Synthesis of 2-nickela(II)oxetanes from nickel(0) and epoxides: Structure, reactivity, and a new mechanism of formation

Article abstract of DOI:10.1021/jacs.5b06735

2-Nickelaoxetanes have been frequently invoked as reactive intermediates in catalytic reactions of epoxides using nickel, but have never been isolated or experimentally observed in these transformations. Herein, we report the preparation of a series of well-defined nickelaoxetanes formed via the oxidative addition of nickel(0) with epoxides featuring ketones. The stereochemistry of the products is retained, which has not yet been reported for nickelaoxetanes. Theoretical calculations support a bimetallic ring-opening/closing pathway over a concerted oxidative addition. Initial reactivity studies of a nickelaoxetane demonstrated protonolysis, oxidatively induced reductive elimination, deoxygenation, and elimination reactions when treated with the appropriate reagents.

Full text of DOI:10.1021/jacs.5b06735

Communication
pubs.acs.org/JACS
Synthesis of 2‑Nickela(II)oxetanes from Nickel(0) and Epoxides:
Structure, Reactivity, and a New Mechanism of Formation
Addison N. Desnoyer, Eric G. Bowes, Brian O. Patrick, and Jennifer A. Love*
Department of Chemistry, The University of British Columbia, Vancouver, British Columbia V6T 1Z1, Canada
S
* Supporting Information
ketene. The authors proposed that these nickelaoxetanes arise
from pathways that proceed via [2 + 2] cyclization.
ABSTRACT: 2-Nickelaoxetanes have been frequently
invoked as reactive intermediates in catalytic reactions of
epoxides using nickel, but have never been isolated or
experimentally observed in these transformations. Herein,
we report the preparation of a series of well-defined
nickelaoxetanes formed via the oxidative addition of
nickel(0) with epoxides featuring ketones. The stereo-
chemistry of the products is retained, which has not yet
been reported for nickelaoxetanes. Theoretical calculations
support a bimetallic ring-opening/closing pathway over a
concerted oxidative addition. Initial reactivity studies of a
nickelaoxetane demonstrated protonolysis, oxidatively
induced reductive elimination, deoxygenation, and elimi-
nation reactions when treated with the appropriate
reagents.
Recently, efforts in our group have been focused on the
development of new chemical transformations based on the
reactivity of well-defined metallaoxetanes of rhodium.6 We have
reported that a 2-rhodaoxetane can undergo transmetalation
with aryl- and alkenylboronic acids6b as well as insertion with a
variety of unsaturated electrophiles to give ring-expanded
metallacycles.6c Ring expansion occurred exclusively into the
M−O bond of the rhodaoxetane, which contrasts with reported
insertion into the M−C bond of nickelaoxetanes. This
complementary reactivity of nickel versus rhodium inspired us
to investigate the synthesis of a well-defined nickelaoxetane and
explore its reactivity. We report our initial findings herein.
Ogoshi has demonstrated that carbonyl groups can be used as
directing groups for a wide array of nickel-mediated trans-
formations.7 Inspired by these reports, we prepared epoxide rac-2
and reacted it with (dtbpe)Ni(COD) (1, dtbpe =1,2-bis(di-tert-
butyl)phosphinoethane, COD = 1,5-cyclooctadiene). Monitor-
ing the reaction by 31P NMR spectroscopy reveals the formation
of a mixture of products over the course of 6 h. The major
product appears as two doublets (δ = 72.7 and 71.5 ppm) with
he 2-metallaoxetanes, most notably invoked as reactive
T
intermediates during catalytic olefin oxidation reactions,
remain relatively rare chemical species.1 The classic example of a
nickelaoxetane in the literature is a report by de Pasquale on the
formation of cyclic carbonates from epoxides and CO2.2 The
nickelaoxetane is proposed to arise from the oxidative addition of
nickel(0) to the epoxide. More recently, the Jamison group has
reported on the reductive coupling of alkynes and epoxides
catalyzed by nickel(0) complexes.3 The proposed mechanism of
this transformation involves SN2-type attack of the nickel(0)
complex onto the epoxide, with inversion of configuration,
followed by ring closure to form a nickelaoxetane (Scheme 1);
2
small JP,P values of 6 Hz, indicative of coupling through a
nickel(II) center.8 Removing the volatiles in vacuo and extracting
the residue with pentanes allowed for the isolation of the major
product 3 as the racemate in 20% yield as an analytically pure
orange powder (Scheme 2). A variety of NMR techniques and X-
ray crystallography were employed to determine the structure of
3, which is the first example of an isolable nickela(II)oxetane
derived from nickel(0) and epoxides.
The 1H NMR spectrum of 3 displays a downfield multiplet for
H3 (δ = 5.78 ppm; see Scheme 2 for numbering). A COSY
Scheme 1. Syntheses of Nickelaoxetanes from Nickel(0)
a
Scheme 2. Initial Synthesis of 3
subsequent insertion of the alkyne occurs exclusively into the
M−C bond of the oxetane. The Doyle group has also suggested
the formation of nickelaoxetanes as intermediates during the
transmetalation of aryl epoxides with boronic acids using a
nickel(0) catalyst.4 Although all of these reports invoke the
formation of a nickelaoxetane from nickel(0) and an epoxide, the
nickelacycles have not been isolated or observed experimentally.
Indeed, to date the only examples of well-defined nickel-
aoxetanes belong to the Hillhouse group, where the oxetanes are
formed by the reaction of a nickel alkylidene5a or imido5b with a
a
Isolated yield in parentheses. Relative stereochemistry shown for
clarity.
Received: June 29, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b06735
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
experiment shows that this resonance is coupled to another at
2.87 ppm, which is assigned as the resonance for H2.
Importantly, a NOESY experiment shows strong correlations
between the H2 and H3 protons, indicative of cis stereo-
chemistry. Each of the methylene resonances of the cyclohexyl
ring are diastereotopic, which complicates the aliphatic region of
the 1H NMR spectrum. However, they can each be assigned on
the basis of COSY, HSQC, and HMBC experiments. The
presence of the free ketone is shown by the downfield shift of the
C1 resonance in the 13C NMR spectrum (δ = 211.7, apparent t, J
= 2 Hz). The slow evaporation of a cold Et2O solution of 3
allowed for the growth of dark red crystals, and the structure of 3
was confirmed by an X-ray diffraction study (Figure 1). The
As 5 proved too reactive to isolate, we sought to prepare an
analogue that did not contain the epoxide moiety. Addition of
cyclohexanone to a solution of 4 resulted in an immediate color
change from red to yellow. Crystallization of the crude product
from pentanes afforded the η2-ketone adduct, 6 (see Figure 1 for
solid-state structure). As expected, 5 and 6 are spectroscopically
similar, in particular with regard to the coupling constants in the
13C and 31P NMR spectra.11 In addition, we found that equimolar
amounts of either 1 or 4 and cyclohexene oxide in C6D6 resulted
in no reaction over the course of several days, which highlights
the importance of the ketone in this transformation (vide infra).
Furthermore, the yield of 3 was found to be unaffected by the
presence of 1,4-cyclohexadiene, which indicates to us that a
radical-based mechanism is unlikely.11
We also prepared complexes 9 and 10 using an analogous
synthetic route to that of 3 (Scheme 4).11 To explore the
a
Scheme 4. Synthesis of 6, 9, and 10
Figure 1. POV-Ray (50% ellipsoids) diagrams of 3 and 6. Hydrogen
atoms omitted for clarity.
a
Isolated yields in parentheses. Relative stereochemistry shown for
clarity.
geometry at the nickel is square planar, with a Ni−O distance of
1.832(1) Å and a Ni−C distance of 2.014(2) Å. The metallacycle
bond lengths are similar to those of Hillhouse’s oxetanes.5
With 3 in hand, we sought to explore the mechanism of
formation and reactivity of this species. We speculated that COD
could be competing with 2 at the nickel(0) center, resulting in
low yields of 3.9 Thus, we explored other sources of
(dtbpe)nickel(0) as starting materials. While (dtbpe)Ni(C2H4)
was found to be unreactive with 2, the arene adduct
[(dtbpe)Ni]2(μ−η2:η2-C6H6) 410 was found to rapidly generate
3 at room temperature in 60% isolated yield. In an attempt to
observe any intermediates during this transformation, we
performed low-temperature NMR studies on the reaction of 4
with 2.7b Even at −50 °C, the main species in solution is 3, which
highlights the facile oxidative addition process. In addition, 31P
and 13C NMR experiments reveal the transient existence of a
nickel(0) complex, which we assign as the asymmetric η2-ketone
complex 5 (Scheme 3). Warming the solution to −15 °C resulted
in the disappearance of the resonances of 5 in the 31P NMR
spectrum and the complete conversion to 3. This shows that 5 is
either an intermediate along the pathway to 3 or possibly in an
off-path equilibrium with some intermediate along the reaction
pathway to 3. In contrast, the reaction of 1 with 2 was found to
not proceed below room temperature.
potential reversibility of nickelaoxetane formation, an excess of 7
was added to a solution of 3 in C6D6. No formation of 9 was
detected after 3 days at room temperature, which indicates that
the reaction is irreversible under these conditions.
The Jamison group has previously proposed that low-valent
nickel can react with epoxides via an SN2-type attack, resulting in
an inversion of configuration at the carbon.3b Hillhouse has also
reported a similar mechanism for the oxidative addition of
(bpy)Ni(COD) (bpy = 2,2′-bipyridyl) to a variety of N-
tosylaziridines to generate azanickelacyclobutanes.9,13 We
believe that in our system 3 is formed by a different mechanism,
as the oxetane retains the cis configuration of the epoxide. In
order to further probe the mechanism of our transformation,
DFT calculations were performed, and two possible mechanistic
pathways are shown in Figure 2.14 Ligand dissociation from A or
B leads to the formation of the unsaturated nickel(0) complex C,
which is trapped by the free epoxide to form an η2-ketone
complex D that is stabilized by 11.7 kcal/mol (ΔG) relative to B.
Two conformers are possible for the η2-ketone complex: D,
where the epoxide oxygen is located on the same face of the six-
membered ring as the Ni center, and D′, where the oxygen atom
is found on the opposing face. Attempts to locate a concerted C−
O oxidative addition pathway directly from intermediate D were
unsuccessful. Instead, an intermediate E (ΔG = 10.2 kcal/mol)
was located on the potential energy surface (PES), best described
as an η5-O,C,C,C,H complex.11 Oxidative addition proceeds
through transition state EFTS (ΔG = 24.5 kcal/mol, D → EFTS)
to generate the metallaoxetane F. The connectivity of D with E
was not established, and therefore it is possible that the η2-ketone
complex is formed in an off-pathway equilibrium, and
intermediate E is formed directly from free C and epoxide.
Although we were unable to locate a transition state for SN2-
type cleavage from conformer D′, the zwitterionic ring-opened
product of such as process was found to be high in energy (+27.9
a
Scheme 3. Low-Temperature Synthesis of 3
a
Relative stereochemistry shown for clarity.
B
DOI: 10.1021/jacs.5b06735
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Figure 2. DFT-calculated reaction pathway for stepwise (blue) and concerted (black) metallaoxetane formation. Energies (ΔG) are reported in kcal/
mol relative to B.
kcal/mol relative to D′). We were surprised to find that
(dtbpe)nickel(0) could act effectively to stabilize this inter-
mediate, lowering its relative energy by 24.7 kcal/mol.11 In light
of this, we explored the possibility of a stepwise ring-opening/
closing mechanism that would allow for retention of stereo-
chemistry. In model calculations involving a truncated dmpe
(dmpe = 1,2-bis(dimethyl)phosphinoethane) ligand, a bimetallic
ketone complex was identified in which the epoxide oxygen binds
a second (dmpe)nickel(0) fragment. For the energetics in this
discussion we assume the second nickel(0) fragment is generated
by dissociation from B.11 Formation of the bimetallic species is
endergonic (+8.0 kcal/mol compared to the η2-ketone complex),
and the intermediate exists in a shallow well on the PES, with
epoxide opening proceeding through a transition state that was
calculated to have a slightly negative ΔG value.11 In calculations
involving the dtbpe ligand, no epoxide complex was located, and
ring opening was found to be barrierless upon approach of the
nickel(0) fragment toward the epoxide oxygen of D′. The
resultant ring-opened structure bmII (ΔG = −17.2 kcal/mol) is
an η3-oxa-allylnickel complex, analogous to those previously
described.7a,15 No transition state was located for subsequent
ring closure leading to metallaoxetane bmIII (ΔG = −26.7 kcal/
mol), but we expect this process to be facile.11 In the formation of
bmIII the oxa-allylic nickel readopts a η2-ketone binding mode.
Formation of the observed product F and regeneration of B by
loss of nickel(0) from the carbonyl is calculated to have a
negligible free energy change. Notably, no conversion to bmIII
was observed experimentally upon addition of arene complex 4
to 3. The lack of reactivity between complexes 1 and 4 with
cyclohexene oxide demonstrates that the ketone group plays a
crucial role in this transformation. DFT analysis reveals that the
transition-state energy (ΔG⧧ = 15.4 kcal/mol) for concerted C−
O oxidative addition of cyclohexene oxide is similar to that for
2.11 This observation is consistent with a bimetallic stepwise
mechanism, as precoordination of one of the metal centers as
well as formation of the η3-oxa-allylnickel motif upon ring-
opening are expected to lower the reaction barrier. In addition,
the lower energetics of the bimetallic pathway are more
consistent with the rapid rate of formation of 3 that we observe.
Although our DFT calculations favor a stepwise mechanism, we
cannot conclusively rule out a concerted mechanism at this time,
and further mechanistic studies are ongoing within our group.
We then sought to probe the reactivity of 3 (Scheme 5).
Protonolysis of 3 with an excess of HCl in C6D6 resulted in the
rapid precipitation of a red solid, identified as (dtbpe)NiCl2
Scheme 5. Preliminary reactivity of 3. (a) isolated yield (b)
NMR yield (c) GC yield.11
(11)16 and isolated in 94% yield. The organic product was
identified by GC-MS and 1H NMR spectroscopy as 2-
cyclohexenone 12 (69% yield), presumably formed via acid-
induced elimination from liberated 3-hydroxycyclohexanone
(13).17 Indeed, using less acidic sources of protons such as
MeOH results in the formation of 13 in 59% yield by 1H NMR
spectroscopy. Treatment of 3 with Ph3CBF4 gave complex 14 as
the major organometallic product via deoxygenation of the
nickelaoxetane ring. Complex 14 was formed in 36% yield, and
singly oxidized phosphine ligand (dtbpeO) was also observed by
31P NMR spectroscopy.18 The crude reaction mixture also
contained 12 (26% yield), free dtbpe, and a dark precipitate
presumed to be Ni black. Alkene extrusions have been reported
for a few other 2-metallaoxetanes.19 Complex 14 was also
prepared independently (see Figure 3 for solid-state structure).11
Figure 3. POV-Ray diagrams of 14 (50% ellipsoids) and 17 (30%
ellipsoids). Certain hydrogen atoms omitted for clarity.
C
DOI: 10.1021/jacs.5b06735
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Stirring a solution of 3 in C6D6 under 1 atm of CO for 16 h
resulted in a color change from orange-red to pale yellow.
Analysis of the reaction mixture by 31P NMR spectroscopy
revealed the complete formation of (dtbpe)Ni(CO)2 (15). In
REFERENCES
(1) Dauth, A.; Love, J. A. Chem. Rev. 2011, 111, 2010.
■
(2) De Pasquale, R. J. J. Chem. Soc., Chem. Commun. 1973, 157.
(3) (a) Molinaro, C.; Jamison, T. F. J. Am. Chem. Soc. 2003, 125, 8076.
(b) Beaver, M. G.; Jamison, T. F. Org. Lett. 2011, 13, 4140.
(4) Nielsen, D. K.; Doyle, A. G. Angew. Chem., Int. Ed. 2011, 50, 6056.
(5) (a) Mindiola, D. J.; Hillhouse, G. L. J. Am. Chem. Soc. 2002, 124,
9976. (b) Mindiola, D. J.; Waterman, R.; Iluc, V. M.; Cundari, T. R.;
Hillhouse, G. L. Inorg. Chem. 2014, 53, 13227.
(6) (a) de Bruin, B.; Boerakker, M. J.; de Gelder, R.; Smits, J. M. M.;
Gal, A. W. Angew. Chem., Int. Ed. Engl. 1997, 36, 2064. (b) Dauth, A.;
Love, J. A. Angew. Chem., Int. Ed. 2010, 49, 9219. (c) Desnoyer, A. N.;
Dauth, A.; Patrick, B. O.; Love, J. A. Dalton Trans. 2014, 43, 30.
(7) (a) Tamaki, T.; Nagata, M.; Ohashi, M.; Ogoshi, S. Chem. - Eur. J.
2009, 15, 10083. (b) Hoshimoto, Y.; Yabuki, H.; Kumar, R.; Suzuki, H.;
Ohashi, M.; Ogoshi, S. J. Am. Chem. Soc. 2014, 136, 16752.
(8) (a) García, J. J.; Jones, W. D. Organometallics 2000, 19, 5544.
(b) García, J. J.; Brunkan, N. M.; Jones, W. D. J. Am. Chem. Soc. 2002,
124, 9547.
20
contrast, reacting 3 with CO2 in CD2Cl2 yields (dtbpe)Ni-
(CO3)21 (16) as the major product (>90% yield by 31P NMR
spectroscopy) along with 12 (75% yield by 1H NMR
spectroscopy).11 If the reaction is performed in aromatic solvents
at 50 °C, orange crystals of 17 which are suitable for X-ray
diffraction analysis (Figure 3) can also be obtained from the
reaction mixture. While the insertion of a CO2 unit does result in
ring expansion, the bicyclic moiety of 3 appears to have
undergone an elimination reaction to form an alkene. This group
is clearly seen in the 1H NMR spectrum of 17.11 We have been
unable to characterize complex 17 fully, as it is formed in low
yields (<5% by 1H NMR spectroscopy) and readily decomposes.
Oxidatively induced reductive elimination has been well-
documented for first-row transition metals, including nickel.13,22
Addition of I2 to 3 results in a color change from red-orange to
dark green. Analysis of the resulting solution by EI-MS and UV−
vis spectroscopy reveals the formation of paramagnetic (dtbpe)-
(9) Reacting 1 and 2 in the presence of added COD results in a
suppressed rate of formation of 3. For an example of COD affecting
nickel catalysis, see: Jensen, K. L.; Standley, E. A.; Jamison, T. F. J. Am.
Chem. Soc. 2014, 136, 11145.
23
NiI2 (18), isolated in 93% yield. GC analysis of the crude
(10) Bach, I.; Porschke, K.-R.; Goddard, R.; Kopiske, C.; Kruger, C.;
̈
̈
mixture shows the presence of 2 in 81% yield.
Rufin
́
ska, A.; Seevogel, K. Organometallics 1996, 15, 4959.
In conclusion, we have demonstrated that ketones act as
directing groups for the oxidative addition of nickel(0) with
epoxides to give rare examples of well-defined 2-nickelaoxetanes.
DFT calculations were performed to probe the mechanism of
formation of these nickelaoxetanes, which support a bimetallic
ring-opening/closing pathway over a concerted oxidative
addition. Complex 3 was found to be susceptible to protonolysis,
oxidatively induced reductive elimination, deoxygenation, and
elimination reactions when treated with the appropriate reagents.
We are currently undertaking further mechanistic studies on the
reactions reported herein and are also exploring the possibility of
catalytic reactivity with these complexes.
(11) See Supporting Information for details.
(12) A similar mechanism has been proposed using platinum(II): Aye,
K.-T.; Gelmini, L.; Payne, N. C.; Vittal, J. J.; Puddephatt, P. J. J. Am.
Chem. Soc. 1990, 112, 2464.
(13) Lin, B. L.; Clough, C. R.; Hillhouse, G. L. J. Am. Chem. Soc. 2002,
124, 2890.
(14) For related DFT studies of nickel-mediated transformations, see:
(a) Plessow, P. N.; Weigel, L.; Lindner, R.; Schafer, A.; Rominger, F.;
Limbach, M.; Hofmann, P. Organometallics 2013, 32, 3327. (b) Atesi̧ n,
T. A.; Li, T.; Lachaize, S.; Brennessel, W.; Garcia, J. J.; Jones, W. D. J. Am.
Chem. Soc. 2007, 129, 7562.
̈
(15) (a) Ho, C.-Y.; Ohmiya, H.; Jamison, T. F. Angew. Chem., Int. Ed.
2008, 47, 1893. (b) Liu, L.; Montgomery, J. J. Am. Chem. Soc. 2006, 128,
5348.
(16) Bach, I.; Goddard, R.; Kopische, C.; Seevogel, K.; Porschke, K.-P.
Organometallics 1999, 18, 10.
(17) Kawakami, T.; Shibata, I.; Baba, A. J. Org. Chem. 1996, 61, 82.
(18) Anderson, J. S.; Iluc, V. M.; Hillhouse, G. L. Inorg. Chem. 2010, 49,
10203.
̈
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.5b06735.
■
S
(19) (a) Szuromi, E.; Wu, J.; Sharp, P. R. J. Am. Chem. Soc. 2006, 128,
12088. (b) Gable, K. P.; Phan, T. N. J. Am. Chem. Soc. 1994, 116, 833.
(20) (a) Aye, K.-T.; Ferguson, G.; Lough, A. J.; Puddephatt, R. J.
Angew. Chem., Int. Ed. Engl. 1989, 28, 767. (b) Guo, C.-H.; Song, J.-Y.;
Jia, J.-F.; Zhang, X.-M.; Wu, H.-S. Organometallics 2010, 29, 2069 See
also ref 12.
Computational coordinates and full experimental details
(PDF)
Crystallographic data (CIF)
AUTHOR INFORMATION
Corresponding Author
*jenlove@chem.ubc.ca
■
(21) Gonzal
metallics 2012, 31, 8200.
́ ́
ez-Sebastian, L.; Flores-Alamo, M.; García, J. J. Organo-
(22) (a) Han, R.; Hillhouse, G. L. J. Am. Chem. Soc. 1997, 119, 8135.
(b) Higgs, A. T.; Zinn, P. J.; Simmons, S. J.; Sanford, M. S.
Organometallics 2009, 28, 6142.
Notes
The authors declare no competing financial interest.
(23) Schultz, M.; Plessow, P.-N.; Rominger, F.; Weigel, L. Acta
Crystallogr., Sect. C: Cryst. Struct. Commun. 2013, 69, 1437.
ACKNOWLEDGMENTS
■
We thank the University of British Columbia, NSERC
(Discovery grant, Research Tools and Instrumentation grants)
and the Canada Foundation for Innovation for supporting this
research. A.N.D. is grateful to NSERC for a CGS-D and to the
Izaak Walton Killam foundation for a doctoral scholarship.
E.G.B. is grateful to NSERC for a CGS-M and to the
Government of Canada for a Vanier Canada graduate scholar-
ship. We thank Prof. Pierre Kennepohl, Prof. Andrei Vedernikov,
Marcus W. Drover, and Dr. Shrinwantu Pal for fruitful
discussions. This work was inspired by the scientific career of
Gregory L. Hillhouse and is dedicated to his memory.
D
DOI: 10.1021/jacs.5b06735
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX