Catalytic, oxidant-free, direct olefination of alcohols using Wittig reagents

Article abstract of DOI:10.1039/c5cc02902c

Reported here is the catalytic, acceptorless coupling of alcohols with in situ generated, non-stabilized phosphonium ylides to form olefins as major products. The reaction uses low catalyst loadings and does not require added oxidants. Hydrogenation of the product is minimized and the reaction leads to Z (aliphatic) or E (benzylic) stereospecificity.

Full text of DOI:10.1039/c5cc02902c

ChemComm
View Article Online
View Journal
COMMUNICATION
Catalytic, oxidant-free, direct olefination of
alcohols using Wittig reagents†
Cite this:DOI: 10.1039/c5cc02902c
ab
b
E. Khaskin and D. Milstein*
Received 8th April 2015,
Accepted 28th April 2015
DOI: 10.1039/c5cc02902c
www.rsc.org/chemcomm
Reported here is the catalytic, acceptorless coupling of alcohols form stilbenes, promoted by Ni nanoparticles which requires a
5i
with in situ generated, non-stabilized phosphonium ylides to form stoichiometric (or greater) amount of nickel, was reported.
olefins as major products. The reaction uses low catalyst loadings Direct catalytic reaction of an ylide with an alcohol to
and does not require added oxidants. Hydrogenation of the product selectively yield an olefin, with liberation of hydrogen gas,
is minimized and the reaction leads to Z (aliphatic) or E (benzylic) avoiding the use of oxidants, would be a significant advance
stereospecificity.
in light of the ubiquitous nature of the Wittig transformation in
modern organic synthesis (eqn (1)). It also might be a valuable
1
Since its discovery in 1953, the Wittig olefination reaction has transformation where the aldehyde is an unstable species
been recognized as one of the most important carbon–carbon under ambient conditions.
bond forming transformations in organic chemistry as it is
carried out under mild conditions and shows good tolerance
towards various functional groups. However, the reaction
0
0
R–CH
2
–OH + R –CHQPPh
3
- R–CHQCH–R + OPPh
3
+ H
1)
2
(
2
requires a carbonyl functionality for the ylide to react, and In the current report, we describe for the first time such a process.
3
good, selective reactivity is limited to aldehydes and ketones.
This new Ru catalyzed reaction is applicable to both benzylic and
A few years ago the Williams group developed an elegant aliphatic alcohols, and a good range of Wittig salts, giving the
borrowing hydrogen’ procedure for the synthesis of alkanes corresponding olefins in high yields (often quantitative conver-
‘
from alcohols, using an indirect Wittig reaction where the sion and 480% isolated yields after chromatography), using
alcohol moiety was dehydrogenated by an iridium or ruthe- 1–0.02 mol% catalyst, with evolution of H . The reaction is
2
nium catalyst into an aldehyde intermediate that reacted with a tolerant of a number of functional groups including nitriles,
3
stabilized ylide, leading to alkanes as major products, rather CF , double bonds, amines, and to a lesser extent, of pyridines
4
than olefins. Aiming at olefins, a number of one-pot examples, and halogens. To our knowledge, the catalytic conversion of
where in situ oxidation by various oxidants is carried out with alcohols to olefins as the major products using unstabilized
an alcohol and it is later coupled to a stabilized Wittig reagent, Wittig reagents is unprecedented. Moreover, no oxidants are
usually a carboethoxymethylene-triphenylphosphorane, were required in the reaction reported here.
5
described. Crucially, the large majority of possible Wittig
We have recently reported a range of Ru pincer complexes
reagents are unstable under these conditions, as they them- that catalyse the acceptorless dehydrogenation of various sub-
7
8
9
10
selves react with oxidants, severely limiting the application of strates to form esters, amides, and carboxylic acid salts,
6
8,9
these methods. Even one of the mildest oxidants, oxygen, is from alcohols, amines and their reactions in organic solvents
incompatible with most ylides and causes formation of carbonyl and water. In the mechanism for many of the transformations,
compounds which subsequently react with remaining ylide. One- it is assumed that an aldehyde intermediate that may, or may
1
0
1
1
pot oxidative coupling of benzyl alcohols with Wittig reagents to not be released to solution reacts with another equivalent of
alcohol, with an amine, or with water, to give an intermediate
a
Okinawa Institute of Science and Technology Graduate University, 1919-1 Tancha,
Onna-son, Kunigami-gun, Okinawa, Japan 904-0495
that is further dehydrogenated towards the final product.
A recent report from our group describes the direct olefination
of alcohols with sulfones. However, the reaction is limited to
benzylic and allylic alcohols, while aliphatic alcohols give
complicated mixtures, and only methyl and phenyl sulfones
b
Department of Organic Chemistry, Weizmann Institute of Science, Rehovot, 76100,
Israel. E-mail: david.milstein@weizmann.ac.il
†
Electronic supplementary information (ESI) available: NMR spectra of pro-
ducts, discussion of experimental outcomes and identification of isomers. See
DOI: 10.1039/c5cc02902c
1
2
were reported. In light of these intriguing results, we decided
This journal is ©The Royal Society of Chemistry 2015
Chem. Commun.

View Article Online
Communication
ChemComm
of some Wittig salts during the reaction. It’s likely that the
bulky environment around the metal center significantly raises
the energy of the PPh or OPPh bound Ru(II) species, leading to
3
3
preferential coordination of the less bulky alcohol species.
The power of the transformation is best demonstrated by
entry 7, where a high isolated yield of the product olefin was
Scheme 1 First reaction demonstrating the validity of the olefination of obtained, with no hydrogenation of the terminal double bond
alcohols using Wittig reagents.
14
or its migration to another position (activated olefins are
1
5
isomerized in entries 5, 12 and 15) during the reaction
and the geometry of the inner bond is mostly Z, ruling out
to explore the reactivity of alcohols with Wittig salts, which are significant isomerization reactivity with the metal complex
ubiquitous and easy to prepare. after the transformation. Notably, nitrile groups do not get
We hypothesized that if an aldehyde is indeed a long-lived hydrogenated under the reaction conditions (entry 9) as Ru
intermediate species in the catalytic process, either liberated complexes are known to be active in the hydrogenation of
1
6
into solution or existing bound to the metal, then it may have nitriles when a hydrogen source is present, and amine
an opportunity to react with an in situ formed ylide to form the moieties are tolerated (entries 17/19). Activated benzyl olefins
product olefin (Scheme 1).
give E stereochemistry, possibly via a catalyzed rearrangement
1
5
Equimolar amounts of an aliphatic alcohol, an aliphatic Wittig after the initial reaction as the products of Wittig reactions in
1
7
salt, and a slight excess of base in dioxane solvent were refluxed the absence of lithium cation are overwhelmingly Z. In
under argon, with efficient removal of the hydrogen gas formed. A Table 2, lower catalyst loadings led to higher amounts of the
slight excess of base is needed in order to deprotonate complex 1 Z isomer (Table 2; entry 4), and at 1% loading the Z isomer is
8–11
to the dearomatized catalyst.
Gratifyingly, the first reaction present after 1 hour, but cannot be detected after a 2 hour
tried, using n-hexanol and n-butyl-triphenylphosphonium bromide reaction time. Purely thermal rearrangement towards the more
t
with 1 mol% of complex 1 and 1.2 eq. KO Bu in refluxing dioxane thermodynamically stable E isomer at the activated benzylic
for 24 h (Scheme 1) was successful, giving B90% conversion of the position and an initial higher preference for the E isomer in the
alcohol and an 85% yield of the corresponding olefin, with no case of benzylic substrates cannot be completely ruled out. In
1
5
alkane formation, as determined by GC and GC/MS, with the NMR light of earlier reports of Ru catalyzed olefin isomerization,
being consistent with 4-decene product only. Interestingly, no it’s notable that our catalyst’s isomerization activity seems
traces of ester or other alcohol-derived products were found in limited to activated benzylic double bonds, as the isomerisation
this reaction. The NMR spectrum of the crude mixture showed reaction is undesirable when attempting to synthesize substrates
that 72%/28% Z/E isomers were formed, as is expected from the such as those in entry 7. Minor products in the case of benzylic
basic stereochemistry of the Wittig reaction.
substrates included the 2-aryl-ene isomerisation product that
With this first result in hand, we explored the scope of the may be accessed kinetically during the reaction. In all cases,
reaction, as well as its stereochemistry. After a short initial these species are readily identifiable by NMR as the peaks shift
screening it was found that in order to get consistent results upfield and the coupling pattern is altered dramatically, with the
and accurately estimate conversion, the alcohol should be full product breakdown being reported in the ESI.†
taken as the limiting reagent, with a slight (1.1 eq.) excess of
the Wittig salt and accordingly a 1.2 eq. of base. In examples the benzylic but electron poor bis-trifluoromethyl Wittig
5–21 in Table 1, it was found that 1.7 eq. of base gave better reagent of entry 4 shows good selectivity towards the Z product.
results, perhaps as a result of some acidic impurities in the A rearrangement towards the more stable benzylic conjugated
In contrast with the above-mentioned isomerisation activity,
1
1
3
substrates.
double bond is seen in entries 5 and 15, while in entry 20 the
The reactions were carried out in refluxing dioxane using activated benzylic double bond is completely hydrogenated
1
mol% of complex 1 for 14 h (Table 1). GC yields were and no trace of diene was observed; however, the resulting
calibrated with an internal standard and were further con- 3-heptene position is too isolated for subsequent rearrangement
firmed by isolated yields. The reaction proved robust, often to the benzylic position. A more detailed discussion of most of
giving full conversion of the alcohol, with the only products of the entries, including spectroscopic NMR evidence, isomer
the reactant alcohol being the resulting olefins and in some breakdown and isolation procedures, is available in the ESI,†
cases small amounts of hydrogenated products, except in entry where applicable.
9, where 14% of hydrogenated product is seen. However, only
With the above results in hand, we chose a reaction that
small amounts of this alkane are observed under milder con- appeared to be the most active, in order to test the limits of
ditions (vide infra Table 2). By comparison, it is notable that in catalyst loading. Entry 9 in Table 1 consistently gave quantitative
4
the iridium catalyzed chemistry earlier reported by Williams
yields, but was prone to some hydrogenation of the double bond,
alkanes are the major products and olefins are trace or minor leading us to try to optimize the outcome by lowering reaction
species. Noteworthy, this high catalytic activity took place temperature, and decreasing the catalyst loading (Table 2).
despite the large amount of OPPh
3
formed, and to a much
Lowering the reaction temperature with THF solvent leads to
lesser extent PPh
3
present in solution due to the decomposition lower yield and rate, and more hydrogenated byproduct (entry 3).
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2015

View Article Online
ChemComm
Communication
Table 1 Catalytic olefination of alcohols using Wittig reagents
t
a
#
1
2
Alcohol
Wittig salt
KO Bu eq.
1.2
Conversion%
Product
Stereochem.
Yield (%)
499%
72%Z/28%E
499
1.2
D
Complex mixture
Complex mixture
—
i
3
1.2
499%
89%E/3 others (cis/aryl-2ene)
499[81]
i
b
b
b
4
5
6
7
1.2
1.2
3.0
1.2
82
92
90
99
86%Z/2 others (trans/aryl-2ene)
82[37]
92[33]
90[32]
99[84]
i
82%Z/2 others (cis/aryl-2ene)
d,i
Mostly E (cis/aryl-2ene/dimer)
87%Z
8
9
1.2
1.2
499%
499%
99%E/o1%Z + H
499[81]
c
86%E/14%H
97%E/3%H
499
c
1
0
1.2
499%
499
i
11
12
13
1.2
1.2
1.2
89
77
68
91%E/2 others (cis/aryl-2ene)
89
77
68
95%E/5%Z
One isomer
i
14
15
16
17
18
19
20
21
1.2
1.7
1.7
1.7
1.7
1.7
1.7
1.7
54
74
3 isomers 90/6/4% (aryl-2-ene)
54
i
88%E/2 others (cis/aryl-2ene)
74
25
60%E/40%Z
25
e
499%
99
Large majority Z
99[32]
f
10%Z/27%E (+debrominated)
87%E/8%Z/7%H
37
g
499%
95
60
Mostly Z (see ESI)
95[71]
i
h
499%
75%E/3others (cis/2aryl-2ene)
85
Conditions: alcohol 1 eq.; Wittig salt 1.1 eq.; 14 h vigorous refluxing in dioxane; 1 mol% complex 1. D = decomposition Z = cis, E = trans, H =
a
hydrogenated, Quantitative yield indicated as 499%. Yield determined via GC/FID, corresponds to conversion unless noted, and refers to all olefin
and minor hydrogenated products unless noted. Isolated yield in brackets is after extraction, column chromatography, vacuum concentration, unless
b
c
noted. 200 eq. of alcohol used for entry 6. Low isolated yields due to separation from large amounts of PPh
3
.
Yield determined after extraction in
d
e
Et
2
O, via NMR of crude with 10% cyclooctene internal standard. Mixture of mono (major) and doubly hexylated products; [see ESI]. Product was
isolated via acid/base extraction; [see ESI]. The rest of the products are debrominated stilbenes 18%Z and 45%E. 40% byproduct, [see ESI]. 15%
oxygenated product, cis and 2-butene isomers, [see ESI]. Other isomers are the Z isomer and 2-aryl-ene rearrangement products; [see ESI].
f
g
h
i
Stopping the reaction after one hour (entry 1), results in Significantly, lowering the catalyst loading to 0.02% results in
minimization of the undesired hydrogenation, indicating that quantitative yield of the products after 24 h, with only 1% of
the hydrogenation process is much slower than olefination. the alkane being formed and the rest being the desired olefin.
This journal is ©The Royal Society of Chemistry 2015
Chem. Commun.

View Article Online
Communication
ChemComm
Table 2 Optimization of entry 9 (Table 1) reaction between benzyl
alcohol and 4-cyano benzyl Wittig salt
This research was supported by the European Research
Council under the FP7 framework (ERC 246837), by the Israel
Science Foundation and by the Kimmel Center for Molecular
Design. D.M. holds the Israel Matz Professorial Chair of
Organic Chemistry. E.K. was supported by a half-salary from
the Okinawa Institute of Science and Technology.
a
Time Catalyst
Yield Stereo-
Entry (h) loading (%) Solvent bath T (%)
chemistry
1
2
3
4
1
2
24
24
1
1
1
0.02
Dioxane 110 1C 499
Dioxane 110 1C 499
4%Z 4%H 92%E
13%H 87%E
21%H 79%E
Notes and references
1 G. Wittig and G. Geissler, Liebigs Ann. Chem., 1953, 1, 44–57.
THF 72 1C
78
b
Dioxane110 1C 499
7%Z 1%H 92%E
2
A. D. Abell and M. K. Edmonds, in Organophosphorus Reagents, ed.
Q = quantitative Z = cis E = trans H = hydrogenated. Yield determined by
P. J. Murphy, Oxford University Press, Oxford, UK, 2004, pp. 99–127.
3 A. D. Abell and M. K. Edmonds, in Modern Carbonyl Olefination, ed.
T. Takeda, Wiley-VCH, Weinheim, Germany, 2004, ch. 1.
a
GC/FID except for last entry. Yield is the same as conversion for this
b
reaction. Yield determined after extraction by ether, a quick silica gel
column in ether (to remove OPPh
spectrum of the resulting crude mixture; 10% cyclooctene internal
standard.
3
), vacuum concentration and an NMR
4 (a) M. G. Edwards and J. M. J. Williams, Angew. Chem., Int. Ed., 2002,
41(24), 4740–4743; (b) M. G. Edwards, R. F. R. Jazzar, B. M. Paine,
D. J. Shermer, M. K. Whittlesey, J. M. J. Williams and D. D. Edney,
Chem. Commun., 2004, 90–91; (c) S. Burling, B. M. Paine, D. Nama,
V. S. Brown, M. F. Mahon, T. J. Prior, P. S. Pregosin, M. K. Whittlesey
and J. M. J. Williams, J. Am. Chem. Soc., 2007, 129, 1987–1995;
The increasing percentage of the E isomer on going from
entries 1 to 2 (Table 2) with increasing reaction time supports
a mechanism where the Z isomer is primarily formed first, and
is subsequently isomerized under the reaction conditions.
Compared to our earlier reported alcohol and amine coupling
(
2
d) T. D. Nixon, M. K. Whittlesey and J. M. J. Williams, Dalton Trans.,
009, 753–762; (e) P. J. Black, M. G. Edwards and J. M. J. Williams,
Eur. J. Org. Chem., 2006, 4367–4378.
(a) A. G. M. Barrett, D. Hamprecht and M. Ohkubo, J. Org. Chem.,
5
1
2
997, 62, 9376–9378; (b) A. Maiti and J. S. Yadav, Synth. Commun.,
001, 31, 1499–1506; (c) A. R. Bressette and L. C. Glover, IV, Synlett,
7
–9
chemistry, where esters and amides are formed with hydrogen
liberation, this reactivity is not observed, and amide is a minor
byproduct in the case of entry 19 in Table 1 while a tertiary amine is
understandably completely inactive for amide formation (entry 17,
Table 1). Based on our earlier results, it is clear that ester should
be formed from the alcohol in an open system where the
hydrogen is driven off. The catalytic reactivity of the alcohol with
the ylide is more comparable to the earlier reported synthesis of
2004, 738–740; (d) F. R. Pinacho Crisostomo, R. Carrillo, T. Martin,
F. Garcia-Tellado and V. S. Martin, J. Org. Chem., 2005, 70,
1
0099–10101; (e) J. M. Vatele, Tetrahedron Lett., 2006, 47, 715–718;
(
f ) S. Shuto, S. Niizuma and A. Matsuda, J. Org. Chem., 1998, 63,
4489–4493; (g) M. Reid, D. J. Rowe and R. J. K. Taylor, Chem. Commun.,
003, 2284–2285; (h) U. Karama, Z. Al-Othman, A. Al-Majid and
2
A. Almansour, Molecules, 2010, 15, 3276–3280; (i) F. Alonso, P. Riente
and M. Yus, Synlett, 2009, 1579–1582.
6 (a) A. I. Carrillo, L. C. Schmidt, M. L. Marin and J. C. Scaiano, Catal.
Sci. Technol., 2014, 4, 435; (b) E. Y. Lee, Y. Kim, J. S. Lee and J. Park,
Eur. J. Org. Chem., 2009, 2943–2946; (c) J. R. Kona, C. K. King’ondu,
A. R. Howell and S. L. Suib, ChemCatChem, 2014, 6, 749–752.
1
0
carboxylic acid salts, where water or hydroxide ion intercepts
the aldehyde intermediate and ester formation is suppressed.
In conclusion, we have presented a direct, catalytic olefina-
7
Reviews: (a) C. Gunanathan and D. Milstein, Acc. Chem. Res., 2011,
44, 588–602; (b) C. Gunanathan and D. Milstein, Top. Organomet.
Chem., 2011, 37, 55–84; (c) D. Milstein, Top. Catal., 2010, 53,
tion reaction of alcohols using Wittig reagents that liberates H
2
915–923; (d) C. Gunanathan and D. Milstein, Chem. Rev., 2014,
gas, does not require an oxidant, and is compatible with both
aliphatic and aromatic alcohols. A number of alcohol sub-
strates were tested and showed selective transformation to
the desired olefin. This type of reaction may be attractive for
industrial processes and fine chemicals as it allows to skip an
oxidation step, often saving time and reagents and avoiding the
use of stoichiometric amounts of potentially toxic oxidants.
Another attractive feature is that the use of non-stabilized ylides
is enabled via this protocol, increasing the diversity of products
that can be generated.
1
14, 12024–12087; (e) C. Gunanathan and D. Milstein, Science,
2013, 341, 1229712.
J. Zhang, G. Leitus, Y. Ben-David and D. Milstein, J. Am. Chem. Soc.,
2
C. Gunanathan, Y. Ben-David and D. Milstein, Science, 2007, 317,
790–792.
0 (a) E. Balaraman, E. Khaskin, G. Leitus and D. Milstein, Nat. Chem.,
8
9
005, 127, 10840–10841.
1
2
013, 5, 122–125; (b) for example relevant to Table 1 entry 21
utilizing similar chemistry, see: E. Khaskin and D. Milstein,
ACS Catal., 2013, 3, 448–452.
1 M. Montag, J. Zhang and D. Milstein, J. Am. Chem. Soc., 2012, 134,
1
2 D. Srimani, G. Leitus, Y. Ben-David and D. Milstein, Angew. Chem.,
Int. Ed., 2014, 53, 11092–11095.
3 Many of the reactants were not specially purified after purchase or
preparation, and out of the alcohols only benzyl alcohol and
hexanol were distilled, showing the versatility of the method.
1
1
1
0325–10328.
The reaction proceeds via dehydrogenation of the alcohol by
the catalyst and subsequent release of H
2
, which is then
released from the system under vigorous reflux. The ylide
captures a putative aldehyde intermediate before it can be 14 S. W. M. Crossley, F. Barabe and R. A. Shenvi, J. Am. Chem. Soc.,
2
014, 136(48), 16788–16791.
5 C. R. Larsen and D. B. Grotjahn, J. Am. Chem. Soc., 2012, 134,
0357–10360.
transformed to the ester product. The transformation is tolerant
of a number of functional groups, requires low catalyst loading,
1
1
results mostly in the Z isomer (and E for benzylic positions), 16 (a) R. Reguillo, M. Grellier, N. Vautravers, L. Vendier and S. Sabo-
Etienne, J. Am. Chem. Soc., 2010, 132(23), 7854–7855; (b) S. Werkmeister,
C. Bornschein, K. Junge and M. Beller, Eur. J. Org. Chem., 2013,
forms only traces of the alkane, and is carried out under mild
conditions. We are currently examining the full substrate scope
3671–3674.
of this transformation.
17 P. A. Byrne and D. G. Gilheany, Chem. Soc. Rev., 2013, 42, 6670–6696.
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2015