2, Table 1). Steric reasons seem to be responsible for
this unexpected result. Thus the use of a sterically more
hindered iPr-group in the o,o’-position of the aryl substituent
leads to a significant shift of regioselectivity toward the ipso
substitution product 5a (Entry 3, Table 1). A further increase
in the size of the substituent at nitrogen was achieved by the
introduction of sp3-hybridized carbon atoms. The increased
steric demand causes the predominant formation of product
5a (Entries 5–7, Table 1). Use of the tert-butyl group in ligand
12 led to product 5a in high selectivity (Entry 7, Table 1).[16]
The varying regioselectivity suggests a reaction path that
differs from the s-allyl mechanism. An alternative is the
unprecedented catalytic reaction with a p-allyl iron com-
plex.[17] As the constitution of the allyl carbonate should only
have a minor influence on the product distribution in this
case, the two regioisomeric carbonates 13 and 14 were
transformed into the products 15/16 in the presence of
ligand 7 and 12. The use of 7 leads to the formation of an
identical mixture of regioisomers 15 and 16 (Entries 1 and 2,
Table 2) and is comparable to the product distribution
obtained in the presence of [Pd(PPh3)4] under identical
conditions. In the presence of ligand 12 however, it is the
position of the leaving group in 13/14 which directs the
regioselective course of the reaction (Entries 3 and 4,
Table 2).
Entry[a] Ligand R1
R2
pKa[b] Product a/b[c] Yield [%][c]
1
2
7
12
9:91
91:9
79
84
CO2iBu CO2iBu
16.4
5
3
4
7
12
15:85
94:6
76
74
CO2iBu C(O)CH3 14.2 17
5
6
7
12
74:26
95:5
85
88
CO2iBu CN
13.1 18
12.0 19
11.1 20
7
8
7
12
60:40
80:20
86
87
SO2Ph CN
9
10
7
12
80:20
99:1
76
85
CN
CN
[a] All reactions were performed on a 1-mmol scale in MTBE under an
atmosphere of N2. [b] Entries 1–6 refer to the corresponding ethyl
esters.[18] [c] Yield of isolated product.
value, significant shifts in the regioselectivity were observed.
Whereas the presence of the sterically hindered ligand 12
results in the predominant formation of the ipso substitution
product (Entries 2, 4, 6, 8, 10, Table 3), the regioselective
course of the reaction in the presence of ligand 7 is directed
mostly by the acidity and nucleophilicity of the carbanion
generated in situ. Hence, the fast deprotonation of malodini-
trile and the high nucleophilicity of the anion formed (s 0.67,
N 19.36)[19] leads to a fast substitution of the allyl iron
complex. It appears that the s–p–s isomerization is not fast
enough under these conditions, and consequently the forma-
tion of the ipso substitution product 20a is favored (Entry 9,
Table 3).[20]
Table 2: Influence of the ligand on regioselectivity.
The results obtained so far are summarized in the
mechanistic model in Scheme 2. Assuming that in the
presence of ligand 7 or 12, a s-allyl iron species such as VII/
X is formed, two subsequent reactions are possible. Species
VII and X could be transformed into the desired product XIII
in a fast substitution reaction, or, if this reaction is slower and
the ligand-created steric environment tolerates a fluctuation
of the metal in the allyl terminus, the formation of the more
easily substituted s-allyl iron complex IX from VII is possible.
A planar aryl substituent in 7 could facilitate such a
fluctuation, whereas a tert-butyl group as in 12 generates
unfavorable steric interactions and thus disfavors the forma-
tion of p-allyl complex XI from X (Scheme 2).
The optimized reaction conditions have a great impact on
the reaction scope. The use of MTBE as an inert solvent
enables the use of preformed nucleophiles in the reaction.[21]
The possibilities connected with this important result are
exemplified in the reaction of azlactone 21 (Scheme 3).
Whereas under salt-free conditions almost no reaction was
observed, after deprotonation of 21, the reaction to form
allylation product 23 was successful, and furthermore the
Entry[a]
Carbonate
Ligand
15/16[b]
Yield [%][b]
1
2
13
14
7
17:83
15:85
67 (72)
63 (68)
3
4
13
14
12
91:9
12:88
71 (78)
64 (66)
[a] All reactions were performed on a 1-mmol scale in MTBE under an
atmosphere of N2. [b] Yield of isolated product; yield determined by GC
in brackets.
The s–p–s isomerization in the presence of ligand 7 might
be the consequence of a slow attack of the nucleophile at the
allyl iron intermediate. Hence, more reactive nucleophiles
should influence the regioselective course of the reaction. To
test this hypothesis, various malonic acid derivatives, differing
in nucleophilicity and acidity, were allylated (Table 3). The
allylation is broadly applicable in the presence of ligand 7 or
12. In each case the reaction proceeded with almost full
conversion. However, depending on the ligand and the pKa
Angew. Chem. Int. Ed. 2008, 47, 198 –201
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
199

Communications
Scheme 2. Mechanistic model for the ligand-dependent dichotomy.
Scheme 4. Ligand influence in allylic substitutions of 27. Reagents and
conditions: a) [Bu4N][Fe(CO)3(NO)] (2.5 mol%), ligand (2.5 mol%),
MTBE, 808C. n.d.=not determined.
isobutoxide 25 generated in situ reacted with 25 in a
subsequent transesterification/ring opening sequence to
yield isobutyl ester 26.
of a significantly improved procedure which
features the use of exact stoichiometric amounts
of the pronucleophile and a solvent change from
DMF to MTBE. The latter aspect is of great
importance, as it allows the allylation of non-
stabilized or reactive nucleophiles. Furthermore,
we were able to show that in the presence of
ligand 12, both regio- and stereoconservative
allylic substitution is possible in which the
double bond geometry stays intact. The use of
aryl-substituted ligand 7 allowed for the first
time allylation by the p-allyl mechanism, which
is complementary to that with 12. The loss of
constitutional information of the starting mate-
Scheme 3. Tandem allylic substitution and trans-
esterification of azlactones. Bz=benzoyl.
rial should, in the case of a fast s–p–s-isomer-
ization, set the stage for the successful develop-
ment of a iron-catalyzed, asymmetric, dynamic–
kinetic allylic substitution.
The consequences of the mechanistic dichotomy on
Received: August 23, 2007
Published online: November 14, 2007
stereo- and regioselectivity in allylic substitutions of 1,2-
disubstituted isomeric carbonates such as 27 were investi-
gated (Scheme 4). The allylic substition of deuterated car-
bonate 27-D occurred only in the presence of ligand 12
selectively at the deuterated carbon atom [Eq. (1), Scheme 4].
Use of the enantiomerically enriched carbonates (E)- and
(Z)-27 using ligand 12 resulted in only a small decrease in
enantioselectivity. The substitution products 28 were formed
with formal retention of the configuration and the geometry
at the double bond stayed intact.[22] In contrast, the same
reactions in the presence of aryl-substituted ligand 7 led to a
loss of constitutional and configurational information of the
carbonates 27 [Eq. (2) and (3), Scheme 4].
Keywords: allylic compounds · iron · nucleophilic substitution ·
.
regioselectivity · stereoselectivity
[1] U. Kazmaier, M. Pohlmann in Metal-Catalyzed Cross-Coupling
Reactions, Vol. 2, 2nd ed. (Ed.: A. De Meijere, F. Diederich),
Wiley-VCH, Weinheim, 2004, p. 531.
[2] B. M. Trost, M. L. Crawley, Chem. Rev. 2003, 103, 2921.
[3] a) B. M. Trost, K. Dogra, I. Hachiya, T. Emura, D. L. Hughes,
S. W. Krska, R. A. Reamer, M. Palucki, N. Yasuda, P. J. Reider,
Angew. Chem. 2002, 114, 2009; Angew. Chem. Int. Ed. 2002, 41,
1929; b) F. Glorius, A. Pfaltz, Org. Lett. 1999, 1, 141.
We have presented a ligand-dependent mechanistic
dichotomy in iron-catalyzed allylic substitutions. The use of
tert-butyl-substituted NHC ligand 12 led to the development
200
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 198 –201

Angewandte
Chemie
[4] For a review on iridium-catalyzed asymmetric allylic substitu-
tions, see: G. Helmchen, A. Dahnz, P. Dübon, M. Schelwies, R.
Weihofen, Chem. Commun. 2007, 675.
[5] a) Y. Morisaki, T. Kondo, T.-A. Misudo, Organometallics 1999,
18, 4742; b) B. M. Trost, P. Fraisse, Z. Ball Angew. Chem. 2002,
114, 1101; Angew. Chem. Int. Ed. 2002, 41, 1059; Angew. Chem.
Int. Ed. 2002, 41, 1059; c) M. D. Mbaye, B. Demerseman, J.-L.
Renaud, L. Toupet, C. Bruneau, Angew. Chem. 2003, 115, 5220;
Angew. Chem. Int. Ed. 2003, 42, 5066; d) R. Hermatschweiler, I.
[12] Reviews on the stoichiometric use of allyl iron complexes:
a) S. V. Ley, R. Liam, G. Meek, Chem. Rev. 1996, 96, 423; b) D.
Enders, B. Jandeleit, S. von Berg, Synthesis 1997, 421; c) D.
Enders, B. Jandeleit, S. von Berg, G. Raabe, J. Runsink, Organo-
metallics 2001, 20, 4312, and references therein.
[13] N-Heterocyclic Carbenes in Transition Metal Cataylsis (Ed.: F.
Glorius), Springer, Hamburg, 2007.
[14] For the influence of monodentate phosphane ligands see the
Supporting Information.
Fernandez, F. Breher, P. S. Pregosin, L. F. Veiros, M. J. Calhorda, [15] Details on the reaction optimization are provided in the
Angew. Chem. 2005, 117, 4471; Angew. Chem. Int. Ed. 2005, 44,
4397.
[6] a) K.-G. Chung, Y. Miyake, S. Uemura, J. Chem. Soc. Perkin
Trans. 1 2000, 15; b) M. T. Didiuk, J. P. Morken, A. H. Hoveyda,
J. Am. Chem. Soc. 1995, 117, 7273.
[7] D. K. Leahy, P. A. Evans, in Modern Rhodium-Catalyzed
Organic Reactions (Ed.: P. A. Evans), Wiley-VCH, Weinheim,
2005, pp. 191.
[8] a) B. Plietker, Angew. Chem. 2006, 118, 1497; Angew. Chem. Int.
Ed. 2006, 45, 1469; b) B. Plietker, Angew. Chem. 2006, 118, 6200;
Angew. Chem. Int. Ed. 2006, 45, 6053.
[9] a) Review on iron catalysis: C. Bolm, J. Legros, J. Le Paih, L.
Zani, Chem. Rev. 2004, 104, 6217; b) Review on iron-catalyzed
cross coupling: A. Fürstner, R. Martin, Chem. Lett. 2005, 34, 624.
[10] a) J.-L. Roustan, M. Abedini, H. H. Baer, Tetrahedron Lett.
1979, 20, 3721; b) J.-L. Roustan, M. Abedini, H. H. Baer, J.
Organomet. Chem. 1989, 376, C20.
Supporting Information.
[16] R. B. Bedford, M. Betham, D. W. Bruce, A. A. Danopoulos,
R. M. Frost, M. Hird, J. Org. Chem. 2006, 71, 1104.
[17] It is known from allyl iron chemistry that the initial substitution
in the starting material does not always occur strictly by an SN2’
anti-type mechanism, but can also occur by a SN2-type mecha-
nism. The subsequent reaction of the allyl iron complex with the
nucleophile would then lead to inversion of regioselectivity.
[18] Taken from the Bordwell pKa table: http://www.chem.wisc.edu/
areas/reich/pkatable/.
[19] a) H. Mayr, B. Kempf, A. R. Ofial, Acc. Chem. Res. 2003, 36, 66;
b) Taken from H. Mayrꢀs s-N-E database: http://cicum92.cup.
uni-muenchen.de/mayr/reaktionsdatenbank/.
[20] The nature of the leaving group had only a minor influence on
the regioselectivity (see Supporting Information).
[21] See Supporting Information.
[22] A similar reaction has only been known in the presence of
rhodium catalysts: B. L. Ashfeld, K. A. Miller, S. F. Martin, Org.
Lett. 2004, 6, 1321.
[11] a) Y. Xu, B. Zhou, J. Org. Chem. 1987, 52, 974; b) B. Zhou, Y.
Xu, J. Org. Chem. 1988, 53, 4421.
Angew. Chem. Int. Ed. 2008, 47, 198 –201
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
201