Iron-Catalyzed Nucleophilic Addition Reaction of Organic Carbanion Equivalents via Hydrazones

Article abstract of DOI:10.1021/acs.orglett.8b01391

Earth-abundant and well-defined iron complexes are found to be cheap and effective catalysts for a series of "umpolung" nucleophilic additions of hydrazones. The new catalytic system not only maintains the broad substrate scope of an earlier expensive ruthenium system but also attains chemoselectivity of different kinds of carbonyl groups. Furthermore, the iron catalyst enables this reaction at ambient temperature.

Full text of DOI:10.1021/acs.orglett.8b01391

Letter
pubs.acs.org/OrgLett
Cite This: Org. Lett. XXXX, XXX, XXX−XXX
Iron-Catalyzed Nucleophilic Addition Reaction of Organic Carbanion
Equivalents via Hydrazones
Chen-Chen Li, Xi-Jie Dai, Haining Wang, Dianhu Zhu, Jian Gao, and Chao-Jun Li*
Department of Chemistry and FRQNT Centre for Green Chemistry and Catalysis, McGill University, 801 Sherbrooke Street West,
Montreal, Quebec H3A 0B8, Canada
S
* Supporting Information
ABSTRACT: Earth-abundant and well-defined iron complexes
are found to be cheap and effective catalysts for a series of
“umpolung” nucleophilic additions of hydrazones. The new
catalytic system not only maintains the broad substrate scope of
an earlier expensive ruthenium system but also attains chemo-
selectivity of different kinds of carbonyl groups. Furthermore, the
iron catalyst enables this reaction at ambient temperature.
arbon-skeleton construction is the basis of organic
synthesis and the core for forming various organic
molecules, in which the nucleophilic reactions of carbanion
and its equivalents play a vital role.1 In classic carbanion
addition reactions, stoichiometric amounts of metals are
usually needed (Scheme 1, a).1 Inspired by the classical
to develop earth-abundant metal-catalyzed related addition
reactions to make the carbanion chemistry safer, cheaper, and
more practical.
C
Iron, an inexpensive and earth-abundant metal, is one of the
ideal catalysts for organic reactions. So far, a considerable
number of iron-catalyzed organic reactions have been studied
by chemists, such as free radical reactions, cross-coupling
reactions, C−H functionalizations, Lewis acid−base reactions,
etc.9 Until now, however, reports about an iron-based
Grignard-type reaction have been limited, among which a
stoichiometric amount of metal still cannot be avoided.9a−h
Early in 1961, a series of iron−bisphosphine complexes were
reported,10 which could coordinate and interact with hydrazine
derivatives with the help of a base as was shown by the later
literature.11 Thus, we hypothesized that an iron−phosphine
complex might also catalyze the nucleophilic addition of
hydrazone (as carbanion equivalents) for different kinds of
transformations of electrophiles.
Scheme 1. Development of Carbanion Chemistry
Moreover, utilizing such complexes to catalyze organic
reactions is of great significance because most earlier studies, as
we mentioned above, on related iron−bisphospine complexes
were limited to stoichiometric transformations of the
complexes themselves instead of being applied to catalyze an
organic reaction.
To proceed, several key challenges were envisioned: first,
iron is a first-row transition metal which is a comparatively
harder Lewis acid and is more likely to coordinate with harder
ligands such as water, hydroxide, and nitrogen-containing
compounds.12 Thus, hydrolysis and hydrazinolysis of the
catalyst may occur in the reaction system and should be taken
into consideration. Second, iron has two commonly stable
oxidation states, iron(II) and iron(III), which can easily
interchange.13 This implies that an iron complex might interact
with other components via oxidation/reduction processes in
Wolff−Kishner reduction,2 we have recently found that
hydrazones, readily generated in situ from aldehydes and
hydrazine, can serve as an alkyl carbanion equivalent3 to
undergo nucleophilic additions.4 In these cases, the hydrazone
was established as an organic carbanion equivalent based on
just a catalytic amount of metal catalyst rather than a
stoichiometric amount of metal reagent. To date, a range of
such hydrazone reactions including carbonyl additions,5 imine
additions,6 conjugate additions,7 and CO28 addition have been
realized with ruthenium catalysts (Scheme 1, b). Nevertheless,
as noted earlier,5 the required catalyst is based on a scarce and
precious ruthenium. Therefore, major efforts have been made
Received: May 2, 2018
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.8b01391
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
the reaction and affect the main transformation. With all these
regards, herein, we report a well-defined iron−phosphine
complex as an effective catalyst for a series of “umpolung”
nucleophilic additions of hydrazones to carbonyls, imines, and
Michael acceptors, which can even proceed at room temper-
ature with a broad substrate scope (Scheme 1, c).
We initiated our research by selecting different iron salts
together with 1,2-bis(dimethylphosphino)ethane (dmpe) as
ligand, K3PO4 as base, and THF as solvent at 100 °C. A trace
amount of the desired product was obtained with FeF2, FeCl3
(8% yield), and FeCl2 as catalyst (Table 1, entries 1−5). With
(15%), which may be due to the inhibition of side reactions
such as hydrolysis of the catalyst. It is also worth noting that at
60 °C the yield was even lower than the one at 100 °C. This
might be because, at 60 °C, the hydrolysis and other side
reactions are still not efficiently inhibited (as at room
temperature), whereas the reactivity of the main reaction is
lower than at 100 °C. In order to facilitate the ligand exchange
step on iron involved in the catalytic process, CsF was added
into the reaction system.14 Delightfully, we found that addition
of 50% of CsF to the reaction system increased the yield to
72% (entry 13). The test of a variety of bases showed that
K3PO4 provided the best yield of 72% (entries 13−16). A close
analysis of the reaction mixture found that significant
quantities of both starting materials remained unconverted,
which suggested the deactivation of the catalyst during the
reaction process. According to the procedure of preparing a
hydrazone solution (see SI), a trace amount of water may
remain in the reaction system and destroy the iron catalyst
slowly via hydrolysis. To test this hypothesis, we added some 4
Å molecular sieves directly to the reaction system under the
optimized conditions. Indeed, the yield was increased to 90%,
and the starting materials were nearly consumed (Table 1,
entry 17). We also tested a series of solvents and found that
the reaction worked well in some polar solvents such as THF
and DMSO (Table 1, entries 17−20). To simplify the reaction
system, we further considered whether the 4 Å molecule sieves
could be used to pretreat the hydrazone solution before
running the reaction (see SI for details). By doing this, the
yield of the product was further increased to 97% (entry 21).
On the basis of these conditions (Table 1, entry 21), a
variety of ketones were examined in this catalytic system.
However, we discovered that compared with aromatic ketones,
aliphatic ketones were much less effective under this catalyst
system, with a large amount of side products being generated.
One of the key side products was azine, which was formed by
the condensation of hydrazones and ketones or hydrolyzed
hydrazones. As Fe(III) has a high Lewis acidity due to a high
positive charge, we asserted that Fe(III) may also assist azine
formation while catalyzing the main reaction. Furthermore,
Fe(III) could oxidize the dmpe (sensitive to oxidant),
decreasing the catalysis efficiency. To avoid these problems,
we considered preparing an Fe(II) dmpe complex, which has a
lower Lewis acidity, to catalyze this reaction. Indeed, using
Fe(dmpe)2Cl2, a stable complex readily generated by the
literature procedure, instead of [Fe(dmpe)2Cl2]+[FeCl4]− as
catalyst, gave nearly quantitative yield of the desired product
under the same conditions (Table 1, entry 22).
a b
,
Table 1. Investigation of Reaction Conditions
a
General reaction conditions: 1a (0.25 mmol), 2a (0.2 mmol), [Fe]
catalyst (5 mol %), ligand (5 mol %), base (50 mol %), additive (50
mol %), solvent (0.2 mL) under N2 atmosphere. See Supporting
b
Information (SI) for details. 1H NMR yield was determined using
c
1,3,5-trimethoxybenzene as an internal standard. 20 mg of 4 Å
d
molecule sieve was added in the reaction system. 1a solution was
treated with 4 Å molecule sieves beforehand. See SI for details.
FeCl3, we then tested different ligands and found that the
desired product could only be formed when dmpe was used as
the ligand (Table 1, entry 4, entries 6−9), suggesting that this
reaction requires an electron-rich and less bulky ligand.
However, further explorations showed that the yield with the
FeCl3 catalyst system was extremely low, and dark particles
always formed at the end of the reaction. We attributed these
observations to the hydrolysis of FeCl3 at the high reaction
temperature (100 °C), which led to catalyst deactivation. To
avoid this problem, we premixed FeCl3 and dmpe (1:1) to
form a stable complex [Fe(dmpe)2Cl2]+[FeCl4]− to catalyze
the reaction. Unfortunately, only a slight improvement of the
yield was observed (entry 10). We then speculated that
lowering the reaction temperature may alleviate the hydrolysis
problem and thus tested the reaction at lower temperatures
(Table 1, entries 10−12). To our surprise, a slightly higher
yield (20%) was obtained at room temperature than at 100 °C
With the newly optimized conditions in hand, the substrate
scope of both electrophiles and hydrazones was tested again,
and much better results were obtained (Scheme 2). First,
different types of carbonyl compounds were studied. For
acetophenone derivatives, most have high reactivities. Other
alkyl aryl ketones also reacted very well, with a slight decrease
in yield for longer alkyl chains 3aa−3ca. When changing the
alkyl to trifluoromethyl, a very strong electron-withdrawing
group, the reaction also gave an extremely high yield (3da).
However, benzophenone has a much lower reactivity (3ea).
Acetophenones bearing either electron-donating or electron-
withdrawing groups gave good yields of the desired products
3fa−3ja. As previously mentioned, iron can also serve as an
efficient Lewis acid to catalyze the nucleophilic (both C- and
N-) addition to carbonyl compounds, and a main side reaction
for aliphatic ketone substrates was the nucleophilic attack by
B
DOI: 10.1021/acs.orglett.8b01391
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
as acrylonitrile also work in this reaction system, which may
greatly broaden the scope of conjugate addition of hydrazones.
Next, we tested different kinds of hydrazones as nucleophiles
in this reaction and found that the scope of hydrazone was
relatively limited. Only aromatic aldehyde hydrazones worked
well in this reaction. For aliphatic hydrazone, no desired
product was observed (3aj). The different behaviors of
aromatic and aliphatic hydrazones may be due to the fact
that for aromatic hydrazones, the aromatic ring can stabilize
the adjacent partial carbanion, while for the aliphatic one, such
a stabilization is absent. Thus, for the same reason, hydrazones
from aromatic aldehydes with electron-withdrawing groups on
the para position generally had high reactivity, and all gave
quantitative yields (3ab−3ac). In contrast, when electron-
donating groups were in the para position, the hydrazone did
not react well in THF. In such cases, the reaction can be
improved by using DMSO as solvent (3ad, 3ah). Substituents
at the meta position did not influence the reaction significantly
(3ae−3ag). Apart from these, other types of aromatic
hydrazones also gave high yields of the desired products (3ai).
Aromatic imines, comparatively weaker electrophiles, can
also undergo similar nucleophilic addition with hydrazone
(Scheme 3). Their reactivity, however, is much lower than
Scheme 2. Substrate Scope of Carbonyl Compounds and
Michael Acceptors
a b
,
a b
,
Scheme 3. Substrate Scope of Aromatic Imines
a
Reaction conditions: For 3ta−3wa, 1 (0.2 mmol), 2 (0.3 mmol),
Fe(dmpe)2Cl2 (5 mol %), base (50 mol %), CsF (50 mol %), and
solvent (0.2 mL), rt under N2. For the rest, 1 (0.25 mmol), 2 (0.2
mmol), Fe(dmpe)2Cl2 (5 mol %), base (50 mol %), CsF (50 mol %),
and solvent (0.2 mL), rt under N2. See SI for details. Condition A:
base: K3PO4, solvent: THF. Condition B: base: K3PO4, solvent:
b
t
DMSO. Condition C: base: BuOK, solvent: THF. Yield of isolated
c
1
product is reported unless noted. These products are volatile. H
NMR yield was determined using 1,3,5-trimethylbenzene as standard.
a
Reaction conditions: 1 (0.25 mmol), 2′ (0.2 mmol), Fe(dmpe)2Cl2
the nitrogen on hydrazone to form azine. For linear aliphatic
ketones, the shorter the aliphatic chain, the less side products
were formed (3ka−3ma). For cyclic ketones, both cyclo-
pentanone and cyclohexanone, as well as some derivatives,
could give good yields of the final products 3oa−3pa. Ketone
substrates containing methoxy or cyclopropyl also gave good
yields and sometimes better yields than simple alkyl ketones
(3na, 3qa). Aldehydes, being much more reactive than
ketones, also worked in this reaction. However, due to their
high reactivity, the competing azine formation was difficult to
prevent. Under the standard conditions, less than 50% yield of
the desired product was obtained. To improve the yield, we
considered two possible solutions: to use a stronger base, such
(5 mol %), K3PO4 (50 mol %), CsF (50 mol %), and solvent (0.2
mL), rt under N2. See SI for details. Yield of isolated product is
b
reported unless noted. *These products cannot be isolated efficiently
1
because of similar polarity to byproduct and the low yield. H NMR
yield was determined using 1,3,5-trimethylbenzene as an internal
standard based on the previous literature.8 **The side reaction was
observed by the standard 1H NMR peaks of side products 4* and 5*.
carbonyl compounds, and the substrate scope is more narrow
than those. Generally speaking, imines with electron-deficient
aromatic substituents favor the reaction (4aa−4ca) with
various aromatic aldehyde hydrazones (4aa−4ac, 4ba, 4bb).
With electron-rich aromatic rings as substituents, the imine
showed low or no reactivity (4da−4ea). On the other hand, a
strong electron-withdrawing group other than an aromatic ring
on imine nitrogen failed to generate the target product (4fa),
and the azine byproduct was formed almost quantitatively.
To study the chemoselectivity of this reaction, we designed
competition experiments shown in Scheme 4. Using a 1:1
mixture of acetophenone and acetone, the nucleophilic
addition of acetophenone was much more favored than that
of acetone, suggesting higher reactivity of acetophenone
derivatives than aliphatic ketones for such reactions (eq 1).
t
as BuOK, or to use DMSO as the solvent (since the azine
product was always harder to form in DMSO).15 After testing
t
these two ideas, we found that using BuOK as the base and
running the reaction in THF gave the desired product in
higher yields for aldehydes (3ra−3sa, 3xa).
In addition to carbonyl 1,2-additions, the iron−phosphine
complex also catalyzes the Michael addition of hydrazones to
activated alkenes (3ta−3wa). To our delight, in contrast to our
previously reported ruthenium-catalyzed conjugate addition,
Michael acceptors lacking an oxygen coordinating group such
C
DOI: 10.1021/acs.orglett.8b01391
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Scheme 4. Chemoselectivity Study
ACKNOWLEDGMENTS
■
The authors acknowledge the Canada Research Chair
Foundation (to C.-J. Li), the CFI, FRQNT Center for Green
Chemistry and Catalysis, NSERC, and McGill University for
support of our research. J.G. thanks China Scholarship Council
for a visiting scholarship. The authors thank Z. Hearne for
proofreading.
REFERENCES
(1) Shirley, D. A. Org. React. 1954, 8, 28.
(2) (a) Kishner, N. J. Russ. Phys. Chem. Soc. 1911, 43, 582. (b) Wolff,
L. Justus Liebigs Annal. Chem. 1912, 394, 86.
■
(3) (a) Baldwin, J. E.; Bottaro, J. C.; Kolhe, J. N.; Adlington, R. M. J.
Chem. Soc., Chem. Commun. 1984, 22. (b) Baldwin, J. E.; Adlington,
R. M.; Bottaro, J. C.; Kolhe, J. N.; Newington, I. M.; Perry, M. W.
Tetrahedron 1986, 42, 4235.
(4) Little, R. D.; Masjedizadeh, M. R.; Wallquist, O.; McLoughlin, J.
I. Org. React. 1995, 47, 315.
(5) Wang, H.; Dai, X.-J.; Li, C.-J. Nat. Chem. 2017, 9, 374.
(6) Chen, N.; Dai, X.-J.; Wang, H.; Li, C.-J. Angew. Chem., Int. Ed.
2017, 56, 6260.
(7) Dai, X.-J.; Wang, H.; Li, C.-J. Angew. Chem., Int. Ed. 2017, 56,
6302.
(8) (a) Yan, S.-S.; Zhu, L.; Ye, J.-H.; Zhang, Z.; Huang, H.; Zeng,
H.; Li, C.-J.; Lan, Y.; Yu, D.-G. Chem. Sci. 2018, 9, 4873. For
olefination reaction, see: (b) Wei, W.; Dai, X.-J.; Wang, H.; Li, C.;
Yang, X.; Li, C.-J. Chem. Sci. 2017, 8, 8193. For cross-coupling, see:
(c) Tang, J.; Lv, L.; Dai, X.-J.; Li, C.-C.; Li, L.; Li, C.-J. Chem.
Commun. 2018, 54, 1750. (d) Lv, L.; Zhu, D.; Tang, J.; Qiu, Z.; Li, C.-
C.; Gao, J.; Li, C.-J. ACS Catal. 2018, 8, 4622.
Replacing acetone with a long-chain aliphatic ketone increased
the product ratio in favor of the acetophenone addition
product (eq 2). The reactivity of benzophenone is also much
lower than acetophenone probably due to its high steric effect
(eq 3). Based on these analysis, we concluded that
acetophenone derivatives are more reactive for this trans-
formation than other kinds of ketone substrates.
In conclusion, we have successfully established the use of
earth-abundant and well-defined iron complexes to catalyze the
nucleophilic additions of hydrazones to a broad scope of
carbonyl compounds, imines, and Michael acceptors at room
temperature. The study not only marks the first abundant-
metal-catalyzed nucleophilic reaction of an organic carbanion
equivalent via hydrazones but also opens up a potentially new
avenue in homogeneous iron catalysis. In addition, unlike
classical Grignard-type reactions, chemoselectivity of various
kinds of carbonyl groups can be realized by this reaction. The
mechanism,16 asymmetric studies, and synthetic applications of
this reaction and this catalytic system are under further
investigation.
́
̈
rstner, A.; Leitner, A.; Mendez, M.; Krause, H. J. Am.
(9) (a) Fu
Chem. Soc. 2002, 124, 13856. (b) Nakamura, M.; Matsuo, K.; Ito, S.;
Nakamura, E. J. Am. Chem. Soc. 2004, 126, 3686. (c) Furstner, A.;
̈
Martin, R. Chem. Lett. 2005, 34, 624. (d) Sapountzis, I.; Lin, W.;
Kofink, C. C.; Despotopoulou, C.; Knochel, P. Angew. Chem., Int. Ed.
́
2005, 44, 1654. (e) Guerinot, A.; Reymond, S.; Cossy, J. Angew.
Chem., Int. Ed. 2007, 46, 6521. (f) Hatakeyama, T.; Nakamura, M. J.
Am. Chem. Soc. 2007, 129, 9844. (g) Fu
H.; Seidel, G.; Goddard, R.; Lehmann, C. W. J. Am. Chem. Soc. 2008,
130, 8773. (h) Sherry, B. D.; Furstner, A. Acc. Chem. Res. 2008, 41,
̈
rstner, A.; Martin, R.; Krause,
̈
1500. (i) Sun, C.-L.; Li, B.-J.; Shi, Z.-J. Chem. Rev. 2011, 111, 1293.
(j) Zhao, D.; Shi, Z.; Glorius, F. Angew. Chem., Int. Ed. 2013, 52,
12426. (k) Shang, R.; Ilies, L.; Asako, S.; Nakamura, E. J. Am. Chem.
Soc. 2014, 136, 14349. (l) Zhong, Y.; Han, W. Chem. Commun. 2014,
ASSOCIATED CONTENT
* Supporting Information
■
̈
50, 3874. (m) Bauer, I.; Knolker, H.-J. Chem. Rev. 2015, 115, 3170.
S
(n) Ludwig, J. R.; Zimmerman, P. M.; Gianino, J. B.; Schindler, C. S.
́
Nature 2016, 533, 374. (o) Saa, C. Angew. Chem., Int. Ed. 2016, 55,
10960.
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.8b01391.
(10) Chatt, J.; Hayter, R. G. J. Chem. Soc. 1961, 5507.
(11) (a) Pasternack, R. F.; Huber, P. R.; Boyd, P.; Engasser, G.;
Francesconi, L.; Gibbs, E.; Fasella, P.; Cerio Venturo, G.; Hinds, L. d.
J. Am. Chem. Soc. 1972, 94, 4511. (b) Whittlesey, M. K.; Mawby, R. J.;
Osman, R.; Perutz, R. N.; Field, L. D.; Wilkinson, M. P.; George, M.
W. J. Am. Chem. Soc. 1993, 115, 8627. (c) Miller, W. K.; Gilbertson, J.
D.; Leiva-Paredes, C.; Bernatis, P. R.; Weakley, T. J. R.; Lyon, D. K.;
Tyler, D. R. Inorg. Chem. 2002, 41, 5453. (d) Hazari, N. Chem. Soc.
Rev. 2010, 39, 4044. (e) Casitas, A.; Krause, H.; Goddard, R.;
General experimental information, experimental proce-
dures, additional control experiments to understand the
mechanism, a plausible mechanism for the iron-catalyzed
nucleophilic reaction of hydrazones, reaction of 1a with
2a at 1 mmol scale, spectroscopic data of products,
reference, and NMR spectra of products (PDF)
Furstner, A. Angew. Chem., Int. Ed. 2015, 54, 1521. (f) Field, L. D.;
̈
AUTHOR INFORMATION
■
Guest, R. W.; Vuong, K. Q.; Dalgarno, S. J.; Jensen, P. Inorg. Chem.
2009, 48, 2246. (g) Field, L. D.; Li, H. L.; Dalgarno, S. J.; McIntosh,
R. D. Inorg. Chem. 2012, 51, 3733.
Corresponding Author
*E-mail: cj.li@mcgill.ca.
ORCID
(12) (a) Koerner von Gustorf, E. A.; Grevels, F.-W.; Fischler, I. The
Organic Chemistry of Iron; Academic Press: New York, 1978.
(b) Silver, J. Chemistry of Iron; Blackie Academic & Professional:
London, NY, 1993. (c) Symons, M. C. R.; Gutteridge, J. M. C. Free
Radicals and Iron: Chemistry, Biology, and Medicine; Oxford University
Press: Oxford, NY, 1998. (d) Miessler, G. L.; Tarr, D. A. Inorganic
Chemistry; Pearson Education: Upper Saddle River, N.J., 2004.
Chen-Chen Li: 0000-0002-3251-5450
Chao-Jun Li: 0000-0002-3859-8824
Notes
The authors declare no competing financial interest.
D
DOI: 10.1021/acs.orglett.8b01391
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(e) Plietker, B. Iron Catalysis in Organic Chemistry; Wiley-VCH Verlag
GmbH & Co. KGaA: Weinheim, 2008.
(13) (a) Mashuta, M. S.; Webb, R. J.; McCusker, J. K.; Schmitt, E.
A.; Oberhausen, K. J.; Richardson, J. F.; Buchanan, R. M.;
Hendrickson, D. N. J. Am. Chem. Soc. 1992, 114, 3815. (b) Feng,
W.; Nansheng, D. Chemosphere 2000, 41, 1137. (c) Williams, A. G. B.;
Scherer, M. M. Environ. Sci. Technol. 2004, 38, 4782.
(14) Clark, J. H. Chem. Rev. 1980, 80, 429.
(15) When we pregenerate hydrazone in THF and some other
solvents, the azine byproduct, more or less, can be observed.
However, in DMSO, we cannot observe any azine product.
(16) A plausible mechanism together with some related control
experiments were included in the SI for reference.
E
DOI: 10.1021/acs.orglett.8b01391
Org. Lett. XXXX, XXX, XXX−XXX