Synthesis and complexes of fluoroalkoxy carbenes

Article abstract of DOI:10.1002/anie.201301503

Gripping ligands: A series of fluoroalkoxy-functionalized imidazol-2-ylidene ligands were synthesized by the alkylation of a range of commercially available azoles with hexafluoroisobutylene oxide. The deprotonation of these species with a strong base cleanly generated the free carbenes, which acted as tridentate chelating ligands (see scheme). Copyright

Full text of DOI:10.1002/anie.201301503

Angewandte
Chemie
DOI: 10.1002/anie.201301503
Fluorinated Ligands
Synthesis and Complexes of Fluoroalkoxy Carbenes
˘
Anthony J. Arduengo III,* Joshua S. Dolphin, Gabriela Gurau, William J. Marshall,
Joseph C. Nelson, Viacheslav A. Petrov, and Jason W. Runyon*
Building on the success of his hexafluorocumylalcohol-
derived bidentate ligand 1 (Scheme 1), Martin and co-work-
ers designed two versions of a tridentate ligand, 2 and 3, which
were used to explore the chemistry of main-group elements in
unusual coordination environments.[1–4] The relatively elec-
(5).[20–26] In an effort to avoid ring-opening polymerization of
the epoxide, initial attempts to prepare the desired tridentate
ligands involved the use of the iodohydrin 6 (derived from
epoxide 5) as an alkylating agent.[27] When a solution of
sodium imidazolide in THF was heated at reflux with two
equivalents of the iodohydrin over a period of days, the
imidazolium salt 7 was formed (Scheme 2).
Scheme 1. Bi- and tridentate Martin-type complexes 1–3 and a similar
architecture, 4, based on an imidazol-2-ylidene. E=main-group ele-
ment, M=metal.
Scheme 2. Synthesis of bis(fluoroalcohol)-substituted imidazolium
iodides.
tropositive equatorial substituent (C or N, respectively), the
À
À
electronegative axial substituents (R C(CF3)2 O), the for-
mation of favorable five- and six-membered rings, and the
Thorpe–Ingold effect all contribute to the efficacy of the
Martin ligand.[5] The use of Martin-type ligands for d-block
metals is limited, partially owing to the relatively poor aryl–
metal bond.
The reaction can be carried out as a one-pot process or as
two successive alkylation steps to form first the mono-
substituted imidazole and eventually the desired imidazolium
salt. The chemical shifts in 7 resemble those of other
À
imidazolium
salts
(dH(C2 H) ꢀ 9 ppm;
dC(C2)
ꢀ 140 ppm).[28] The crystal packing of 7 (Figure 1) displays
a hydrogen-bonded network of cations and anions. Hydrogen
bonds are formed between the alcohol hydrogen atoms and
the iodide, and a third hydrogen bond is formed between the
The abundance of adducts of hypervalent main-group
elements with carbenes and of transition-metal–carbene
complexes suggests that replacement of the central aryl
group in the Martin-type structures 2 and 3 with an imidazol-
2-ylidene (as in 4) as the central binding moiety should
provide a more suitable ligand for d-block metal complexes
while taking advantage of all of the other attractive aspects of
the Martin ligands.[6–9] A number of popular tridentate ligands
already exist, but the architecture of 4 offers a remarkably
robust structure.[10–19]
Introduction of an CH2C(CF3)2OH group on an azole is
simplified by the availability of hexafluoroisobutylene oxide
˘
[*] Prof. Dr. A. J. Arduengo III, J. S. Dolphin, Dr. G. Gurau, J. C. Nelson,
Dr. J. W. Runyon
The University of Alabama, Department of Chemistry (USA)
E-mail: aj@ajarduengo.net
jwrunyon@crimson.ua.edu
Homepage: http://ajarduengo.net
W. J. Marshall, Dr. V. A. Petrov
DuPont Central Research and Development
Wilmington, DE 19880-0500 (USA)
Figure 1. Diagram of the X-ray crystal structure of the imidazolium salt
7. Thermal ellipsoids are drawn at the 50% probability level; hydrogen-
bond distances [ꢀ]: H1’–I1’ 3.07, H1/H2–I1 2.75.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201301503.
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
hydrogen atom at C2 (H1’) of another cation and the iodide.
Electrostatic interactions between the imidazolium ions and
iodide counterions are also evident (r= 3.47 ꢀ).
transition metals render this approach undesirable. It was
reported that conditions of phase-transfer catalysis (PTC)
were successful for the preparation of a pyrazole bearing
a CH2C(CF3)2OH nitrogen substituent.[25b,26]. Therefore, the
analogous azoles 9, 13, and 14 were prepared under PTC
conditions (Scheme 4). These adducts were obtained by the
treatment of imidazole or triazole with only one equivalent of
The electron-withdrawing nature of the geminal CF3
groups renders the alcohol hydrogen atoms more acidic
than those on the azolium ring. These acidities are expected to
be very different from those found in the tridentate bis-
(alkoxy) carbene ligands introduced by Arnold and co-
workers.[29–34] The acidity and stability of the fluoroalkoxy
carbene ligands reported herein offer distinct advantages over
their nonfluorinated counterparts. Treatment of 7 with one
equivalent of a base results in the deprotonation of an alcohol
group to form the stable zwitterion 8 (Scheme 3). These
fluoroalkoxy zwitterions are unique, because more basic
alkoxides have been shown to attack the imidazolium carbon
atom (C2).[33] The solubility of compound 8 in nonpolar
organic solvents facilitates separation from the imidazolium
salt.
Scheme 4. Alkylation of azoles, including substituted azoles, with
epoxide 5.
epoxide 5. Some over-alkylation occurred during the prepa-
ration of 9 and 13. Thus, a mixture of mono- and bis-
substituted azoles (zwitterion 8 or 15) was obtained, whereby
the monoalkylated species was isolated as the major product.
Selective alkylation of 1,2,4-triazole proved difficult owing to
the multiple nucleophilic centers on the triazole ring. The
reaction of 1,2,4-triazole also competes with the oligomeriza-
tion of epoxide 5, which leads to a complex mixture. A minor
(uncharacterized) symmetrical adduct was observed in the
NMR spectra which may be the 4-substituted regioisomer.
This compound may be the 4-substituted regioisomer. Adduct
13 could be isolated and was fully characterized (Figure 3);
Scheme 3. Synthesis of imidazolium fluoroalkoxides.
The growth of crystals of 8 under different solvent
conditions resulted in different packing motifs. Both a chain
of hydrogen-bonded molecules (Figure 2a) and hydrogen-
bonded “dimers” could be obtained (Figure 2b). The hydro-
gen-bond distances lie in the range of 1.4–2.2 ꢀ. This hydro-
gen bonding is quite dynamic, as illustrated by the two low-
energy crystal motifs and rapid proton tautomerism in fluid
1
solution (the H NMR signals for the alcohol and imidazole
À
C2 H hydrogen atoms are broad, whereas the methylene
groups appear as a sharp singlet).
The iodohydrin-based synthetic approach to the triden-
tate ligand is practical. However, complications associated
with an iodide counterion in subsequent reactions involving
Figure 3. X-ray crystal structures of 9, 13, and 14 with thermal
ellipsoids drawn at the 50% probability level.
however, the product was obtained in low yield (10%).
Although only one isomer was isolated, the existing data are
not sufficient to make a conclusion regarding the regioselec-
tivity of this reaction. In stark contrast to the reactivity of
1,2,4-triazole with 5, the reaction of 1,2,3-triazole with 5 led to
only one regioisomer, 14. No evidence for bisalkylation was
found, even when an excess of the epoxide 5 was used.
Figure 2. Packing motifs in 8: a) linear hydrogen-bonding network;
b) hydrogen-bonding “dimers”. Thermal ellipsoids are drawn at the
50% probability level.
2
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
X-ray single-crystal diffraction studies on 9, 13, and 14
(Figure 3) confirmed the substitution patterns of the azole
rings and indicated that the preferred tautomeric structure is
neutral (the proton lies on the oxygen atom), rather than
a zwitterionic imidazolium alkoxide.
Since some bis-alkylation occurred in the preparation of
monosubstituted azoles even under PTC conditions, it was
envisioned that zwitterions 8, 10–12, and 15 could be obtained
in a one-pot reaction without the isolation of an intermediate.
Indeed, the addition of exactly two equivalents of epoxide 5
to the initial azoles led to the formation of zwitterions 8, 10–
12, and 15 in good yields with none of the by-products found
under PTC conditions. Alkylation occurred exclusively at the
nitrogen atoms: no ethoxy ethers derived from oxygen
alkylation were observed. In the case of 1,2,4-triazole,
alkylation with two equivalents of epoxide 5 afforded only
the 1,4-disubstituted isomer. An X-ray crystal structure of 15
(Figure 4) confirmed this substitution pattern and revealed
the same hydrogen-bonding tendencies as observed for 8.
Scheme 5. Preparation of bis(fluoroalkoxy) carbenes.
Table 1: 13C NMR chemical shifts of the C2 carbon atom of carbenes 16–
20 in [D8]THF.
Compound
dC [ppm]
16
17
18
19
20
218.34
220.25
219.27
229.75
212.76
solution for short periods, they decompose when left in
solution over days.
Addition of the anionic fluoroalkoxy carbene 16 to
[NiCl2(PPh3)2] afforded the tridentate complex 21 and the
bis(carbene) complex 22 (Scheme 6). The latter complex was
isolated after workup on neutral alumina. In spite of the 1:1
stoichiometry and dropwise addition of the carbene to
Figure 4. X-ray crystal structure of 15 with thermal ellipsoids drawn at
the 50% probability level.
Two different methods were developed for the epoxide-
ring-opening alkylation (the reactions were carried out in
acetonitrile or toluene). Although some small specific advan-
tages for particular substrates differentiate these two epoxide-
based methods, both approaches provide the products in
higher overall yields, involve fewer steps, offer broader scope
for the synthesis of ligands, and enable simpler purification of
the products than the iodohydrin-based method described
above.
The treatment of zwitterions 8, 10–12, and 15 with two
equivalents of a strong base generated the free bis(fluoroal-
koxy) carbenes 16–20 (Scheme 5). The existence of transient
carbenes was unambiguously confirmed in situ by 1H, 13C, and
19F NMR spectroscopy. Deprotonation with KOtBu is rapid
and complete within minutes. Spectroscopic evidence (multi-
ple broad 19F signals and broad methylene resonances in the
1H NMR spectra) suggests that the alkoxy groups may form
ionic aggregates with alkali-metal cations in solution.
The chemical shifts of the carbene carbon atom in 16–20
(Table 1) are in good agreement with those of other nonionic
carbenes; this similarity suggests little or no carbene–potas-
sium coordination.[35–37] Thus far it has not been possible to
isolate single crystals of any of these anionic carbene ligands.
Although the free carbenes are conveniently stable in
Scheme 6. Synthesis of a tridentate nickel carbene complex 21 and
a bis(carbene) nickel complex 22.
[NiCl2(PPh3)2], both complexes were isolated from the
reaction mixture in a ratio of 2:1 (21/22). The formation of
this mixture of products results from a combination of factors,
including solvent effects, the greater donor strength of CD
relative to that of PD, a chelation effect, and differences in the
ligand-dissociation kinetics of the nickel–phosphine adducts
in solution. The different physical properties of 21 and 22
enable facile separation of these compounds.
The resonances of both the carbene carbon atom and the
phosphorus atom in 21 (dC = 159.5 ppm, dP = 6.03 ppm)
appear upfield in the corresponding NMR spectra relative
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
to those of nonchelated carbene analogues ([IMesNiCl2PPh3]:
dC = 165.3 ppm, dP = 13.74 ppm; IMes = 1,3-bis(2,4,6-trime-
thylphenyl)imidazol-2-ylidene).[38] The methylene and tri-
fluoromethyl groups of the ligand appear as singlets in the
1H and 19F NMR spectra, respectively. This latter observation
suggests flexibility of the somewhat twisted geometry for
a planar tridentate arrangement (see below).
An X-ray crystal structure of 21 (Figure 5) revealed the
expected square-planar geometry around the nickel center.
The formation of a 6,5,6 tricyclic fused ring system about
that two alkoxy groups from one carbene ligand are bound to
nickel, whereas two hydroxy groups from the other carbene
ligand are hydrogen-bonded to the adjacent oxygen atoms.
The two ligands adopt an antiplanar relationship to one
another with a dihedral angle ranging from 83 to 938 as
a result of the rotation of the dihydroxy carbene ligand to
form hydrogen bonds that complete two distorted eight-
membered rings through H-bonding. The bonds of the
chelated ligand in 22 are elongated by 0.05 ꢀ relative to
those in complex 21. The bond lengths and angles around the
nickel center do not significantly deviate from those of other
trans-bis(carbene) nickel complexes.[38–41]
In conclusion, we have described a versatile new series of
fluoroalkoxy tridentate carbene ligands. Studies of novel
organometallic complexes of these ligands with a variety of
metals, including Fe, Ta, Ti, Ru, and W, in various catalytic
reactions will be reported subsequently.
Received: February 21, 2013
Published online: && &&, &&&&
Keywords: carbenes · fluorinated ligands · imidazol-2-ylidenes ·
.
nitrogen heterocycles · tridentate ligands
Figure 5. X-ray crystal structure of 21 with thermal ellipsoids drawn at
the 50% probability level. Selected bond lengths [ꢀ] and angles [8]:
C2–Ni1 1.853, P1–Ni1 2.289, O1/O2–Ni1 1.83; N1-C2-N3 105.17, C2-
Ni1-P1 172.51, O1-Ni1-O2 171.65.
[1] T. T. Nguyen, R. L. Amey, J. C. Martin, J. Org. Chem. 1982, 47,
1024 – 1027.
[2] C. W. Perkins, J. C. Martin, J. Am. Chem. Soc. 1986, 108, 3211 –
3214.
a square-planar metal center slightly distorts the structure by
twisting the imidazol-2-ylidene ring out of the O-Ni-O plane
(ftwist = 13–208). This variability in solid-state torsions sup-
ports the observations from NMR spectroscopy in solution of
[3] C. W. Perkins, R. B. Clarkson, J. C. Martin, J. Am. Chem. Soc.
1986, 108, 3206 – 3210.
[4] E. F. Perozzi, R. S. Michalak, J. D. Figuly, W. H. Stevenson III,
D. B. Dess, J. C. Martin, J. Am. Chem. Soc. 1981, 103, 1049.
[5] J. C. Martin, Science 1983, 221, 509 – 514.
[6] A. J. Arduengo III, H. V. R. Dias, R. L. Harlow, M. Kline, J. Am.
Chem. Soc. 1992, 114, 5530 – 5534.
[7] A. J. Arduengo III, J. R. Goerlich, R. Krafczyk, W. J. Marshall,
Angew. Chem. Int. Ed. 1998, 37, 1963 – 1965.
À
a flexible and a conformationally dynamic process. The C Ni
À
and P Ni bond lengths are typical of those in other nickel
complexes despite this twisting; this similarity suggests that
tridentate chelation does not substantially distort the car-
bene–nickel interaction.[38]
[8] B. Bostai, Z. Novak, A. Benyei, A. Kotachy, Org. Lett. 2007, 9,
3437 – 3439.
The chelated and hydrogen-bonded carbene ligands in 22
1
give separate signals in the H, 13C, and 19F NMR spectra,
[9] D. Enders, K. Breuer, G. Raabe, J. Runsink, J. H. Teles, J.-P.
Melder, K. Ebel, S. Brode, Angew. Chem. 1995, 107, 1119 – 1122;
Angew. Chem. Int. Ed. Engl. 1995, 34, 1021 – 1023.
[10] L. P. Spencer, C. Beddie, M. B. Hall, M. D. Fryzuk, J. Am. Chem.
Soc. 2006, 128, 12531 – 12543.
which suggests that their interconversion is slow on the NMR
timescale. An X-ray crystal structure of 22 (Figure 6) showed
[11] L. P. Spencer, B. A. MacKay, B. O. Patrick, M. D. Fryzuk, Proc.
Natl. Acad. Sci. USA 2006, 103, 17094 – 17098.
[12] B. A. MacKay, R. F. Munha, M. D. Fryzuk, J. Am. Chem. Soc.
2006, 128, 9472 – 9483.
[13] L. P. Spencer, S. Winston, M. D. Fryzuk, Organometallics 2004,
23, 3372 – 3374.
[14] S. S. Rocks, W. W. Brennessel, T. E. Machonkin, P. L. Holland,
Inorg. Chim. Acta 2009, 362, 1387 – 1390.
[15] C. C. Hojilla Atienza, C. Milsmann, E. Lobkovsky, P. J. Chirik,
Angew. Chem. 2011, 123, 8293 – 8297; Angew. Chem. Int. Ed.
2011, 50, 8143 – 8147.
[16] S. K. Russell, C. Milsmann, E. Lobkovsky, T. Weyhermꢁller, P. J.
Chirik, Inorg. Chem. 2011, 50, 3159 – 3169.
[17] A. M. Tondreau, C. Milsmann, A. D. Patrick, H. M. Hoyt, E.
Lobkovsky, K. W. Wieghardt, P. J. Chirik, J. Am. Chem. Soc.
2010, 132, 15046 – 15059.
Figure 6. X-ray crystal structure of 22 with thermal ellipsoids drawn at
the 50% probability level: a) side-on view; b) view down the axis of the
chelated carbene. Non-hydroxy hydrogen atoms are omitted for clarity.
[18] P. J. Chirik, K. W. Wieghardt, Science 2010, 327, 794 – 795.
4
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
[19] A. C. Bowman, C. Milsmann, C. C. H. Atienza, E. Lobkovsky,
K. W. Wieghardt, P. J. Chirik, J. Am. Chem. Soc. 2010, 132, 1676 –
1684.
[32] I. S. Edworthy, M. Rodden, S. A. Mungur, K. M. Davis, A. J.
Blake, C. Wilson, M. Schrçder, P. L. Arnold, J. Organomet.
Chem. 2005, 690, 5710 – 5719.
[20] V. A. Petrov, W. J. Marshall, C. G. Krespan, V. F. Cherstkov,
E. A. Avetisian, J. Fluorine Chem. 2004, 125, 99 – 105.
[21] V. A. Petrov, J. Fluorine Chem. 2004, 125, 531 – 536.
[22] V. A. Petrov, U.S. Pat. Appl. Publ., US 20080177097A1
20080724, 2008.
[33] P. L. Arnold, I. J. Casely, Z. R. Turner, C. D. Carmichael, Chem.
Eur. J. 2008, 14, 10415 – 10422.
[34] P. L. Arnold, M. S. Sanford, S. M. Pearson, J. Am. Chem. Soc.
2009, 131, 13912 – 13913.
[35] A. J. Arduengo III, F. Davidson, H. V. R. Dias, J. R. Goerlich, D.
Khasnis, W. J. Marshall, T. K. Prakasha, J. Am. Chem. Soc. 1997,
119, 12742 – 12749.
[36] A. J. Arduengo III, R. Krafczyk, R. Schmultzer, H. A. Craig,
J. R. Goerlich, W. J. Marshall, M. Unverzagt, Tetrahedron 1999,
55, 14523 – 14534.
[37] F. E. Hahn, L. Wittenbecher, R. Boese, D. Blꢂser, Chem. Eur. J.
1999, 5, 1931 – 1935.
[38] K. Matsubara, U. Keit, S. Youhei, Organometallics 2006, 25,
3422 – 3427.
[39] A. J. Arduengo III, S. F. Gamper, J. C. Calabrese, F. Davidson, J.
Am. Chem. Soc. 1994, 116, 4391 – 4394.
[40] A. J. Arduengo III, T. K. Prakasha, W. J. Marshall, 1996, unpub-
[23] V. A. Petrov, Synthesis 2002, 2225 – 2231.
[24] V. A. Petrov, W. J. Marshall, J. Fluorine Chem. 2011, 132, 41 – 51.
[25] a) V. A. Petrov, A. E. Feiring, J. Feldman, U.S. Pat. Appl. Publ.,
WO 00/66575 PTC/US00/11746, Patent no.: US 6,653,419, 2003;
b) A. S. Ionkin, W. J. Marshall, Y. Wang, V. A. Petrov, U.S. Pat.
Appl. Publ., WO 07/133523, Patent no.: US 7,737,277, 2010.
[26] A. S. Ionkin, Y. Wang, W. J. Marshall, V. A. Petrov, J. Organo-
met. Chem. 2007, 692, 4809 – 4827.
[27] W. A. Herrmann, L. J. Gooßen, M. Spiegler, J. Organomet.
Chem. 1997, 547, 357 – 366.
[28] D. T¸apu, D. A. Dixon, C. Roe, Chem. Rev. 2009, 109, 3385 – 3407.
[29] P. L. Arnold, M. Rodden, C. Wilson, Chem. Commun. 2005,
1743 – 1745.
lished
results;
http://www.ajarduengo.net/Jmol/htmlfiles/
[30] P. L. Arnold, C. Wilson, Inorg. Chim. Acta 2007, 360, 190 – 196.
[31] P. L. Arnold, A. J. Blake, C. Wilson, Chem. Eur. J. 2005, 11,
6095 – 6099.
0253Viewer.html.
[41] C. H. Lee, T. R. Cook, D. G. Nocera, Inorg. Chem. 2011, 50, 714 –
716.
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!

.
Angewandte
Communications
Communications
Fluorinated Ligands
˘
A. J. Arduengo,* J. S. Dolphin, G. Gurau,
W. J. Marshall, J. C. Nelson, V. A. Petrov,
J. W. Runyon*
&&&& — &&&&
Synthesis and Complexes of Fluoroalkoxy
Carbenes
Gripping ligands: A series of fluoroalkoxy-
functionalized imidazol-2-ylidene ligands
were synthesized by the alkylation of
a range of commercially available azoles
with hexafluoroisobutylene oxide. The
deprotonation of these species with
a strong base cleanly generated the free
carbenes, which acted as tridentate che-
lating ligands (see scheme).
6
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
These are not the final page numbers!