Regio- and chemoselective rearrangement of terminal epoxides into methyl alkyl and aryl ketones

Article abstract of DOI:10.1039/C8CC06503A

The development of the highly active pincer-type rhodium catalyst 2 for the nucleophilic Meinwald rearrangement of functionalised terminal epoxides into methyl ketones under mild conditions is presented. An excellent regio- and chemoselectivity is obtained for the first time for aryl oxiranes.

Full text of DOI:10.1039/C8CC06503A

ChemComm
View Article Online
View Journal
COMMUNICATION
Regio- and chemoselective rearrangement of
terminal epoxides into methyl alkyl and aryl
Cite this:DOI: 10.1039/c8cc06503a
ketones†
Received 9th August 2018,
Accepted 18th September 2018
Yingying Tian,
Eva Jurgens and Doris Kunz
*
¨
DOI: 10.1039/c8cc06503a
rsc.li/chemcomm
The development of the highly active pincer-type rhodium catalyst addition of a methyl-Grignard reagent to the aldehyde, followed
2 for the nucleophilic Meinwald rearrangement of functionalised by alcohol oxidation.
terminal epoxides into methyl ketones under mild conditions is
So far all nucleophilic Meinwald reactions require a pre-
presented. An excellent regio- and chemoselectivity is obtained for activation of the epoxide by Lewis acids. In the first report in
the first time for aryl oxiranes.
1962 by Eisenmann,7a the Co(II)–Lewis acid co-catalyst and the
nucleophilic [Co(CO)4]ꢀ catalyst are formed in situ by dispro-
Epoxides are important intermediates in organic synthesis as portionation of [Co2(CO)8] in methanol. A similar principle is
they undergo various useful transformations.1 Among those, likely applicable to the case of Pd-catalysis, in which the Pd(0)
the isomerisation to carbonyl compounds (Meinwald reaction) catalyst is formed in situ from Pd(II).7e The presence of the latter
has gained increased attention, and various kinds of catalysts can explain the Lewis acid catalysed selectivity in favour of
have been reported so far.2,3 Most abundant is the Lewis acid aldehydes in the case of aryl oxiranes. Nucleophiles like iodide
catalysed isomerisation in which the selectivity is determined and bromide can also serve as catalysts in combination with
by formation of the most stable carbenium intermediate followed Lewis acids like Li+ 7b–d or Sm2+.7f Recently, we have shown that
by an H-shift that leads to aldehydes as the major product in the the combination of the nucleophilic Rh-pincer complex 18
case of terminal epoxides (Scheme 1a).4–6 The inverse selectivity (Fig. 1) and LiNTf2 is suitable for the regioselective isomerisation
is obtained with nucleophilic catalysts that attack the more of terminal alkyl epoxides to methyl ketones9 and soon after
electrophilic site (usually the less substituted one) of the epoxide. Coates reported the highly active complex [Al(porphyrin)]+
After rearrangement via a formal H-shift, methyl ketones are [Co(CO)4]ꢀ.7g In both cases, the isomerisation of phenyl oxirane
formed from terminal epoxides (Scheme 1b).7 This selectivity is led only to mixtures of phenyl ethanal and phenyl methyl
very attractive, as it could substitute the Wacker oxidation by a ketone (2 : 3 (Rh), 3 : 2 (Co)), which indicates a competition
two-step epoxidation-isomerisation sequence for temperature between the Lewis acidic and the nucleophilic pathway. Therefore,
and/or Lewis acid-sensitive olefins.7d Moreover, combining the enhancing the selectivity for phenyl oxiranes as well as increasing
Johnson–Corey–Chaykovsky epoxidation with the isomerisation the activity of the catalyst was the primary goal of our investigations,
would provide a very mild and oxidation-free transformation of whose results are reported in the following.
aldehydes into methyl ketones, which is usually carried out by
Catalyst 1 requires elevated temperature (60 1C) and a co-catalyst
(20 mol% LiNTf2) to achieve full conversion within 2 h in the
Scheme 1 Catalytic isomerisation pathways of terminal epoxides: formation
of aldehydes versus methyl ketones.10
¨
¨
¨
Institut fur Anorganische Chemie, Eberhard Karls Universitat Tubingen, Auf der
¨
Morgenstelle 18, 72076 Tubingen, Germany. E-mail: Doris.Kunz@uni-tuebingen.de
Fig. 1 Examples of nucleophilic epoxide isomerisation catalysts.
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c8cc06503a
This journal is ©The Royal Society of Chemistry 2018
Chem. Commun.

View Article Online
Communication
ChemComm
case of terminal alkyl oxiranes.9 An additional substrate scope coupling of 33.9 (carbene), 6.7 (C15) and 11.3 Hz (C14). DFT
(see ESI†) revealed a high functional group tolerance for many calculations also predict a pentacoordinating ligand with
substrates, but a substantial drop in conversion and/or yield was both olefin moieties contributing to the trigonal bipyramidal
recognised. To overcome these drawbacks we supposed that a coordination mode at the Rh centre and an orientation in line
CO-free Rh-catalyst would further increase the nucleophilicity with the trigonal base.14 The four 1H NMR signals for the
of the metal centre and thus reduce the need for a Lewis acid methylene protons (H-12, H-13) of the homoallyl moiety were
co-catalyst. However, all attempts to synthesise a CO-free version assigned according to their coupling constants and NOE to the
of 1 by using [Rh(m-Cl)(COD)]2 or [Rh(m-Cl)(C2H4)2]2 as starting axial and equatorial protons of the six membered metallacycles.
material failed and revealed a complex product mixture in the The signals of the olefinic protons are strongly shielded
NMR spectra. Increasing the steric bulk at the carbene by N-iPr (4.18–4.10 (H-14), 2.41 (H-15cis), 1.67 ppm (H-15trans)) and the
groups lead only to a bridged dinuclear complex.11 Therefore, reduced coupling constants of only 8.0 (cis) and 9.9 Hz (trans)
we tried to reversibly coordinate olefin moieties intramolecularly, can be explained by the reduced s-character due to the coordination
which could also stabilise the complex during catalysis and with the Rh-centre.
increase its lifetime.
With the rhodium complex 2 in hand, we tested the iso-
We found the earlier reported N-homoallyl substituted merisation of 1,2-epoxyhexane (3b) as the model substrate. An
ligand bimcaHomo 12 to fulfil these criteria (Scheme 2). Prepara- initial test applying 5 mol% of complex 2LiX in THF-d8 revealed
tion of 2 was achieved by deprotonation of HbimcaHomoꢁ2HBr only low conversion at room temperature (Table 1, entry 1),
with LiHMDS and subsequent addition of [Rh(m-Cl)(COD)]2. which improved by raising the reaction temperature to 60 1C
Removal of COD in vacuo yielded 2 along with LiBr and LiCl and adding LiNTf2 as co-catalyst (Table 1, entry 2). Interestingly,
quantitatively. As the lithium salts turned out to be necessary full regioselectivity and quantitative yields were obtained in
for the catalytic activity of 2 (vide infra), we did not remove short reaction times at room temperature when toluene-d8
them, but used this mixture 2LiX for catalysis. To obtain the pure (3 h) or C6D6 (2 h) were used as solvent (Table 1, entries 3 and 4).
complex 2, we transmetalated from the potassium complex We suppose that the poor conversion in THF-d8 derives from
[K(bimcaHomo)] generated in situ by deprotonation of HbimcaHomo
2HBr with KHMDS.13 The potassium halides formed were readily Lewis acid co-catalyst.15 To test whether the residual lithium
removed by filtration. halides are still necessary as a Lewis acid co-catalyst we applied
ꢁ
competitive binding between THF and the substrate to the
The molecular structure of 2 (Fig. 2) is revealed by NMR the isolated catalyst 2 in C6D6. No catalytic activity was found
spectroscopic analysis and DFT calculations. The single signal under these conditions (Table 1, entry 5). Surprisingly, addition
set confirms a symmetric coordination mode of the ligand. of 10 mol% of LiBr achieved no improvement, but small
Coordination of both double bonds and both carbene moieties amounts of THF (20 mL) restored the catalytic activity fully,
to the Rh-centre is confirmed not only by the up-field shift of
the respective signals in the 13C NMR spectrum to 185.5 ppm
1
Table 1 Rh-Catalyzed isomerisation of 1,2-epoxyhexane (3b): optimisa-
tion of the reaction conditionsa
(carbene), 55.9 (C15) and 51.1 ppm (C14), but also by the JRhC
Catalyst [mol%] +
Entry additive [mol%]
Conc. 3b
Yieldb
Time [%]
Solvent
[mol Lꢀ1
]
1
5 (2LiX
5 (2LiX
5 (2LiX
5 (2LiX
5 (2)
5 (2) + 10 (LiBr)
5 (2) + 10 (LiBr)
10 (LiBr)
)
)
)
)
THF-d8
THF-d8
Toluene-d8 0.1
0.1
0.1
2 h
5.5 d
3 h
2 h
2 h
1 h
3 h
24 h
24 h
3 h
8 h
9 h
7 d
4 d
o5
88
499
499
0
0
499
0
2c,f
3
4
5
C6D6
C6D6
C6D6
C6D6
C6D6
C6D6
C6D6
C6D6
C6D6
C6D6
C6D6
C6D6
C6D6
C6D6
C6D6
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.2
0.4
1.0
6
Scheme 2 Synthesis of the rhodium pincer complex 2.
7d
8
9d
10
11
12
13
14e
15f
16
17
18
10 (LiBr)
0
4 (2LiX
3 (2LiX
2 (2LiX
1 (2LiX
1 (2LiX
1 (2LiX
1 (2LiX
1 (2LiX
1 (2LiX
)
)
)
)
)
)
)
)
)
499
499
499
499
87
3 d
93
24 h
14 h
3 h
499
499
98
a
b
Carried out in J. Young NMR tubes. Yield (1H NMR) of 4b calibrated
Fig. 2 Calculated structure of 2 with tert-Bu groups depicted in wire-
frame for clarity.
to 1,3,5-trimethoxybenzene (internal standard); conversion = yield.
With 20 mol% LiTNf2. With 20 mL THF. At 40 1C. At 60 1C.
c
d
e
f
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2018

View Article Online
ChemComm
Communication
which can be explained by the enhanced solubility of LiBr16 acidic protons, nonetheless 4f and 4g were formed in 99% and 97%
(entries 6 and 7). As lithium halides are known catalysts for the yields. Notably, catalyst 2LiX is well suited for the isomerisation of
Meinwald reaction, albeit at elevated temperature,7b–d we tested terminal epoxides 3h–k with ester groups and afforded 4h–k with
them under our conditions at room temperature. Neither high regio- and chemoselectivity. 3j reacts a bit slower than 3h,
lithium bromide nor solubilised lithium bromide (by 20 mL which might be due to its a-acidity. Furthermore, 2-benzyloxirane
THF) were able to catalyse this reaction (entries 8 and 9). These (3l) and its methoxy (3m) and trifluoromethyl (3n) congeners were
results clearly show that catalyst 2 is much more active than isomerised to the methyl ketones 4l–n quantitatively. Internal
catalyst 1 and that at least a weak Lewis acid co-catalyst is still epoxides react considerably slower.§
necessary.
The selective rearrangement of aryl oxiranes to methyl ketones
Lowering the catalyst loading from 5 mol% to 1 mol% at has not been achieved so far due to the Lewis acid catalysed side
constant concentration of the substrate did not affect the reaction.7e, f,9 Therefore, we applied our new catalyst 2LiX also in
regioselectivity but required a longer reaction time (Table 1, the isomerisation of aryl oxiranes (Table 2). With styrene oxide
entries 10–13). A higher temperature accelerated the reaction, (5a) we found an unprecedented good ratio between aceto-
but it still needed several days for completion using 1 mol% of phenone (6a) and 2-phenylacetaldehyde (7a) of 40 : 1. To our delight,
catalyst (Table 1, entries 14 and 15). Epoxide concentrations of acetophenones 6b–d bearing the strong electron-withdrawing
0.4 mol Lꢀ1 and 1.0 mol Lꢀ1 (Table 1, entries 16–18) shortened trifluoromethyl substituent were obtained regioselectively in
the reaction time substantially and the reaction is about excellent yield. This counts also for the terminal epoxides 5e–j
10 times faster compared to literature.7g Even though 1 mol% possessing fluoro, chloro and bromo substituents (499 : 1). Aryl
of catalyst loading was sufficient for the model reaction at high oxiranes containing electron-donating groups (methyl or methoxy)
concentrations, further experiments showed that it may lead to were converted into 6k and 6l with still very good regioselectivities
polymerisation of functionalised epoxides (especially with ester of 21 : 1 and 10 : 1 favouring the methyl ketones (6k, 6l) over the
groups). Thus, the optimised reaction conditions (5 mol% 2LiX
0.1 M epoxide, C6D6, room temperature) were applied to explore extremely sensitive towards Lewis acids in C6D6 at room
,
aldehydes (7k, 7l) in good to moderate yields. Notably, epoxide 5l is
the generality of this protocol.‡
temperature. In two control experiments without catalyst 2,
Using complex 2LiX as the catalyst, 13 functionalised terminal 10 mol% of LiBr and 5 mol% of LiCl as well as 10 mol% LiI were
epoxides were converted successfully into the desired methyl added to the aryl oxirane 5l in C6D6. After few days (LiBr/Cl) or
ketones (Scheme 3). The alkyl substituted epoxides 3a–d and the
methoxy epoxide 3e were transformed into the desired products
4a–e almost quantitatively with full regioselectivity. The only
Table 2 Regio- and chemoselective isomerisation of aryl oxiranes 5a–ma
exception was tert-butyloxirane (3c), which could be obtained in
only 11% yield after 10 days, possibly due to steric effects. In the
case of hydroxyl (3f) and sulfonamide groups (3g) the isomerisation
progressed slowly, which might reveal a competing influence of the
6 yieldb [%]
6a, 95
Ratio (6 : 7)
Epoxide 5
Entry
1
5a
40 : 1
2
5b
5c
5d
m-
o-
p-
6b, 98
6c, 91
6d, 99
499 : 1
499 : 1
499 : 1
3c
4
5
6
7
5e
5f
5g
F-
Cl-
Br-
6e, 99
6f, 99
6g, 99
499 : 1
499 : 1
499 : 1
8
9
10
5h
5i
5j
F-
Cl-
Br-
6h, 85
6i, 92
6j, 82
91 : 1
499 : 1
499 : 1
11
12
5k
5l
6k, 77
6l, 57
21 : 1
10 : 1
7 : 1
13d
5m
6m, 75
Scheme 3 Rh-Catalyzed regio- and chemoselective isomerisation of
terminal epoxides. Standard reaction conditions: substrate (50.0 mmol),
2LiX (5 mol%), C6D6 (0.5 mL), r.t. at the given time. Carried out in J. Young
NMR tubes. Yield (1H NMR) calibrated to 1,3,5-trimethoxybenzene (internal
standard). a 14% of unreacted epoxide left. b Isolated yield after column
chromatography of a 1 mmol scale experiment.
a
Standard reaction conditions: epoxide (50.0 mmol), 2LiX (5 mol%),
C6D6 (0.5 mL), r.t. Carried out in J. Young NMR tubes. Yields (1H NMR)
calibrated to 1,3,5-trimethoxybenzene (internal standard). At 80 1C for
2 days. 22 h.
b
c
d
This journal is ©The Royal Society of Chemistry 2018
Chem. Commun.

View Article Online
Communication
ChemComm
Er3+: (e) A. Procopio, R. Dalpozzo, A. De Nino, M. Nardi, G. Sindona and
A. Tagarelli, Synlett, 2004, 2633; Sn: ( f ) M. Banerjee, U. K. Roy,
23 h (LiI), only formation of the aldehyde 7l and other, possibly
polymerisation products (Fig. S5, ESI†), was observed. The oxirane
5m bearing a +I substituent in ortho position can be transformed
into 6m with still good regioselectivity and in good yield after 22 h.
In conclusion, the new CO-free rhodium complex 2 led to
significant improvement of the nucleophilic Meinwald reaction
with respect to lower reaction temperature, catalyst loading and
the absence of additional Lewis acids. It also shows a high
functional group tolerance. The stronger nucleophilicity of
complex 2 is crucial for the excellent regioselectivity achieved
in the case of aryl oxiranes, which can get isomerised for the
first time almost exclusively to the methyl ketones. Thus, we
have broadened the scope of this reaction, which should be very
valuable for organic synthesis, especially in combination with
the Johnson–Corey–Chaykovsky reaction.
P. Sinha and S. Roy, J. Organomet. Chem., 2005, 690, 1422; Cu2+
:
(g) M. W. C. Robinson, K. S. Pillinger, I. Mabbett, D. A. Timms and
A. E. Graham, Tetrahedron, 2010, 66, 8377; [ReBr(CO)3]: (h) R. Umeda,
M. Muraki, Y. Nakamura, T. Tanaka, K. Kamiguchi and Y. Nishiyama,
Tetrahedron Lett., 2017, 58, 2393.
5 The chelation enhanced formation of methyl ketones from a-keto
oxiranes by a Ru2+ catalyst: C.-L. Chang, M. P. Kumar and R.-S. Liu,
J. Org. Chem., 2004, 69, 2793.
6 Pd2+: (a) D. J. Vyas, E. Larionov, C. Besnard, L. Guenee and C. Mazet,
´ ´
J. Am. Chem. Soc., 2013, 135, 6177; (b) N. Humbert, D. J. Vyas,
C. Besnard and C. Mazet, Chem. Commun., 2014, 50, 10592;
(c) S. Kulasegaram and R. J. Kulawiec, J. Org. Chem., 1997, 62, 6547;
(d) S. Kulasegaram and R. J. Kulawiec, Tetrahedron, 1998, 54, 1361; Ni in a
reaction sequence: (e) A. N. Desnoyer, J. L. Geng, M. W. Drover,
B. O. Patrick and J. A. Love, Chem. – Eur. J., 2017, 23, 11509.
7 (a) J. L. Eisenmann, J. Org. Chem., 1962, 27, 2706; (b) B. Rickborn
and R. M. Gerkin, J. Am. Chem. Soc., 1968, 90, 4193; (c) B. Rickborn
and R. M. Gerkin, J. Am. Chem. Soc., 1971, 93, 1693; (d) Z.-w. An,
R. D’Aloisio and C. Venturello, Synthesis, 1992, 1229; (e) S. Kulasegaram
and R. J. Kulawiec, J. Org. Chem., 1994, 59, 7195; ( f ) J. Prandi, J. L. Namy,
G. Menoret and H. B. Kagan, J. Organomet. Chem., 1985, 285, 449;
(g) J. R. Lamb, Y. Jung and G. W. Coates, Org. Chem. Front., 2015, 2, 346.
8 M. Moser, B. Wucher, D. Kunz and F. Rominger, Organometallics,
2007, 26, 1024.
Yingying Tian thanks the China Scholarship Council (CSC)
for a predoctoral fellowship and Eva Ju¨rgens the MWK-BW for
¨
funding (Landesgraduiertenforderung). We thank Alexander
Klaiber, Nicolas Wiedmaier and Mario R. Rapp for help in
synthesising some of the substrates and in the catalyst testing.
¨
9 E. Ju¨rgens, B. Wucher, F. Rominger, K. W. Tornroos and D. Kunz,
Chem. Commun., 2015, 51, 1897.
10 The nucleophilic pathway (Scheme 1) follows an SN2 mechanism.
Formation of metallaoxetanes by insertion into the s-carbon oxygen
bond: (a) A. Dauth and J. A. Love, Chem. Rev., 2011, 111, 2010. For
Ni(0) see Lit. 4a and 6e as well as: (b) A. N. Desnoyer, E. G. Bowes,
B. O. Patrick and J. A. Love, J. Am. Chem. Soc., 2015, 137, 12748; Rh:
(c) M. J. Calhorda, A. M. Galvao, C. Unaleroglu, A. A. Zlota, F. Frolow
and D. Milstein, Organometallics, 1993, 12, 3316; and ; (d) B. de
Bruin, M. J. Boerakker, J. J. J. M. Donners, B. E. C. Christiaans,
P. P. J. Schlebos, R. de Gelder, J. M. M. Smits, A. L. Spek and
A. W. Gal, Angew. Chem., Int. Ed., 1997, 36, 2063; (e) B. de Bruin,
M. J. Boerakker, J. A. W. Verhagen, R. de Gelder, J. M. M. Smits and
A. W. Gal, Chem. – Eur. J., 2000, 6, 298. Without additional Lewis
acid (20-oxoalkyl)(hydrido)rhodium(III) intermediates were isolated
reacting RhCl(PMe3)3 in neat propylene or styrene oxide after 16 h:
( f ) D. Milstein, J. Am. Chem. Soc., 1982, 104, 5227; Ir: (g) D. Milstein
and J. C. Calabrese, J. Am. Chem. Soc., 1982, 104, 3773. We tested
RhCl(PMe3)3 under our conditions and found it considerably slower
than complex 2LiX: 5% yield for styrene oxide after 2 h (with or
without addition of Brꢀ); after 24 h still 5% yield without Brꢀ and
16% in presence of 10 mol% LiBr. For nucleophilic mechanisms via
C–H bond activation: (h) D. Milstein, O. Buchman and J. Blum,
J. Org. Chem., 1977, 42, 2299; (i) J. Wu and R. G. Bergman, J. Am.
Chem. Soc., 1989, 111, 7628.
Conflicts of interest
The authors declare no conflict of interest.
Notes and references
‡ Experimental procedure for in situ generation of catalyst 2LiX: Li(N(SiMe3)2)
(7.4 mg, 44.0 mmol) was added to a suspension of HbimcaHomoꢁ2HBr
(10.0 mg, 14.7 mmol) in 0.5 mL of THF-d8 at room temperature. After
10 min, [Rh(m-Cl)(COD)]2 (3.6 mg, 7.3 mmol) was added and the solution
was stirred for another 10 min. After checking the successful formation
of the catalyst by NMR, 85.0 mL (containing 2.5 mmol 2LiX) of the freshly
prepared catalyst solution were used for each NMR experiment and the
THF was removed in oil-pump vacuum prior to the addition of the
epoxide (see ESI† for details).
§ At 80 1C, but otherwise identical conditions, 77% cis-2,3-epoxybutane
and 34% trans-2,3-epoxybutane were converted into the ketone after
10 d. No further optimisation was attempted.
1 J. G. Smith, Synthesis, 1984, 629.
2 (a) B. C. Hartman and B. Rickborn, J. Org. Chem., 1972, 37, 943;
(b) B. Rickborn, in Compr. Org. Synth., ed. B. M. Trost, Pergamon, 11 E. Ju¨rgens, PhD thesis, University of Tu¨bingen, Tu¨bingen, 2017.
Oxford, 1991, vol. 3, pp. 733–775; (c) Y. Kita, S. Kitagaki, Y. Yoshida, 12 Bimca (3,6-di-tert-butyl-1,8-bis(imidazolin-2-ylidene)-9-carbazolide).
S. Mihara, D.-F. Fang, M. Kondo, S. Okamoto, R. Imai, S. Akai and
E. Ju¨rgens and D. Kunz, Eur. J. Inorg. Chem., 2017, 233.
H. Fujioka, J. Org. Chem., 1997, 62, 4991; (d) C.-Y. Huang and A. G. Doyle, 13 In contrast to [Li(bimcaHomo)], complex [K(bimcaHomo)] decomposes at
Chem. Rev., 2014, 114, 8153.
room temperature into butadiene and the bisimidazole carbazolide
[K(imi)2carb] (see ESI†).
14 See ESI† for another, but energetically disfavoured minimum.
3 J. Meinwald, S. S. Labana and M. S. Chadha, J. Am. Chem. Soc., 1963,
85, 582.
4 Ni2+: (a) A. Miyashita, T. Shimada, A. Sugawara and H. Nohiya, Chem. 15 DCM and acetonitrile react already with 1: M. Moser, PhD thesis,
Lett., 1986, 1323; In3+: (b) B. C. Ranu and U. Jana, J. Org. Chem., 1998,
Heidelberg University, Heidelberg, 2007.
63, 8212; Fe3+: (c) K. Suda, K. Baba, S.-I. Nakajima and T. Takanami, 16 2LiX is generated in THF, which is then removed (with COD) in
Tetrahedron Lett., 1999, 40, 7243; Bi3+: (d) A. M. Anderson, J. M. Blazek,
P. Garg, B. J. Payne and R. S. Mohan, Tetrahedron Lett., 2000, 41, 1527;
vacuo. Some THF remains coordinated to Li+ so that the solubility of
LiX is enhanced during catalysis in C6D6 or toluene-d8.
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2018